Factor Financial Analytics Pty Ltd v Xero Limited - [2021] APO 32

IP AUSTRALIA
AUSTRALIAN PATENT OFFICE
Factor Financial Analytics Pty Ltd v Xero Limited [2021] APO 32

Patent Applications: 2017305494

Title:Network-based automated prediction modeling

Patent Applicant/Patentee:     Xero Limited
Requestor:  Factor Financial Analytics
Delegate:  Dr N. R. Madsen – Deputy Commissioner of Patents
Decision Date:  24 August 2021
Hearing Date:  25 May 2021, in Canberra (by videoconference)

Catchwords: PATENTS – generating scores on the basis of performance of entities in a transaction network – section 36 – section 113 – entitlement – application for a declaration of an eligible person – request for a direction for application to proceed under joint applicants – confidential material under sub-regulation 4.3(2)(b) – application under s36 and request under s113 refused – costs awarded against requestor with a variation to the scale

Representation: Counsel for the applicant: Angus Lang SC
Solicitor for the applicant: Gilbert & Tobin
Patent attorneys for the applicant: FB Rice

Patent attorneys for the requestor: Krouzer IP

IP AUSTRALIA
AUSTRALIAN PATENT OFFICE
Patent Applications: 2017305494

Title:Network-based automated prediction modeling

Patent Applicant/Patentee: Xero Limited
Date of Decision: 24 August 2021
DECISION
The applicant, Xero Limited, is the sole person entitled in respect to the present application.
I refuse the application for a declaration under s36(1) and refuse the requested direction under s113(1).
I provide the parties two (2) months from the date of this decision to provide detailed reasoning as to the relevant parts of this decision and the identified list of documents provided prior to the hearing, that should not be open to public inspection.
I award costs according to Schedule 8 against Factor Financial Analytics. Further, I apply a variation to the scale of costs against Factor Financial Analytics to the extent of five times item 11 of Schedule 8 Part 1 ($5000), and full costs for two days of attendance for item 13 ($1800) and item 14 ($2700).
REASONS FOR DECISION
BACKGROUND

Xero Limited (“the applicant/Xero”) filed patent application 2017305494 on 3 August 2017 as an international application under the Patent Cooperation Treaty. The application claims priority from a US application filed on 4 August 2016. Prior to national phase entry, on 8 October 2018, Factor Financial Analytics Pty Ltd (“the requestor/Factor”) filed an Application to the Commissioner for a Declaration of an Eligible Person under s36(1) of the Patents Act, along with a Request for a Direction for the Application to Proceed in the Name of Joint Applicants under s113(1). While a direction of the Commissioner issued on 11 May 2021 directing the applicant to request examination, the application is yet to be examined.
The requestor filed evidence in support of their application for a declaration shortly following their request. Without elaborating upon the details at this stage, the matter has been subject to extensive correspondence and procedural steps in relation to evidence gathering, consideration of confidential material in the evidence, and general procedural issues regarding extensions of time and opportunities for filing various stages of evidence. The resolution of some of these issues occurred in an interim decision of the Deputy Commissioner of Patents in Factor Financial Analytics Pty Ltd v Xero Limited [2021] APO 2 (“interim decision”). In that decision the Deputy Commissioner determined that certain material was not to be considered as evidence in the present matter, while also refusing to give a direction that further evidence could not be filed without leave.
Prior to the hearing, the requestor sought to adduce further evidence that reflected the evidence considered by the delegate in the interim decision and sought to amend both the application for a declaration under s36(1) and the request for a direction under s113(1) that were filed initially on 8 October 2018. I invited discussion of these issues along with the substantive issues at the hearing, noting that the hearing notice sent to the parties, and early correspondence by a Delegate of the Commissioner, indicated that the hearing was to be in relation to s36 only. As it turns out, it is possible for me to provide a final determination of both the application and the request in this decision.
Xero was represented in person at the hearing by counsel, and Factor chose not to appear to present oral submissions, instead relying upon their summary of submissions filed before the hearing.
SUB-REGULATION 4.3(2)(b)
Throughout the progress of this matter the parties have sought to invoke sub-regulation 4.3(2)(b) for certain elements of the filed evidence. This sub-regulation provides for the Commissioner to prevent certain materials from being open to public inspection should she consider there to be reasonable grounds for doing so. Interim directions were put in place by the Deputy Commissioner to the extent that certain documents would be considered non-OPI until such time that a final determination could be made as to appropriate matter for a permanent direction. Prior to the hearing and in light of the large volume of documents filed throughout the process, on 6 May 2021, a delegate of the Commissioner wrote to the parties referring generally to the evidence noting that in light of the various request for information to be made non-OPI, to ensure that any further direction under regulation 4.3(2)(b) is comprehensive:
“…the parties are requested to provide the Commissioner with a complete list of any filed material they consider should be made non-OPI (consistent with the Commissioner’s previous indications in this regard) but which has not yet been the subject of a permanent regulation 4.3(2)(b) direction.”
The applicant obliged in filing their summary of submissions for the hearing, providing a list of documents to be the subject of such a direction. Following this, the Deputy Commissioner again issued further interim directions covering this complete list. At the hearing I made it clear that in my considerations following the hearing, I may write to the parties to progress final determination of non-OPI documents, or alternatively, I may leave determination of this matter until after any final decision regarding the s36 matter, potentially issuing this decision itself non-OPI.
In the course of deciding the matter, I did not see an appropriate opportunity to seek clarification on the list of documents that were sought to be non-OPI, nor any particular elements of the decision that the parties considered should be non-OPI. Thus, the decision is issued to the parties with no material redacted and the parties are invited, before publication of the decision, to identify certain material in the decision that should be redacted along with certain material in the list of documents and provide detailed reasons as to the need to make the information not open to public inspection.
The APPLICATION FOR A Declaration under s36(1)
The following is the grounds for the application for a declaration currently before the Commissioner, being filed on 10 May 2021. While this application seeks to amend the fundamentals of Factor’s case when compared to the originally filed application, I consider it sensible to consider this amended application in this decision as it reflects the submissions made by Factor, and Xero substantively addressed the relevant issues in the hearing. The amended application for a declaration is as follows:
“We, Factor Financial Analytics Pty Ltd (Requestor), Paul Anthony Reynolds (Reynolds) and Aaron Wallace (Wallace), apply for a declaration referred to in Section 36(1) of the Act stating that:
we are eligible persons along with the nominated person
in respect of Australian Patent Application 2017305494 (Application) in the name of the applicant Xero Limited, who is the nominated person, which has not been granted.”
The particulars upon which the requestor makes its case are as follows:
a. The inventive concept of the invention of the Application was initially conceived during the period March to July 2014 (first phase) and then expanded upon the inventive concept in more detail from August 2014 to January 2015 (second phase). The inventive concept was further expanded upon from February 2015 to August 2016 (third phase) and then was developed to a working prototype from September 2016 to July 2017 (fourth phase).
b. Reynolds and Martin Kemka (Kemka) initially conceived the invention and became partners of a partnership called Decise during the first phase. The partners of Decise then dealt with Alistair Grigg (Grigg) of the nominated person during the second phase. The partners of Decise by way of an agent and individually, including Wallace entered a formal arrangement with the nominated person during the third phase. Wallace joined Decise during the third phase. The Requestor was established during the third phase by the partners of Decise with the agent consulting for them and then entered into a formal arrangement with the nominated person to undertake the development of the invention on behalf of the Decise partners during the fourth phase.
c. Kemka assigned all of his rights in relation to the invention to the nominated person on 3 August 2017 and relinquished his entitlement as a partner of Decise to the Application, leaving only Reynolds and Wallace having entitlement by way of Decise.
d. Grigg was always an employee of the nominated person and never had any entitlement to the invention, but provided entitlement to the nominated person.
e. The Requestor established its own entitlement to aspects of the invention conceived by Kemka and Reynolds for its own benefit that were excluded from the agreement with the nominated person and attributed other aspects of the invention conceived by Kemka during the development of the invention for the benefit of the nominated person in accordance with the agreement.
The Requested DIRECTION under s113(1)
Factor similarly filed an amended request for a direction under s113(1) on 10 May 2021. Again, I consider it sensible to consider this request in this decision for the reasons above.
“We, Factor Financial Analytics Pty Ltd (Requestor), Paul Anthony Reynolds (Reynolds) and Aaron Wallace (Wallace), interested parties (Interested Parties) in respect of Australian Patent Application 2017305494 (Application) in the name of the applicant Xero Limited (Applicant), request a direction of the Commissioner under Section 113 (1) of the Act that the Application proceed in the name of the Interested Parties and the Applicant.”
APPLICABLE LAW
Section 36 applies to the application in this case and opens in the following terms.
(1)If:
(a)   a patent application has been made and, in the case of a complete application, the patent has not been granted; and
(b)   an application for a declaration by the Commissioner is made by one or more persons (the section 36 applicants) in accordance with the regulations; and
(c)   the Commissioner is satisfied on the balance of probabilities, in relation to an invention disclosed in the specification filed in relation to the application for the patent:
(i)that the nominated person is not an eligible person, but that the section 36 applicants are eligible persons; or
(ii)that the nominated person is an eligible person, but that the section 36 applicants are also eligible persons;
the Commissioner may declare in writing that the persons who the Commissioner is satisfied are eligible persons are eligible persons in relation to the invention as so disclosed.
Schedule 1 of the Patents Act defines an eligible person as follows:-
“eligible person”, in relation to an invention, means a person to whom a patent for the invention may be granted under section 15.
Subsection 15(1) provides that a patent may only be granted to a person who:
(a)is the inventor; or
(b)would, on the grant of a patent for the invention, be entitled to have the patent assigned to the person; or
(c)derives title to the invention from the inventor or a person mentioned in paragraph (b); or
(d)is the legal representative of a deceased person mentioned in paragraph (a), (b) or (c).
Accordingly, the primary person to whom a patent may be granted is the inventor. All other persons may derive title to the patent, and the invention claimed therein, via the inventor; Stack v Davies Shephard Pty Ltd, [2001] FCA 501 at [21].
In University of Western Australia v Gray (No 20), [2008] FCA 498 (“UWA”), French J commented, at [1426], that the inventive concept marks a boundary between invention and verification. French J then referred to the US Court of Appeals Federal Circuit decision, Burroughs Wellcome Co v Barr Laboratories Inc, [1994] USCAFED 1125, (1994) 40 F.3d 1223. He stated that this decision set out the following principles, amongst others, as established by US appellate courts:
1.Conception is the touchstone of inventorship, the completion of the mental part of inventions.
2.Conception is the “formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention as it is hereafter to be applied in practice”. It is complete only when the idea is so clearly defined in the inventor’s mind that only ordinary skill would be necessary to reduce the invention to practice without extensive research or experimentation.
On appeal, the Full Court of the Federal Court, in University of Western Australia v Gray, [2009] FCAFC 116, did not appear to disturb French J’s position on the above points. The Full Court, at [253], also appeared to accept the following steps from the primary judge’s decision as appropriate:-
1.identify the “inventive concept” of each relevant invention as defined by the claims;
2.determine inventorship including the person responsible for the inventive concept and the time of conception as distinct from its verification and reduction into practice; and
3.determine how many contractual or fiduciary relationships give rise to proprietary rights in the invention.
In JMVB Enterprises Pty Ltd v Camoflag Pty Ltd, [2005] FCA 1474, (2005) 67 IPR 68 (“JMVB”), Crennan J stated as follows at [132]:
“Rights in an invention are determined by objectively assessing contributions to the invention, rather than an assessment of the inventiveness of respective contributions. If the final concept of the invention would not have come about without a particular person’s involvement, then that person has entitlement to the invention. One must have regard to the invention as a whole, as well as the component parts and the relationship between the participants. The fact that the parties were in collaboration can be a major consideration.”
The passage was cited with apparent approval by the Full Court of the Federal Court in Polwood Pty Ltd v Foxworth Pty Ltd, [2008] FCAFC 9 at [54]. Furthermore, in Polwood at [34] - [36] (emphasis in original):-
“The entitlement to the grant of a patent as the inventor is not determined by quantitative contribution. The role of joint inventors does not have to be equal; it is qualitative rather than quantitative. It may involve joint contribution or independent contributions. The issue is whether the contribution was to the invention. What constitutes the invention can be determined from the particular patent specification which includes the claims. In some cases, evidence can assist. In some cases, the reduction of a concept to a working apparatus by a person may not be part of the invention, in other cases it may be. For example, the construction of an apparatus may involve no more than carrying out the instructions in the specification. This would not normally entitle that person to joint inventorship. On the other hand joint inventorship may arise where the invention is in the apparatus itself, or where the person constructing the apparatus contributed to a different or better working of it which is then described and claimed.
One criterion for inventorship may be to determine whether the person’s contribution had a material effect on the final invention. It may be that an invention is made as part of a collaborative effort. In those circumstances, it would ordinarily follow that the collaborators are joint inventors of the product of the collaboration. ….
To ascertain the inventor for the purposes of entitlement to the grant of the patent it is therefore necessary to determine the contributions to the invention described in the patent application. The claims may assist in that determination, bearing in mind that the claims may be to less than the totality of the invention. It may also be appropriate to investigate the contributions to the inventive steps giving rise to the invention.”
Section 113 provides a mechanism by which a person may claim entitlement to a patent before grant under assignment or agreement, and largely functions to enable the correct recording of applicants, facilitated by the filing of assignment documents. In the context of the present matter, the request under section 113 is essentially procedurally redundant in that I must perform the same considerations to determine whether the evidence establishes entitlement, where the potential outcomes of positive findings as to entitlement of the requestor under section 36 and section 113 differ slightly. For completeness, section 113(1) is as follows:
(1) Where, before a patent is granted, a person would, if the patent were then granted, be entitled under an assignment or agreement, or by operation of law, to:
(a) the patent or an interest in it; or
(b) an undivided share in the patent or in such an interest;
the Commissioner may, on a request made by the person in accordance with the regulations, direct that the application proceed in the name of the person, or in the name of the person and the applicant or the other joint applicant or applicants, as the case requires
EVIDENCE
A large amount of evidence has been filed by both parties. Both parties have provided evidence in support of their position and reply evidence. Shortly before the hearing, and outside of any direction for the filing of evidence, the requestor also filed a further declaration seeking consideration consistent with Regulation 5.23. I wrote to the parties before the hearing noting that consideration of admission of the evidence could be discussed at the hearing with any decision potentially being deferred as part of final decision making arising from the main hearing.
The Deputy Commissioner usefully summarised the roles of the various parties in the context of the entitlement dispute at [3] of his interim decision. I add to this below.
·     Alastair Grigg was a General Manager at Xero at the relevant times and is a named inventor.
·     Martin Kemka was a product research and development analyst at the requestor and is also a named inventor.
·     Benjamin Styles was an employee of Xero from November of 2014 and became General Manager, Product – Financial Services at Xero around October of 2017
·     Paul Reynolds was the Chief Executive Officer and Managing Director of the requestor and is argued by the requestor to be entitled to the invention.
·     Aaron Wallace was a Director of the requestor. He described himself in his first declaration as an "accounting data specialist" with "accounting data knowledge" and is also argued by the requestor to be entitled to the invention.
Requestor’s evidence in support
The requestor’s evidence in support consists of:
·     A first declaration of Aaron Wallace dated 23 October 2018 (Wallace #1)
·     A first declaration of Paul Reynolds dated 22 October 2018 (Reynolds #1) with exhibit PAR-1
·     A second declaration of Aaron Wallace dated 15 April 2019 (Wallace #2) with exhibits AW-1 to AW-3
·     A second declaration of Paul Reynolds dated 15 April 2019 (Reynolds #2) with exhibits PAR-1 to PAR-11
Applicant’s evidence in support

The applicant’s evidence in support consists of:
·A declaration of Alastair Grigg dated 10 September 2019 (Grigg) with exhibits AG-1 to AG-15
·A first declaration of Martin Kemka dated 10 September 2019 (Kemka #1) with exhibits MK-1 to MK-26
·A declaration of Benjamin Styles dated 10 September 2019 (Styles) with exhibits BS-1 and BS-2
Requestor’s evidence in response
The requestor’s evidence in response consists of:
·A third declaration of Aaron Wallace dated 28 January 2020 (Wallace #3) with exhibits AW-1 to AW-11
·A third declaration of Paul Reynolds dated 10 January 2020 (Reynolds #3) with exhibits PAR-A to PAR-P
·A fourth declaration of Paul Reynolds dated 7 February 2020 (Reynolds #4) with exhibits PAR-1 to PAR-22
Applicant’s evidence in response
The applicant’s evidence in response consists of:
·A second declaration of Martin Kemka dated 30 April 2020 (Kemka #2) with exhibit MK-1.
Requestor’s further evidence
Factor’s above mentioned further evidence request provides the following declaration:
·A fifth declaration of Paul Reynolds dated 8 May 2021 (Reynolds #5) with exhibits PAR-1 to PAR-13
Regulation 5.23 enables the Commissioner of Patents to consult a document that is relevant to an opposition, has not been filed in evidence, and is available in the Patent Office. Since the present matter is not strictly an opposition, regulation 22.24 is appropriate in these circumstances. The latter regulation enables the Commissioner to determine the practice and procedure to be followed for the purposes of enabling an application or matter to be decided that is not an opposition. This includes the filing of evidence. In his interim decision, the Deputy Commissioner outlined the relevant principles in considering information under regulation 5.23 (see [7]-[12]). In short, the following are relevant matters for me to balance in deciding wither the Reynolds #5 declaration forms further evidence in the present matter. These are:
1. The circumstances leading up to the evidence not being filed earlier
2. What does the evidence show
3. Is the information likely to be crucial to the delegate's decision
4. The public interest in having the information considered
5. The balance of convenience of the parties if the information is considered
For reasons that will become clearer later in the decision, I determine that the fifth declaration of Paul Reynolds is not evidence that can be relied upon by the parties in the present matter. At this point it is useful to note that the balance of convenience and the delay weighs strongly against incorporation of the evidence, with the content of the evidence being most relevant to my decision to not consider the evidence formally as part of the matter.
Weight to give to certain evidence
Factor’s written submissions seek to identify a conflict of interest for Mr Kemka and suggest that this should be taken into account in considering the probative value of his evidence. Factor suggest that he has conflicting fiduciary duties that should diminish the weight given to it in determining any question of fact. To put it briefly, Factor state that:
“In particular, Mr Kemka’s opinion concerning the inventive contributions of Mr Reynolds and Mr Wallace in both of his declarations forming part of the Applicants PEIA and AEIR need to be disregarded or given very low weight, especially having regard to Mr Kemka’s association of invention contribution (sic) being coloured by the degree of ‘technical’ and scientific contribution.”
It is important to note that a clear understanding of this allegation of a conflict of interest is intertwined with the considerations regarding the key elements of the entitlement considerations being those identified above in accordance with UWA. That aside, and before I seek to address these elements, I will provide a summary of what appears to be undisputed in relation to the general history between the parties. That being said, as it will be seen, I see no reason to view Mr Kemka’s relevant evidence to be of lesser probative value as posed by Factor.
COMMON NARRATIVE REGARDING THE PARTIES
A useful point in time at which to begin the narrative is late 2013 to early 2014 where Martin Kemka was introduced to Paul Reynolds (Kemka #1 at [11]). At the time, Mr Reynolds had been working for the International Finance Corporation (IFC) investigating extending financial services to Indonesia’s micro-small medium enterprises segment, which evolved into the development of a business model that leveraged new digital technology. This technology included inventory management systems, mobile-based sales ordering applications and mobile-based e-money accounts. During this period, Mr Reynolds discusses that he developed the concept of using newly created data captured to support integrated e-invoice payments, credit scoring for trade finance, and for marketing insights (Reynolds #3 at [13]).
Kemka’s introduction to Reynolds coincides with Kemka joining the IFC on a consultancy basis, in relation to modelling and credit scoring capabilities. This included conducting a feasibility study to build a trade credit risk model using Fast Moving Consumable Goods (FMCG) data (Reynolds #3 at [22(f)-(g)]). Kemka and Reynolds subsequently travelled and worked together as part of this endeavour before Kemka ceased his work with the IFC (Reynolds #3 at [23(a)-(b)]). Kemka’s work had involved data analysis, applied credit scoring, data research and model training (Kemka #1 at [13]). Around this same time, Kemka had independently commenced building a decisioning platform which he had named “Decise”, the platform being an application programming interface for hosting models. He was particularly interested in using machine learning to develop models that could be hosted on the platform which could in turn be effectively maintained and upgraded as data relating to each of the models evolved. This was work Kemka undertook himself without involvement of Reynolds (Kemka #1 at [14]).
Early in 2014, Kemka indicated that he was curious how the work he was doing for IFC could be applied to a local business and considered that data held by Xero may be useful as it represented a significant interconnection of business entities and actions (Kemka #1 at [15]). Kemka and Reynolds engaged in a number of discussions between early and mid-2014 regarding the potential for developing tools for analysing payment data for credit purposes. During the period March -July 2014, Reynolds discussed with Kemka a business idea he had relating to trade credit programs utilising big data (Reynolds #3 at [24]) following which a range of discussions occurred as to the formation of a business together, identifying business problems and potential solutions for trade credit (Reynolds #3 at [25]). They discussed that Australian companies were rapidly adopting accounting software and that this could be a solution to the access of valuable data for scoring purposes. At this stage Kemka introduced Reynolds to the “Decise” software platform and they furthered this discussion to the idea of potentially using the platform in solving big data credit model challenges (Reynolds #3 at [26]), with Kemka and Reynolds subsequently identifying Xero as holding significant trade network data that was suitable for trade network concepts. Following this they agreed to present to Xero under the business name of Decise, which is posed by the requestor as a partnership (Reynolds #3 at [27]).
Later, Kemka introduces Reynolds to Mr Wallace who was Kemka’s accountant and a former employee of Xero (Kemka #1 at [17]; Reynolds #3 at [30]). Mr Wallace subsequently assists in reaching out to Xero and between August and September of 2014, Kemka and Reynolds meet with Mr Grigg of Xero to discuss the potential for Xero network data to be used to enhance credit approval algorithms and help small business secure lending (Kemka #1 at [19]). [Passage not published by reason of confidentiality].
Between February and April of 2015, Kemka performed the work of the feasibility study with the assistance of Grigg on site at Xero premises using a sample set of New Zealand Xero customer data (Kemka #1 at [40]-[50]). It appears Wallace provided answers to questions of Kemka during the process (Kemka #1 at [51]), and Reynolds had little involvement in this particular work (Kemka #1 at [22]). Kemka prepared a report on the study in April of 2015 and presented it to Xero (Kemka #1 at [56]).
After the feasibility study, with business potential increasing in relation to the relationship between Reynolds and Kemka, on 20 October 2015, Reynolds registered Factor Financial as a company (Kemka #1 at [63]). Kemka registered the logo and trademark for Factor Financial in the name of Northraine in May of 2016 (Kemka #1 at [65]). Recognising Wallace’s accounting knowledge and value to the business, Reynolds, Kemka and Wallace entered into a shareholder’s agreement for Factor (Kemka #1 at [67]). Kemka notes at this stage of the evidence that intellectual property relating to his developed “Decise engine” was assigned to Factor (Kemka #1 at [67]). All three shareholders of Factor then agreed that Factor Financial, rather than Northraine, would be engaged to undertake future work of Xero. This arrangement was reflected in a professional services agreement (PSA) dated 1 September 2016 between Factor and Xero having associated statements of works (SoWs) (Kemka #1 at [70]). A range of work was delivered under these statements of works.
On or about 6 October 2017, Kemka ceased as a director of Factor Financial. He subsequently consulted with Xero under Northraine around November 2017 (Kemka #1 at [88]), and following this, around August of 2018 was employed by Xero four days per week remaining working for Northraine one day per week (Kemka #1 at [91].
This narrative is useful in understanding the conception of the relevant invention and the nature of the flow of entitlement therefrom. The narrative appears to outline events and issues that are not in dispute.
SPECIFICATION
First half of specification
The specification is particularly extensive covering 100 pages and referring to 33 figures. It introduces, by way of background information, that computer systems generate numerical representation models for individuals, groups or business, and these models may be used in determining whether to take a particular action in relation to an entity for which the model was generated. At paragraph [0003] of the specification it is noted that known numerical representations may lack context, and that they lack predictive value for actions taken by the individual for which the representation was generated, particularly regarding the timing of actions. It is then noted that there is a need in the art to improve generation of predictive models capable of identifying probabilities of actions and probabilities of actions occurring with respect to identified periods of time. As I understand this more specifically with reference to the body of the specification, this may include probabilities that an invoice will be paid by an entity within a certain period of time.
After summarising the content of the various figures, the specification provides a detailed description of the invention from paragraph [0043]. Issues are discussed in recognising that sales receipt forecasting is important knowledge that is useful for businesses, and that timeliness of invoice payment is a major factor relevant to a business’s cash flow and overall performance. Small businesses typically possess payment history of a particular account and from this they may forecast the probability of on-time payment. However, if a business has limited interaction history with a client, then forecasting becomes difficult. They may check known “credit scores” but this is not necessarily a good indicator of the likelihood of paying invoices.
Paragraph [0046] appears to articulate a solution to the problems identified above by using not only payment histories made to the business, but payments made to other businesses. This is made possible because an accounting platform (denoted in the specification as an “action history system”) that contains data spanning many businesses and networked interactions between these businesses may be used to generate metrics (denoted by the specification as “action scores”) that represent payment histories for a particular entity. These metrics may have a variety of uses in assisting in the making of business decisions, one of those being an indicator useful to a lending institution in assessment of potential loans.
The invention is said to enable the analysis of historical and live date that is also particularly voluminous, comprising potentially thousands of millions of users and inevitably an even higher volume of transactions. The collection of data is made possible by connecting multiple servers and data providers (e.g., accounting platforms, banks, lenders, and users). With the collection and analysis of this data, it is said that the action scores can paint a live picture based on expected cash flow and other financial parameters, the level of risk associated with the cash flow, and the timeliness of interactions of a user/entity. At this stage, the discussion of the invention in the specification remains rather nebulous, before reciting consistory statements for the claims and then discussing Figs. 1 – 4 which prima facie appear to represent standard computer architecture and programming methodology used to implement a networked data management system.
At paragraph [0092] the specification discusses that the accounting system (action history system) comprises algorithms that can gather and process large amounts of information and communicate actions scores to various entities. A general discussion is provided in paragraph [0093] as to the calculation of an action score. Here, the action score is described as a value that provides an indication of the past performance of an entity with respect to a set of performance values. More specifically:
“For example, the action history system 100 may track payment history of the entity to generate an action score that measures average time for paying invoices, percentage of times that invoices are paid after 30 days (or some other time), percentage of times that invoices are paid after 60 days, or other performance values.”
The specification then discusses other functions of the action history system in tracking incoming payments and calculating new action scores. Performance values/metrics may be generated based on types of events such as accounts receivable and accounts payable. From here an action score may be calculated on the basis of a specific event and the evaluated performance metrics. At this point it appears worth nothing that it is not entirely clear how action scores are being calculated. It is clear that they are based on certain performance metrics or values, but no specific mathematical relationship is yet to be defined. I will turn later in my discussion of the specification to a specific equation that is identified for calculation of action scores but for now I will continue with the flow of the specification.
The specification follows by exemplifying a number of action scores that provide useful business intelligence. I summarise these as follows:
·     The network predictive cashflow (NPC) is an action score utilized to assess the future cash flow for the entity. It predicts sales cash flow and in particular, the incoming cash resulting from the sales of the entity ([0098]). This score is based on a predictive model that utilises a network view of contact payments to predict the likelihood of late payments ([0099]). To this extent is seems clear that the probability of late payment of related entities is data required to calculate the NPC. However, at this point it is not entirely clear exactly how the NPC is mathematically calculated. Nonetheless, the NPC is discussed as a useful indicator upon which one may set triggers for action or monitor for understanding of business health.
·     The invoice trade score (ITS) (discussed at [0102] to [0105]) is a score generated by a predictive statistical model that looks at payment performance of sales contracts across all entities. There is general discussion that it is calculated by means of an average score of each invoice score across all the accounts receivable invoices, weighted by debtor values. However as per the NPC, the mathematics of this calculation is not entirely clear. This score provides a measure of risks to cash flow and general financial stability.
·     Another score discussed in the specification (at [0106] and [0107]) is the health action score or HAS. This score is built on the transactions with direct entities (e.g. siblings or cousin companies) and predicts the probability that the entity will have a revenue below a predetermined threshold for a predetermined amount of time. For example, the probability of three consecutive months of zero revenue.   Again, how this is calculated is not yet clear in the specification.
After discussing these scores the specification discusses Fig. 5B. For convenience at this point I also provide Fig. 30 for discussion at it appears to depict similar subject matter with both figures being reflective of the content of the independent claims of the specification. Thus, their discussion appears relevant to preliminary considerations of the inventive concept. The figures are largely self-explanatory, but I will nonetheless provide commentary below.
Fig. 5B shows the calculation of action probabilities from action scores. Initially the action history system (or more simply identified as an accounting system) contains data regarding transactions. These transactions are analysed to produce performance metrics and action scores for each entity in the system. A request is made for action probabilities for an event, and related entities are identified. Action scores are evaluated for each of the entities and then these action scores are used to determine action probabilities of each related entity. This information is then presented. Fig. 30 is a little more detailed than the stepped schematic process of Fig. 5B, however it appears to reflect a similar process with the figure being more specific about automatic actions, and also more specific about the presentation of action probabilities on a graphical user interface of a display. As has been discussed earlier, the specification remains elusive about precisely how performance metrics, action scores and action probabilities are calculated. It is noted at [0113] that in some example embodiments:
…the action scores may be based on different parameters or a combination thereof, such as invoice amount, invoice terms, number of invoices with one entity, abridged value of invoice, minimum value of invoices, maximum value of invoices, number of unpaid invoices, average payment terms, average payment performance, number of connections of the entity, number of changes in the connections over time, or distribution of sales volume by connection.
The specification does not describe with clarity, how these parameters may be used. At this point it is useful to accelerate through the specification to the single disclosure of a mathematical relationship between an action score and probability. Suffice to say, there is little between paragraphs [0113] and [0206] that appears to provide further valuable relevance to the inventive concept to be discussed in this decision. In these paragraphs is discussion of a range of other variations to business analytics using gathered data. In this light, I turn to paragraph [0206] which discloses that in some embodiments, the action score is calculated with an equation as follows:
Score = K₁ + Ln(P/(1-P))*(K₂/Ln(2))

This equation is discussed as facilitating the calculation of a score (action score) using P as the probability that an item is in good health (e.g., a receivable). More specifically P might represent a 75% chance of an invoice being paid on time. Variables K₁ and K₂ are a base score and point multiplier respectively, the point multiplier being the number of points that have half or double the odds. As an example calculation, with K₁ and K₂ both set at 100, and a P of 0.75, an action score is calculated as 258. Notably, the specification does not describe a basis for selecting K₁ other than to say numbers other than 100 can be selected. The specification continues at [0210] noting:
In some embodiments, the analysis of the score includes comparing to certain thresholds. A first threshold 1214 is 250, which means that action scores at 250 or less are at higher risk. A second threshold 1218 is 500, which means an average level of risks. A third threshold 1220 is 750 and indicates that action scores above the third threshold will have low risk and a high probability of being healthy.
Paragraph [0211] indicates that real time calculations of action scores are possible. New data may arrive at the action history system which generates a new set of performance values. A scoring component may generate a new action score based on these regenerated performance values and in turn, this new action score can update the action probability based on the new action score (presumedly using the equation above, although this is not made clear). There is no specific detail at this stage describing how the performance values are used to affect the action scores absent the equation above.
It is at this point that I have discussed the specification to a degree that is closely reflective of the claims. Remaining in the specification are paragraphs [212] to [0391] constituting 50 pages and discussion of multiple figures. I will discuss this information further as necessary, but it is useful for understanding at this point to introduce the claims. Claims 1, 10, and 16 are independent representing the same subject matter couched as method, system, and non-transitory machine-readable storage medium comprising instructions. Claims 1-9 are representative of the entirety of the 20-claim set with dependent claims of 2-9 being repeated as dependencies to claims 10 and/or 16.
Claims 1-9
Claims 1-9 of the specification are as follows:
1.   A method, comprising:
accessing, by one or more processors of an action history system and via a network connection, data stored on a database, the data comprising transactions exchanged between a set of entities;
generating, by the action history system, a set of performance values for each entity based on the data;
for each entity of the set of entities, generating an action score based on at least one performance value of the set of performance values, the action score being an indication of performance of the entity with respect to the set of performance values;
receiving a request, from a user, for one or more action probabilities regarding an event associated with a first entity, each action probability being a probability that a second entity responds to the event associated with the first entity within a predetermined time frame;
automatically determining related entities associated with the event;
automatically determining the action probability for each entity of the related entities, the action probability being determined based on the action score of the related entity; and
automatically causing presentation of the action probabilities within a graphical user interface (GUI) of a display device.
2.   The method of claim 1, wherein generating the action score for the first entity further comprises:
tracking the transactions of the first entity with other entities;
evaluating a performance of the first entity with reference to the event based on the     transactions of the first entity with other entities; and
calculating the action score for the event based on the evaluated performance.
3.    The method of claim 2, wherein the action score is a network predictive cashflow (NPC), wherein the transactions include payments made by the other entities, wherein evaluating the performance comprises analyzing a history of payments by the other entities, wherein calculating the NPC comprises predicting a cash flow for the first entity for a predetermined time period based on the history of payments of the other entities.
4.    The method of claim 2, wherein tracking the transactions of the first entity further comprises accessing, from an accounting platform providing accounting services, accounting data of the first entity, wherein evaluating the performance further comprises accessing, from the accounting platform providing accounting services, accounting data of the other entities.
5.    The method of claim 1, wherein causing presentation of the action probabilities further comprises:
identifying groups of related entities based on values of the action probabilities; and
presenting the action probabilities according to the identified groups.
6.    The method of claim 1, further comprising:
receiving a request from an institution for the action score of the first entity; and
transmitting the action score to the institution, wherein the institution determines granting a loan to the first entity based on the action score of the first entity.
7.    The method of claim 1, further comprising:
calculating a second action score for the first entity based on a second event; and
calculating a third action score for the first entity based on the action score and the second action score.
8.    The method of claim 1, further comprising:
generating a response recommendation based on the action probabilities; and
causing presentation of the response recommendation in the GUI.
9.    The method of claim 1, wherein the transactions include a respective time deadline.
Second half of specification
From paragraph [0212] there is discussion of a range of data presentations including tables and graphs. Here Figs. 13-19 depict data such as: probability of invoices being paid late as a function of ranges of action scores; the amount of expected money to be received in various risk category groupings as a function of the month; the amount of money receivable on a weekly basis; the amount of money receivable in different risk categories on a weekly basis; the amount of money receivable on a weekly basis divided into payments expected to be on time and payments expected to be late; and also the number of invoices and total revenue on a monthly basis divided into high and low risk categories. Each of these graphical depictions of Figs. 13-19 appear useful data for showing the status of a business to assist in decision making. I provide below Fig. 14 as an example where such data may also be presented in graph form.
Fig. 20 depicts a flow of an example method of analysing lines of credit in response to the generation of an action score. Figs. 21, 22, 23, 26, 27 and 28 expand upon this figure by providing a range of use cases for the interaction between the action history system and financial institution. Fig. 20 is reproduced above.
The various use cases involve situations such as: a financial services prospecting case where consenting user information is sent to an institution which determines credit worthiness of a user and sends them a direct offer (Fig. 21); a loan assessment data use case where user financial data and action scores are pushed to the institution which determines credit worthiness and credit terms, sends an offer to the user and writes back any liabilities and payments (Fig. 22); a health and trade score analytics case where an institution requests users that fit certain criteria. Following this, a list of consenting users is returned with the NPC score and data is used by the institution for decision making and portfolio management. The financial institution continues to monitor the NPC in connection with the action history system (Fig. 23): a real-time invoice factoring and discounting case where upon invoice creation, alerts are sent to a financial institution with relevant details including action scores and amounts. The financial institution determines credit worthiness and makes credit offers in line with the invoice raised. The user may confirm the offer and details of the loan are returned to the system with funds provided to the users bank account (Fig. 26); a dynamic lending portfolio and monitoring case where a change in NPC is used by the financial institution to trigger a pull of data to try to do things such as predict default or anticipate early repayment, and thereby take proactive steps (Fig. 27); and a payer financing case where a payer is inserted as an intermediary between the action history system and the financial institution, and the payer applies for financing by pushing their data to the financial institution for consideration of loan option (Fig. 28).
Other use cases are envisaged where data is shared by the action history system with entities other than a financial institution such as: with a supplier in a trade credit use case (Fig. 24); data records such as contact records for dynamic setting of payment terms (Fig. 25); and data records of a client management ruleset for dynamic monitoring and modification of accounting actions (Fig. 29). While these figures are discussed in detail [0236] to [266], further discussion at this stage appears unnecessary. Following this there is a discussion of Fig. 30 ending at [0279], Fig. 30 being already summarised above.
The specification then discusses a range of what appears to be standard computing and processing implementation means which while substantial (paragraphs [0281] to [0291]) is exemplified by statements such as:
Example embodiments may be implemented in digital electronic circuitry, or in computer hardware, firmware, software, or in combinations of them.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communications network.
From [0295] to [0299] the specification discusses example financial services arising by connecting the system to bank accounts that may include the automated provisioning of accounts via internet banking, bank-provided services called by accounting software, and the secure transfer of data via feed services. Automated account provisioning may include a user identifying, within internet banking software, which accounts they would like shared with accounting software. This account provisioning can facilitate automated retrieval of statements by the accounting system ([0302]). Following this the specification discusses the management of accounts and authorisation steps to activate accounts between internet banking and accounting software systems. Connection of accounts enables a data feed that can assist in providing suggestions for ledger entries. Connection of accounts also facilitates batch payments and update of terms and services associated with various accounts. A range of data structures and algorithms are also provided to implement the connectivity between financial institution and accounting software via a browser. However, to the extent that they are explained, it appears that they may be standard programming techniques applied to particular information (see specification [316] to [372], and importantly do not relate to content that is pressed by the requestor as being deficient in entitlement within the application.
Of note, the specification substantively ends with discussion of Figs. 32 and 33. Fig. 32 illustrates a process for training and using a machine learning program according to certain embodiments. Fig. 33 is a block diagram of a machine in the example form of a computer system with instructions for performing the described invention. As is self-evident, the figures are quite generic representations of the implementation, and I note there is no specific detail in the specification as to the functioning of the machine learning program aside from a general description of what would appear to be standard aspects of machine learning ([0376] to [0389]).
Having considered the specification for the purposes of understanding the inventive concept, it does strike me as to the nature of the described and claimed subject matter in that the invention focusses upon business analytics, and the specifics of implementation were quite difficult to discern. To this extent, prima facie it appears there may be serious issues to be faced by the specification when examined, in particular under grounds of s40(2) or (3), and s18(1)(a). Although in this decision I make no determination under these grounds, it not being in my power to do so.
INVENTIVE CONCEPT(s)
Requestor’s identification of the inventive concept
Factor generally argues that the invention was developed across a number of phases. Primarily, they argue that the inventive concept of the application was initially conceived during the period March to July 2014 (first phase). This phase coincides with the common narrative discussed earlier where Kemka and Reynolds come together and discuss a business idea to create trade credit programs utilising big data. At this point, Reynolds and Kemka discuss the adoption of accounting platforms in Australia and Kemka introduces Reynolds to his “Decise” software platform. Xero is identified by the pair as holding significant trade network data suitable for the discussed trade network concepts, and as a result they agree to present to Xero under the name of Decise. From Factor’s submissions it is stated that:
“The invention originally conceived by Mr Reynolds and Mr Kemka, which we submit is disclosed in the specification filed in relation to the Application, is referred to in the evidence as the “Trade Network Concept” and was conceived during the first phase identified in the Requestor’s Summary of Case and expanded upon in the second, third and fourth phases.
The principal focus of the Trade Network Concept was creating credit scores for small business and accessing big data that matched invoice and bank transaction data in the one location (Requestor’s AU PA).”
Referring to the Requestor’s evidence in reply, Reynolds notes at 38f) of Reynolds #4 that:
“…Kemka and I contacted the Applicant in the second half of 2014 and presented the concept of a ‘Trade Network” based credit model. In these interactions with the Applicant I was the person who created the content of the presentations and explanation of the ‘Trade Network Concept’, which had evolved in discussions between Kemka and myself as described in previous paragraphs.”
More specifically, in terms of the contribution Reynolds alleges to have made to the invention he notes at 38d) of Reynolds #4 that:
“…as part of undertaking the World Bank research into FMCG distribution, I
observed that each individual small shop entity had multiple FMCG distributor
trade relationships, and to more accurately understand the risk of late or nonpayment of an invoice would be achieved by identifying and tracking the data of each entity of small shop’s payment behaviour with each FMCG distributor in the Trade Network. I presented this concept to Kemka at the very beginning of our relationship in March 2014…”
In his third declaration (Reynolds #3), Reynolds diagrammatically illustrates this trade network concept at exhibit PAR-C as below. In this exhibit Reynolds also discusses that he was investigating a business which involved looking at data to create credit models to profile the risk of extending their downstream trade credit programs involving the sourcing of private money to fund the trade credit programs based on risk. The diagram is a high-level graphic of that idea to which Kemka was responding noting he did work on a trade credit program and had developed risk models for small business as well.
Following this, Factor suggest that the inventive concept within the specification correlates with several elements I discussed above in the specification. In particular, Factor points to the content identified by: Figs. 5A, 5B and their associated description relating to the calculation of action scores and probabilities based on these action scores; the idea of a Network Predictive Cashflow (NPC) as a specific action score; Fig. 5C being a system for implementing algorithms allegedly incorporating the Decise platform; and Fig. 32 and associated description relating to a Machine Learning Credit Score Generator (MLCSG).
Applicant’s identification of the inventive concept
Xero’s simplified position is that the inventive concept of the specification relates to content that was developed in a period roughly coincident with the third phase. To arrive at this position, Xero discusses the nature of the specification and the background to the invention. They submit that the key inventive concept of the patent application was that:
“(a) a score could be generated by a network platform dynamically at any point in time for an entity (entity A) based on its network data, that data relating to an aspect or aspects of the entity’s performance; and (b) that score could be used to predict the probability of timeliness of some future action of that entity with respect to another entity in the network (entity B). This probability could then be displayed to a user, eg an accountant for entity B”
Clearly this equates very closely with Figs. 5B and 30 and is further embodied in terms of the NPC action score. Xero argue that neither the platform for hosting the scoring algorithms (Fig. 5C) or the MLCSG argued by Factor to be represented by Fig. 32 and associated text in the description, form part of the inventive concept of the application. Xero’s argument is initially premised on the background knowledge to the invention and the content of the specification. There appears agreement to the timelines of discussion between Kemka and Reynolds at the “first phase” and “second phase” of development, noting that the discussions at the “second phase” manifested a Draft Presentation presented to Grigg at Xero (Kemka #1 at exhibit MK-3):
“What it [the Draft Presentation] shows is that Kemka and Reynolds had discussed the idea of using “network data” for the purposes of a business, that is, data concerning entities in the business’s network such as suppliers and customers. Kemka, being a user of the Xero accounting platform and being aware of the broad uptake of that platform, had the idea that Xero could provide the relevant network data. They also had discussed the idea that this data might be able to be used somehow to create a “credit score” for the business, that is, a score which the business could then provide to a lender for the purposes of seeking a loan. Thus, the document [Draft Presentation] stated “Xero network data has the potential to enhance credit approval algorithms supporting small business lending” and identified in very broad terms the notion that “the performance of a small business is also impacted by the on-going performance of their key suppliers and their buyers”.
Regarding this “Trade Network Concept” the applicant discusses that these insights were not unique. Grigg discusses in his evidence (Grigg at [12]) that he “[Passage not published by reason of confidentiality]”. Kemka also notes that the concept of commercialising an over-arching trade network risk analysis to which Reynolds refers to at [38] of Reynolds #4, referred to as a trade network, was one which he “had come across while working on trade data projects with various credit bureaus and banks” and that “I did work on the trade program when I was at D&B and have developed risk models for small business as well” (Kemka #2 at [12]).
Turning to the specification, Xero discusses the nature of the problems described and the solutions provided by the application. They point to the background discussion in the specification which I have summarised above where it is stated that numerical representation models for a business are known and they usually lack context (e.g. context which can be provided by data relating to the business’s network), further noting that these representations lack predictive value for actions taken by the individual for which the numerical representation was generated, especially with respect to the time of actions taken by the individual. As such, the specification points to a need in the art to improve generation of predictive models capable of identifying probabilities of actions and probabilities of actions occurring with respect to identified periods of time. Following this Xero submit that the gap identified by the specification is the use of an action score to predict probability identified essentially as (b) above at paragraph [66].

The alleged advance over the prior art is said to be further depicted in the specification as follows at [0045], [0046] and [0050]:
Small businesses typically only have the payment history of an account to forecast the probability of on-time payment from that account. If the business has limited or no sales history with a client, then it is difficult for the business to forecast when the bill would be paid. The business may check the credit score with a credit bureau, but credit reports are biased towards scoring for representing the risk of non-payment or default, rather than identifying the risk of late payment. For example, a company with an excellent credit report may still pay its creditors an average past thirty days overdue.
Embodiments allow users to obtain accurate forecasting for sales receipts based on the payment history of clients, where the payment history includes not only the payments made to the business, but also the payments made to other businesses. This is possible because an accounting platform, also referred to herein as an "action history system," may provide metrics (referred to herein as "action scores") regarding payment history for the customer.
…
Traditional credit scores only paint an overall picture, created over a long period of time, of the history of companies paying bills on time. However, the action scores paint a live picture, of real-time data, based on the expected cash flow and other financial parameters (e.g., receivables of a customer), as well as the level of risk associated with the cash flow.
Following this, Xero points to the discussion in the specification of the NPC score and refers to Fig. 5B, before also discussing the claimed invention with a focus on the calculation of probability and predictive cashflow/NPC. Conveniently, Kemka makes the following statement at [16] of Kemka #2.
“…the trade score algorithm described and claimed in the patent application is more specific than the generalised concept put forward by Paul. In particular, the trade score algorithm which embodies claim 1 incorporates a number of features. It assesses payment histories of small and medium business (SMBs) or other entities in a network, and allocates each SMB with a score. This score is based on certain performance metrics, such as percentage of times that invoices (e.g. accounts receivable invoices) are paid on time to the SMB. When a first SMB issues invoices to a group of SMBs in the network, the algorithm determines the entity or entities related to the invoices, and then determines the probability of each of those entities paying the invoice within a certain number of days based on the action scores of those entities. The information is then presented to the first SMB.”
In summary, it is clear that the inventive concept being posited by Xero is not as broad as that being defined by Factor, with Factor essentially appearing to suggest the inventive concept may be “credit scores for small business and accessing big data that matched invoice and bank transaction data in one location”. This was allegedly a highly generalised idea that they (Kemka and Grigg) had already come across, the specification allegedly bearing this out by its posing of the problem and discussion of the contribution made to the art.
For completeness I also note Xero’s submissions with respect to other elements identified by Factor as forming part of the inventive concept. In short, these elements can be referred to as the “Decise’ platform (Fig. 5C) and the MLCSG (Fig. 32).
The Decise platform is software that I have discussed earlier as being developed by Kemka prior to him engaging with Reynolds. It was designed to host different scoring algorithms. As above, it is alleged by Factor to be present in the specification by way of Fig. 5C. Similarly, reference is made by Factor to the MLCSG arguing that Fig. 32 and associated paragraphs [0376] to [0384] incorporate intellectual property of Factor. Xero points out that this allegation is made with reference to a patent application filed by Factor incorporating a specific “Decise” platform and specific associated machine learning system. On this Xero refers to Kemka #2 at [21] and [21]:
“Paul appears to believe that the patent application describes and claims the validation module of the Factor patent application (AU 2017374966). This is not true. The validation module of the Factor patent application compares a predicted outcome given by a generic model with the actual outcome and automatically and iteratively adjusts the generic model in order to improve its performance. In contrast, the validation module of the Xero patent application simply compares the predicted score against the actual score. It doesn’t make any adjustments to the model based on the comparison and is not concerned with continuous improvement. The validation module described in the patent application is a common method of comparing performance. As a data scientist, I believe it would have been hard to find any example of a predictive model of credit risk where the estimate isn’t compared to the actual outcome.
Claims 2, 4, 11, 13, 17 and 19 of the patent application relate to calculating the scores for the network entities based on their transactions. They do not describe the validation module of the Factor patent application. Paragraphs 115, 120, 121, 160, 161 and 163 and Figs. 5C and of the patent application describe comparing predicted performance for the entities with actual performance. They do not describe automatically and iteratively adjusting the model to improve performance, as performed by the validation module of the Factor patent application.”
To put it more succinctly, the content of the Factor application that is said to involve a MLCSG constitutes an algorithm deployed on the Decise system which itself may host a library of models. It is argued to involve the selection of a generic predictive model and the subsequent refinement via machine learning of the predictive model based on a comparison of the predicted outcomes for the customer with the actual outcomes for the customer. Xero argues that on the contrary, the application in question does not implement such a platform nor machine learning approach, instead using dynamic performance measures, and static models generated by well-known machine learning.
Consideration of Inventive Concept
Clearly the specification points out that the primary problem that the invention is presented to solve is to improve generation of predictive models capable of identifying probabilities of actions and probabilities of actions occurring with respect to identified periods of time. I consider it indubitable that the solution to this problem is best depicted in general form with reference to Figs. 5B and 30. Furthermore, the independent claims are precisely reflective of the subject matter in Figs. 5B and 30, with the dependant claims clearly convergent with this subject matter. I see there to be little contention arising from Factor’s submissions with this articulation forming relevant subject matter to the requested declaration under s36. In fact, Factor specifically identified Fig 5A and 5B in their submissions as is shown on the following page.
The specification does not appear to set out to identify there to be a contribution in the art in the idea of merely accessing transactions between entities on a database and using this data to create some kind of score indicative of performance. To this extent, the general idea that appears represented by Fig. 5A of tracking transactions and evaluating performance to generate a score appears reflective of a general state of the art of methods such as creating credit scores based on an entity’s history of paying off debts, these traditional methods are seemingly lacking, on the face of the specification, in predictive power. The present invention seeks to improve upon the standard scoring methodology by improving predictive models capable of identifying probabilities of actions occurring with respect to identified periods of time. In other words, the invention appears to constitute the use of particular scores in a more detailed algorithm to make probabilistic predictions.
Xero point out in their submissions that Factor misconceive the inventive concept, and, in line with my understanding, Xero suggests Factor conceives of an inventive concept in terms of overly broad and known subject matter. They note that Factor suggests that the inventive concept was “creating credit scores for small business and accessing big data that match invoice and bank transaction data in one location”. They further point out that this is not what the specification presents as the inventive concept. They also argue that the key concept of the specification is that network data could be used by a platform to create a score indicating the cash flow performance of an entity A [e.g. an NPC], and then that score could be used by entity B at any given time to determine whether entity A would pay an invoice of entity B’s with for example, 30 days. I agree with the applicant that the inventive concept of the specification involves this more detailed “improved” algorithm involving the first stage of creating a score on the basis of performance data, but importantly using this score in the prediction of timeliness probability.
It clearly follows that a relevant heart of the inventive concept is that which is defined by the applicant, and provided in this decision above at [66].
I now turn to the other elements of the specification identified by Factor as being subject matter relevant to my determination under s36 being the Decise platform and MLCSG. As noted above these elements are essentially represented by Figs. 5C and 32 (Fig. 32 provided earlier in the decision). Figure 5C is an extremely basic block diagram of a system. It clearly does not show any specific hardware arrangement. The associated description from [0115] to [0122] generally describes each of the elements of a client device 480. The device includes:
·A receiver component 510 that accesses communication data. This communication data may be financial transactions flowing from an account to another user or financial transactions flowing to an account from third parties. The data might be received automatically or manually
·A generation component 520 that generates performance values based on these financial transactions
·A scoring component 530 that generates an action score based on the performance values
·A prediction component 540 that generates predictions based on some model and these predictions may be a set of action probabilities that indicate that a user will perform respective actions within a predetermined time frame
·A validation component 550 that validates actions scores by verifying the set of predictions based on data indicative of prior transaction trends. As such, the validation element tests the quality of the predictions/scores.
·A presentation component 560 which causes presentation of an action score and action probability on a user interface.
Clearly, the elements described in this device are primarily focussed upon the mere system or device-based implementation of the method described above relating to the prediction of the probability of certain actions occurring within some timeframe. What is the only notable addition in Fig. 5C is the idea that there is a step where scores are validated by comparing the predictions of a model with predictions based on historical information regarding transactions. Reference is made to the idea that this may be done “statistically” but there is no specific detail developed beyond this general idea of validation by comparison with historical data. To this extent, the described content does not appear to be directed to any particular improvement of the predictive capability of a model as is posed to be the problem addressed by the specification, but instead merely appears to be playing the role of validation. I cannot see this content as contributing to the inventive concept in the specification, and instead see the content represented in Fig 5C as a mere implementation of the inventive concept defined in relation to Fig. 5B.
I do not see Fig. 5C being a platform of any specific arrangement for improving the generation of predictive models. The specification does not focus any detail upon the construction or function of this system aside from pointing out a rather generic step of data verification by comparison with historical results. Similarly, the claims place no focus upon any element of validation as being part of the invention sought for protection. The focus of the specification appears to be the algorithms for generating scores and predictions, the particular scores and predictions themselves, and particular business case uses of the scores and predictions. It follows that I do not see any aspect of Fig. 5C forming a heart of the inventive concept.
I think similar can be said of Fig. 32 as was said in relation to Fig. 5C. Most importantly, the associated discussion of Fig. 32 at paragraphs [0376] to [0384] appears to me as nothing more than a generic description of standard machine learning technology and approaches that may be applied to create algorithmic models for the calculation of action probabilities. In Fig. 32 data is fed into machine learning to create a trained program (algorithm) that is capable of generating an action probability. New events can occur that can be provided to the trained algorithm to generate new action probabilities, but, again, this appears to me to amount to nothing more than a routine manner of identifying, automatically, rather than by brute-force intellectual analysis, that there is a relationship between certain data that can be encoded into an algorithm to produce a prediction action probability. This appears to be the mere implementation of the general inventive concept and the content therein of Fig. 32 does not form a heart of the inventive concept.
It does appear important at this point to refer to the evidence, in particular that of Mr Kemka, as Factor seek to identify him as possessing a conflict of interest that would require me to give less weight to his statements. However, I note that the observations that I have made above regarding the inventive concept are quite consistent with the evidence provided by Mr Kemka. To this extent, both Mr Grigg and Mr Kemka make the following observations regarding the broad ideas of: creating credit scores for small business and accessing big data that match invoice and bank transaction data in one location; and the content of Figs. 5C and 32 as not relating to a dynamic machine learning approach to iteratively improving a model but merely a generic implementation of the inventive concept. The observations support a finding that the related content of the specification is not part of the inventive concept.
“During the initial introductory and exploratory discussions with Martin and Paul they expressed the view that [Passage not published by reason of confidentiality]” (Grigg at [12])
“In paragraphs 17 d) and 24 of Paul’s third declaration, and paragraph 38 of Paul’s fourth declaration, Paul states that while working on the World Bank project he identified the challenge for creating and the opportunity for commercialising an over-arching trade network risk analysis which he refers to as the “Trade Network”. This was not a new concept to me at the time. The idea of a trade network was one I had come across while working on trade data projects with various credit bureaus and banks. This was what I was referring to when I responded to Pauls email of 18 March 2014 by saying ‘I did work on the trade program when I was at D&B and have developed risk models for small business as well”, as shown in Annexure Par-C.” (Kemka #2 at [12])
“…the validation module of the Xero patent application simply compares the predicted score against the actual score. It doesn’t make any adjustments to the model based on the comparison and is not concerned with continuous improvement. The validation module described in the patent application is a common method of comparing performance. As a data scientist, I believe it would have been hard to find any example of a predictive model of credit risk where the estimate isn’t compared to the actual outcome.” (Kemka #2 at [20])
Further, I see no basis, at this stage of consideration, to believe the evidence of Mr Kemka to be of little probative value. In fact, his evidence is supportive of my construction of the inventive concept formulated on the basis of the specification. Furthermore, the assessment of the nature of the inventive concept defined by the specification is entirely unaffected by the additional evidence filed by Factor in the form of Reynolds #5. Instead, this further evidence focuses primarily upon circumstantial events leading into the hearing, alleged fiduciary duties, and various contractual arrangements. Thus, it is clear at this stage of consideration that Mr Kemka’s evidence is of substantial probative value, and that the Reynolds #5 declaration is not relevant as evidence to the question of “What is the inventive concept?”.
Inventive Concept - Conclusion
In accordance with Xero’s submissions, I consider the relevant inventive concept for the purpose of the requested declaration under s36 can be presented as follows:
“(a) a score could be generated by a network platform dynamically at any point in time for an entity (entity A) based on its network data, that data relating to an aspect or aspects of the entity’s performance; and (b) that score could be used to predict the probability of timeliness of some future action of that entity with respect to another entity in the network (entity B). This probability could then be displayed to a user, eg an accountant for entity B.”
As an aside, I do note that this concept is particularly nebulous and does not appear described with much detail in the specification, and as such may present issues for Xero during examination in consideration of issues related to enablement. Similar challenges may be faced as regarding the subject matter itself as it does prima facie present as an abstract calculation and prediction.
INVENTORSHIP
The position put forward by Xero is rather simply summarised as being:
“…the invention was conceived in the course of a feasibility study (the Feasibility Study) conducted by Kemka with the assistance of Grigg in February to April 2015, and the report of that study (the Feasibility Report) sets out the inventive concept in a manner consistent with the patent specification as subsequently filed… Reynolds and Wallace were not involved in the conception of the invention...”
This period in which the Feasibility Study was performed corresponds to the initial part of a third phase identified by Factor. A summary of their position is as follows:
“a. The inventive concept of the invention of the Application was initially conceived during the period March to July 2014 (first phase) and then expanded upon the inventive concept (sic) in more detail from August 2014 to January 2015 (second phase). The inventive concept was further expanded upon from February 2015 to August 2016 (third phase) and then was developed to a working prototype from September 2016 to July 2017 (fourth phase).
b. Reynolds and Martin Kemka (Kemka) initially conceived the invention and became partners of a partnership called Decise during the first phase. The partners of Decise then dealt with Alistair Grigg (Grigg) of the nominated person during the second phase. The partners of Decise by way of an agent and individually, including Wallace entered a formal arrangement with the nominated person during the third phase. Wallace joined Decise during the third phase. The Requestor was established during the third phase by the partners of Decise with the agent consulting for them and then entered into a formal arrangement with the nominated person to undertake the development of the invention on behalf of the Decise partners during the fourth phase.”

From the narration of events I have already provided earlier in this decision, it appears to not be in dispute that certain general stages of progress correspond with the four “phases” identified by Factor. In this “first phase”, Kemka and Reynolds were having preliminary discussions and beginning to formulate general ideas for a business that would initially be considered to operate under the name of Decise. In outlining the common narrative earlier in this decision, I referred to the ideas expressed by Reynolds regarding trade credit programs utilising big data, and that Kemka and Reynolds agreed to seek to form a business together which would solve problems and provide solutions for trade credit. They would seek to operate using data gathered by accounting software and deploy Kemka’s “Decise platform” to implement the solutions. They approached Xero as a company holding significant volumes of useful data to seek to develop useful solutions, and this contact with Xero was facilitated with the help of Wallace. Factor submit that the initial inventive concept was conceived during this first phase and then outlined to Mr Grigg of Xero at an initial set of meetings during the second phase.
The “second phase” of work corresponds to the period in which a proposal embodied in a presentation (Draft Presentation) is presented to Xero and discussed between key individuals including Reynolds, Kemka and Grigg. Ultimately these discussions lead to the preparation of a draft proposal for a Feasibility Study for the creation of models around January of 2015. A contract for services was arranged between Xero and Kemka’s company Northraine Pty Ltd for the completion of the proposed Feasibility Study (which begins the third phase). At this point it is important to discuss the progress of the invention and the content of the Draft Presentation and related discussions with Xero.
A primary submission of Factor’s with regard to the Draft Presentation is that in early August of 2014:
“[Passage not published by reason of confidentiality]”
A copy of the Whiteboard screenshot is provided at Reynolds #3 in exhibit PAR-G representing the discussions between Reynolds and Kemka prior to the formulation of the Draft Presentation. I have not included the screenshot in this decision due to its low image quality and poor reproducibility. Nonetheless, as identified by Factor, it provides a general overview of relevant elements forming the task of using Xero’s network data [Passage not published by reason of confidentiality].
It is clear that this information does not provide a depiction showing the formulation of an idea for a particular scoring algorithm that is the relevant inventive concept for the purposes of this decision. What it shows is that there was a plan to look to business characteristics in order to look for patterns that could potentially be translated into useful scores for credit models. As noted, this general plan was translated into the Draft Presentation that represents the discussions between the key parties prior to the performance of the Feasibility Study. There doesn’t appear to be any knowledge generated at this point in relation to the potential for scoring algorithms using networked data of business that was not already known to those in the art.
In this regard, in response Xero argues in relation to this earlier work and the Draft Presentation that:
·The idea of using network data for the purposes of a business was discussed between Kemka and Reynolds
·Xero was realised to be a potential provider of the relevant network data
·Kemka and Reynolds had discussed that it might be possible to somehow create a “credit score” for a business which could be provided to a lender for the purposes of loan provisioning.
·The Draft Presentation stated that “Xero network data has the potential to enhance credit approval algorithms supporting small business lending” and identified in very broad terms the notion that “the performance of a small business is also impacted by the on-going performance of their key suppliers and their buyers” (Kemka #1 exhibit MK-3).
·These insights were not unique or new (referring to evidence of Grigg at [12] and Kemka #2 at [12]).
Particularly apposite at this point in my consideration is the law in regard to the conception of an invention. It appears accepted as Australian law that conception as the touchstone of inventorship can be equated to the completion of the mental part of invention. It is the “formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention as it is hereafter to be applied in practice”. It is complete only when the idea is so clearly defined in the inventor’s mind that only ordinary skill would be necessary to reduce the invention to practice without extensive research or experimentation.
I have reviewed the Draft Presentation which is provided at exhibit MK-3 of Kemka #1 and agree with Xero that the ideas discussed with Xero were preliminary in nature and required an analysis of Xero data to provide insights into networks and potential use for new credit algorithms. The use of network data as a potential for credit algorithms does not appear to be a new idea. At this stage of the process (second phase), what appears to me to be the state of progress is the conceptualisation of a business idea and structure that would use Kemka’s Decise platform to create new credit models on the basis of Xero’s network data by exploring relationships and patterns in the data. Thus, to this point, no operative idea of how a credit model would function had been developed. While at this stage of development a valuable proposition for a business model appears to have been put forward in the context of a hosted decisioning engine running unique credit models with the use of network-based data, what remained was to determine how this was to be done algorithmically. On this basis, it is clear to me that at the end of the second phase, conception of the inventive concept of the specification was yet to occur. To this extent, the second phase of progress ended with the preparation of a draft proposal for a feasibility study by Kemka. Here Xero submit:
“In November 2014, Kemka prepared a draft proposal for a feasibility study in relation to creating a credit risk score model and credit health score model. This proposal was presented to a number of Xero executives, including Grigg, in January of 2015. It did not advance the concept beyond what was stated in the Draft Presentation, that is, it spoke in general terms of the use of network data to generate a credit score which could be provided to banks who could use it for lending purposes, and underlined the need to investigate Xero’s data as a next step in order to consider whether it might be possible to create models and what they might be.”
This draft proposal is provided with Kemka #1 at MK-4. Having reviewed this document I concur that the proposal was couched in general terms with a desire to identify useful data held by Xero and incorporate this into a particular algorithm to create certain “credit scores”. It appears clear that the inventive concept had not been formulated by the end of the second stage. In other words, there was no clear definition in the mind of an inventor as to a complete and operative invention. Instead, what appears to have been achieved is the clear identification of a problem which if solved, would provide for potentially profitable business products. More specifically, the draft proposal refers to the need “[Passage not published by reason of confidentiality]”. The aim was “to create algorithms” following “actions taking in similar investigative projects in trade credit and consumer risk areas”. This is consistent with the drafting of the specification which seeks to improve generation of predictive models for networked entities, capable of identifying probabilities of actions occurring with respect to identified periods of time.
The “third phase” includes the time in which Kemka conducted the Feasibility Study with the assistance of Grigg. As submitted by Xero and shown in evidence the Feasibility Study was conducted at Xero’s premises by Kemka and Grigg using Xero facilities and a sample set of New Zealand Xero data (Grigg exhibit AG-5). Regarding this work Xero submit:
“Kemka prepared the Feasibility Report on the Feasibility Study in April 2015, which was presented to Xero. For the first time, the Feasibility Report set out the key concept of the invention. They key concept was that (a) a score could be generated by a network platform dynamically at any point in time for an entity (entity A) based on its network data, that data relating to some aspect of the entity’s performance; and (b) that score could be used to predict the probability of timeliness of some future action of that entity with respect to another entity in the network (entity B)…”

100. Turning to the Feasibility Report at exhibit MK-9 of Kemka #1 it is stated under the heading “Proposed Algorithms” that provided in the report is “A predictive algorithm [that] takes specific information that is associated with a specific entity at a specific point in time and generates a single-number summary (a score)”. Xero points out that this corresponds to element (a) of the inventive concept with the key aspect of invention being that this element (a) can be used to “predict some specific future behaviour associated with the entity”. Three proposed algorithms are presented in the Feasibility Report and a first and most relevant of these, is called “Trade”, which predicts the chances that an invoice will be paid back in 30 days, and this can be an invoice associated with a member of the network such as a sibling, cousin or stranger (MK-9 of Kemka #1). This is the first recorded instance I can see in the evidence of the manifestation of the prediction of timeliness in the form of probability of an entity’s action with respect to another entity of the network, on the basis of use of some score derived from past performance.
101. The second part of the feasibility study is entitled “Network” and discusses "Network Level Trade Scores” that incorporate characteristic interactions between certain related businesses. The study notes that there are characteristics such as average trade scores of related and unrelated businesses over periods of 3/6/9/12 months, and average high-risk receivables to low-risk receivables of related and unrelated businesses that can be incorporated into current trade scores. These scores “are used as a distilled summary of the behaviour of different connections within the network” and “their fluctuations in the network can act as indicators of a sibling’s behaviour in the future”. Here it appears that quantified risk associated with receivables, and trade scores of other businesses is being used as a predictor for the timely payment of an invoice between entities. In other words, as summarised by Xero, “the cash flow performance of an entity in a network could be used to predict probability of timely payment of an invoice by that entity to another entity in the network”. The report ends with an ‘Overview” which Xero fairly summarise as follows:
“[Passage not published by reason of confidentiality]”

102. What is clear to me is that the preliminary ideas posed and discussed in the first two phases did not amount to anything more than element (a) of the inventive concept alone, which was essentially the provision of a “credit score” on the basis of networked business trade data. The purpose of the work in the Feasibility Study appears to be directed towards a search for more useful scoring algorithms. Importantly in this regard, the inventive concept is the realisation that a score can be generated related to an entity’s past performance and that this score can be used to predict the timeliness probability of a future action of the entity with respect to another entity in the network. While the Feasibility Report does not appear to contain precise clear detail regarding mathematics of score and probability calculations it does appear to me to clearly represent the first studied realisation that predictive algorithms could be generated that used a score based on past behaviour of an entity in a network to predict the probability of a future action of that entity with respect to another entity in the network. It would appear that reduction of the inventive concept to practice was further work to be conducted.
103. For completeness, the “fourth phase” of the invention relates to the development of a working prototype from September 2016 to July 2017 under the operation of a professional services agreement (PSA) dated 1 September 2016 between Factor and Xero having associated statements of works (SoWs). A range of work was delivered under these statements of works. At this point of my decision, I need not discuss this further as it seems to relate to the reduction to practice of ideas of the feasibility study using a larger set of Australian data (Kemka #1 at [71]) and furthermore, is work conducted after the priority date of the patent application in question.

104. It follows that I am sufficiently satisfied that the completion of the conceptualisation of the relevant inventive concept occurred during the Feasibility Study, and its reduction to practice therefrom would not appear to require extensive research or experimentation. The inventive concept had not been developed before the Feasibility study. I further refer to the evidence of Grigg at [26] where the following technical aspects are summarised by Xero to have been developed in discussions between Grigg and Kemka:
“[Passage not published by reason of confidentiality]”

105. In summary, I do not consider any aspect of the inventive concept as having been developed during initial discussion between Kemka and Reynolds, or in the second stage when Kemka and Reynolds initiated discussions with Xero. Relevant concepts were already known, such as the idea to score the performance of an entity in a network for the purposes of credit provision. The initial discussion leading into the Feasibility Study appears to provide foundation and motivation for further work and in particular the desired identification of valuable potential products in the form of improved credit scoring algorithms. This involves data analysis to identify relationships and trends in data of an interconnected network of business. It is the search for improved credit scoring algorithms that appears to come to fruition in the Feasibility Study.
106. Having determined “when” the inventive concept was conceived, I now must turn to consider “who” was responsible for this inventive concept.
107. It is not disputed that Kemka and Grigg are responsible for the inventive concept. Thus, it remains for me to consider whether Reynolds or Wallace were also responsible for the inventive concept that was arrived at during the Feasibility Study.

108. Firstly, regarding the work of Wallace, Factor submit that his contributions were made in or around June of 2015 and from November of 2015 to April of 2016. Notably this is after the Feasibility Study. The content identified by Factor as being the result of a contribution of Wallace can be summarised as follows:
a)   Mr Wallace highlighted issues with a customer health of a business can be identified through data on employee payroll and superannuation liabilities, payments to directors, tax liabilities; and posits solutions using such.
b)   Mr Wallace’s contributions are reproduced in the specification at [0107] and [0136]- [0138].
c)   In exhibit AW-7 of Wallace #2, Mr Wallace provided insights to Mr Kemka on how to interpret and characterise certain data, this being reproduced at [0125] – [0130] of the specification.

109. The relevant paragraphs of the specification cited appear nothing more than known information indicative of transactions and communications occurring between entities, known indicators of business health, and known key performance indicators for businesses. While most significantly this contribution is made after the Feasibility Study, it appears the case that the advice provided by Wallace could not form part of the inventive concept but instead represents more general accounting advice representing information that would have been known to an accountant as common general knowledge. The advice provided by Wallace would have been of assistance in Mr Kemka’s further development and implementation of the invention, but I cannot see how the information provided by Wallace could have constituted any form of inventive contribution, regardless of its timing. Clearly, Wallace cannot be considered an inventor of the relevant inventive concept.
110. Regarding the contribution of Reynolds, I have already determined that the inventive concept was conceived in the Feasibility Study, and that before such time, the idea to create “credit scores” on the basis of a network of trading entities was not a new concept. As submitted by Xero:
“The trade score algorithm as embodied with the Patent Application was only conceivable following the Feasibility Study which assessed available Xero data, how entities were connected in the Xero network, where there was a mechanism for predicting the payment of an invoice and whether Xero data could be used to make a prediction of payment and define the scores.”

111. Reynolds concedes that Kemka was on site during the Study writing code and creating a model (Reynolds #3 at [44(a)-(b)]; Kemka #2 at [22]). It is clear that Reynolds did not make any contribution to the analysis of Xero data sources which was essential to the conception and development of the relevant inventive concept (Kemka #2 at [22]). Kemka did keep Reynolds informed of the progress of work in the Feasibility Study, but due to his lack of experience in any technical data science work, Reynolds focussed in that time upon creating a Business User Assessment Report (Reynolds #3 at [44(a)-(b)]). In short, there is no evidence demonstrating involvement of Reynolds in the invention conceptualisation performed at Xero during the Feasibility Study. As such, I cannot be satisfied that Reynolds can be considered an inventor of the relevant inventive concept of the specification.
112. Thus I consider only Kemka and Grigg as inventors for the purposes of the relevant inventive concept. While there is no doubt that Reynolds was a key participant in identifying business value in the creation of new and improved credit scores, and that he provided clear drive and business support to seek improved solutions, I do not see him as contributing as an inventor to the particular solution provided by this specification.

113. I also again refer to the allegation of Factor as to the evidence of Mr Kemka being of little probative value. I note at this stage it remains that I see no facts of contention that suggest this to be so. The determination of inventorship flows essentially from my identification of the inventive concept and from the undisputed circumstances experienced by the parties. Also, my identification of the relevant inventors is entirely unaffected by the additional evidence filed by Factor in the form of Reynolds #5. As noted above, this further evidence focuses primarily upon circumstantial events leading into the hearing, alleged fiduciary duties, and various contractual arrangements.
CONTRACTUAL OR FIDUCIARY RELATIONSHIPS
114. It is clear that there is no question in this matter regarding the assignment of rights in the inventive concept with respect to the work conducted by Grigg. Entitlement clearly flows with Grigg being an employee of Xero. Thus, I only need to consider the contractual or fiduciary relationship in place in respect of the work done by Kemka during the Feasibility Study. [Passage not published by reason of confidentiality].

115. [Passage not published by reason of confidentiality]. Thus, there appears to me to be no doubt that intellectual property created during the Feasibility Study was to be assigned to Xero.
116. The arguments presented by Factor focus upon the formulation of various concepts of partnership and fiduciary duties that are said to arise in the early stages of the forming of a relationship between Kemka and Reynolds and subsequently the formation of Factor. These submissions do not, in any substantial way, focus upon the determination above that the inventive concept was conceived only during the Feasibility study and only by Kemka and Grigg. The agreements mentioned above appear to clearly show the assignment of intellectual property of the Feasibility Study from the participants of that study to Xero. Essentially, the submissions put forward by Factor rely on an imprecise identification of the inventive concept, with the result being: a) the inclusion of irrelevant material in the inventive concept (i.e., the “Decise Engine” and the “Machine Learning Credit Score Generator”; b) the identification of an incorrect time for conception of the invention (i.e., being before and/or after the Feasibility Study); and c) the identification of relationships and arrangements that were irrelevant to the particular inventive concept (i.e., event, agreements, and business arrangements that happened after the inventive concept was conceived).
117. At this point I note that the flow of entitlement ultimately facilitated by the NDS and the CFS is unaffected by the additional evidence filed by Factor in the form of Reynolds #5. While this further evidence focuses upon alleged fiduciary duties and various contractual arrangements, none of the information disposes of the agreements and the clear agreed terms therein with respect to the Feasibility Study. Furthermore, this evidence does not bring into question any of the evidence provided by Kemka that I have considered in this decision, to the degree that I would consider it of less probative value. I have arrived at my conclusion via what appears to be commonly agreed facts regarding the circumstantial history of the present matter and a construction of the specification following accepted legal principles. Thus, there is no basis to further consider Reynolds #5 for the purpose of this decision and I herein refuse the request to do so in accordance with Regulation 5.23.
CONCLUSION
118. I conclude that the inventors of the inventive concept are Kemka and Grigg and the invention was conceived during the period February to April 2015. The applicant is the sole eligible person by reason of a Contract for Services between the Applicant and Northraine Pty Ltd and a Confidentiality Agreement between the Applicant and Martin Kemka, each executed by him on or about 11 February 2015, and by means of employment by the applicant of Alistair Grigg.
119. It follows that I must refuse the application for a declaration under s36(1). Noting there were no specific discussions at the hearing regarding s113(1), I also consider it expedient to refuse the requested direction under s113(1) as it agitates the same issues. In this regard, there is no contribution I can see from a relevant person that would facilitate another party as being a joint applicant. None of Reynolds, Wallace or Factor are eligible persons and as such, cannot be joint applicants.

120. Furthermore, given the prevalence of potentially confidential material in this matter, I provide the parties two (2) months from the date of this decision to provide detailed reasoning as to the relevant parts of this decision and the identified list of documents provided prior to the hearing, that should not be open to public inspection.
COSTS
121. Both parties sought their costs and in fact both parties sough a variation above the scale of costs stipulated in Schedule 8 of the Patent Regulations. With Factor being unsuccessful and given it normal for costs to follow the event, as a starting point, I herein award costs against Factor Financial Analytics. To address the issue of a variation to Schedule 8, I invited submissions from the parties following the hearing. Both Xero and Factor generally argue that the other party’s conduct has resulted in unnecessary delay leading to inefficient and complex proceedings. To this extent, I summarise the relevant submissions of Factor as follows:
·When both parties agreed to an initial stay of proceedings and committed to an attempt to reach amicable agreement, despite many attempts by Factor, Xero allegedly did not engage with these efforts
·The applicant allegedly has used the failure of a stay as a tactic to waste time and delay
·Factor alleges Xero has used their comparative size to its upmost advantage along with the threat of breaches of confidentiality

122. Similarly, regarding the general conduct of Factor, I summarise relevant submission of Xero as follows:
·The proceeding has long been pending since August 2017, and due to Factor’s conduct the evidence phase took over two years to complete. Xero point to the voluminous correspondence over the period and argue many of these correspondences were irrelevant.
·Xero point to the reframing of Factor’s case on three separate occasions: first with the statement of grounds and particulars; second by direction of the Deputy Commissioner; and third, shortly prior to the hearing.
·Factor refiled evidence shortly before the hearing that was earlier refused for consideration under Reg 5.23 by the Deputy Commissioner, this evidence was accompanied by allegedly serious, unfounded, and irrelevant allegations against FB Rice, the attorneys for Xero.

123. Specific conduct of Factor in relation to the setting and progress of this hearing itself also warrants consideration. Xero make submissions on these events as follows from points c) to k) appearing in their submissions of 28 May 2021:
c)At the interim hearing in December 2020, Dr Barker requested that the parties provide IP Australia with its availability for a two-day hearing in the first half of the year at the parties’ “earliest convenience”.
d)Immediately following the interim hearing, Xero obtained available dates for senior counsel and its advisors and requested confirmation of Factor’s availability.
e)On 24 December 2020, Factor provided its counsel’s availability. Thus, Factor, the moving party, sought an oral hearing, to be conducted by senior counsel. The hearing was subsequently set down for 25-26 May 2021.
f)Notwithstanding having almost 10 weeks’ notice, Factor only filed its complete submissions on 17 May 2021 nearly a week overdue.
g)In accordance with the request by IP Australia, Xero made itself available for a test call to test the technology on 21 May 2021. By way of contrast, Factor failed to provide IP Australia with any indication of its availability or any update on whether it would be appearing at the final hearing.
h)In anticipation of the hearing, Xero’s advisors, including counsel and a representative from Xero’s legal team set aside three days in their schedules, travelled to convene together for a day of preparation and the conduct of the hearing.
i)In the days before the hearing, Xero’s representatives made more than one request as to whether Factor would be represented and appear at the final hearing. It was not until 1:33pm on the eve of the two-day hearing that Factor indicated it would not be appearing at the hearing at all.
j)Nevertheless, Xero had to proceed and present its case. It did so with a degree of speculation due to Factor’s continuously changing position and vague written submissions. Both Xero and the Hearing Officer were left to speculate as to Factor’s position.

k)Factor’s engagement in the hearing consisted of a 30-minute appearance on the platform by an observer named “Paul”.

124. While the Commissioner will not normally award costs divergent from a Schedule 8 basis, a varied scale of costs will be appropriate “where the justice of the case so requires or where there may be some special or unusual feature in the case” (Re Wilcox; Ex parte Venture Industries Pty Ltd and Others, 141 ALR 727). Complexity or importance of the case may be a relevant circumstance, or the conduct of a party may be relevant, where one is guided by a basic question of “whether the justice of the case requires variation to the scale” (E.I. Du Pont De Nemours and Company v ICI Chemicals & Polymers Limited [2007] APO 12).
125. Relevant to the present matter, a variation to the scale of Schedule 8 may be appropriate where the conduct of one (or both) of the parties suggests that variation of the scale is appropriate. In this regard it is necessary to consider the degree to which a party has engaged in for example, misconduct such as unnecessary protraction in proceedings which may disentitle them “to the beneficial exercise of the discretion” (Oshlack v Richmond River Council [1998] HCA 11; 193 CLR 72 at [69]). Similarly misconduct that causes lost time and wasted money, or abuse of process would be considered entirely unreasonable (Rouse v Shepherd (No 2) (1994) 35 NSWLR 277 at 279-280) and lead to a variation of costs on an indemnity basis. Importantly conduct need not be misconduct per se, but conduct that may be seen to have brought about or substantially contributed to the costs being considered (Keddie v Foxall (1955) ALR 835, the note in 30 ALJ 27, Weir Pumps v CML Pumps 2 IPR 129 at page 154). As noted in Australia Patent Office decision Sterling Drug Inc v Boots Company PLC (1996) 35 IPR, “the scale of costs can be varied to allow the recovery of costs reasonably and properly incurred, if the scale provides inadequate recompense”.
126. In such circumstances it will be necessary to search for a "fair estimate" (H Lundbeck A/S v Alphapharm Pty Ltd (No 2) [2009] FCAFC 118 at [12] and [14]) of the variation to the scale of an award. Such an assessment will necessarily be based on the particular facts as a whole, recognising that "the assessment will often be rough and ready" having the virtues of “fairness and reasonableness” (Bowen Investments Pty Ltd v Tabcorp Holdings Ltd (No 2) [2008] FCAFC 107 at [5]).
127. Several decisions of the Commissioner have awarded costs that have varied from the Schedule to reflect conduct of the parties. For example, in MelbourneWater Corporation v Green Edge [2009] APO 26 costs were varied beyond the Schedule because the applicant, shortly before the hearing, sought to postpone the hearing due to their patent attorney going on leave. In Amgen, Inc v Genentech, Inc [2016] APO 15 (Amgen) costs were varied above the Schedule where the applicant engaged in conduct in relation to a request to withdraw an application under opposition. In Amgen, costs were awarded above the scale to the degree of 50% of actual costs occurred during a relevant period where the opponent was preparing their case for hearing. In Super Internet Site System Pty v Sensis Pty Ltd [2006] APO 27 (Sensis) the Deputy Commissioner varied costs to the degree of $3000 being a reasonable amount of indemnity costs considering the applicants conduct in relation to an opposed application that was withdrawn. Finally, in Simco Mining Products and Services Pty Ltd v CQMX Ltd [2019] APO 5 (Simco) the delegate made an award above the Schedule for item 11 – “Preparation of case for hearing” to the degree of five times that listed in the Schedule to compensate for the circumstances of a failure to define boundaries of the case and a large amount of evidence that was not relied upon. The delegate did not see any sufficiently compelling reasons to vary any other amount in the scale in Simco.
128. In the present case it is clear on inspection of the case file that the matter has been long, drawn out, complicated, and filled with large volumes of correspondence. Some of this correspondence has been led by the complications of the case sought to be put forward by Factor, and some of via issues imprecisely defined and associated with confidentiality led primarily by Xero. It is reasonable to suggest that an entitlement dispute, wherein the case to be determined is largely dictated by circumstantial facts, can easily become complicated, and it can be difficult to identify the key issues at hand. In this light, it would also appear that it is easy for a case to turn into a rather litigious affair with each party having problems or complaints about the other’s conduct. To this extent, my analysis of the general conduct of the parties does not raise major questions of an abuse of process, or anything of such a serious nature. While it is alleged that Xero have not engaged as fully as they could have with certain elements of the process, may have used their size to comparative advantage, and may have introduced complexity via confidentiality issues; Factor’s large amount of evidence and correspondence, and two separate attempts to introduce certain further evidence has also contributed to the significant pendency of these proceedings.
129. However, what stands out to me in particular is the general vagueness of the case put forward by Factor in the presence of this voluminous documentation. As I have noted above, Xero point to the reframing of Factor’s case on three separate occasions: first with the statement of grounds and particulars; second by direction of the Deputy Commissioner; and third, shortly prior to the hearing. The second of these instances involving the direction of the Deputy Commissioner was issued in the interim decision upon request by Xero. This appeared to be to aid Xero, and me as the Hearing Officer in deciding the case, due to the complexity and volume of the information already on file. This direction was complied with by Factor on 28 January 2021. Following this and shortly before the hearing, on 10 May 2021, Factor filed a request to amend both the application for a declaration under s36(1) and the request for a direction under s113(1) thereby reframing their case a second time.

130. This late reframing of the case, coupled with the ongoing complexity of issues led by the relevant documentation filed by Factor, has in my opinion clearly led to significant challenges over and above the normal challenges. These challenges were posed for Xero, the delegates involved in the progress of the casefile, and me in understanding issues and preparing for the hearing. To this extent, I consider the present case strikes a similarly with the issues of Simco where the delegate saw it appropriate to award costs beyond the schedule against the relevant party to the tune of five times the value for item 11 – “Preparation of case for hearing”. I consider it reasonable to apply the same variation to the present case as the Schedule provides inadequate recompense.

131. Generally, Xero argue that considering Schedule 8 being inadequate recompense for “AUD six figures” of expenditure, 50% of total actual costs incurred could be awarded. Briefly, I do not see any other elements of Schedule 8 that are applicable to an earlier point in time than preparation of the hearing itself because there are no sufficiently compelling reasons put forward by Xero in respect of any specific costs in that Schedule. Absent any numerical account of expenditure or further detailed description of the nature of expenditure resultant of particular actions of Factor, I see no reason to take such an approach to the variation.

132. Of note however are further items of Schedule 8 that refer to the costs of appearing at the hearing. Items 12 – 14 are respectively: “Attendance at hearing by registered patent attorney or solicitor without counsel”; “Attendance at hearing by registered patent attorney or solicitor instructing counsel”; and “Counsel fees for attendance at a hearing”. Xero was represented by counsel instructed by a registered patent attorney or a solicitor. As such items 13 and 14 are applicable to the present case.
133. Turning to the conduct of Factor leading into the hearing identified above at points c) to k), I consider the conduct is clearly far from ideal. Factor clearly progressed with the understanding of the necessity of a two-day hearing with counsel being involved and failed to engage with relevant timelines effectively and efficiently, leading to significant inconvenience. Without Factor’s attendance, the hearing was extremely abbreviated (only approximately 2 hours in length) with the arrangements made by Xero’s representation in view of the uncertainty of the hearing progress being inconvenient and excessive in hindsight. Factor only advised of their non-attendance on the eve of the hearing with their participation constituting only a brief non-participatory observation of the videoconference by “Paul”. It is clear to me that the inconvenience generated by this conduct, including the complete filing of their written summary of submissions almost one week after the regulatory deadline, is also unreasonable conduct that warrants a variation to the scale. In the present case, considering the arrangements for the hearing to be for a period of two days, I consider a fair variation to the scale of Schedule 8 to be the awarding of a full costs under items 13 and 14 covering a complete two days of the hearing. As such, costs in accordance with Schedule 8 regarding these items would constitute $1800 and $2700, this being in addition to the cost of $5000 reflecting a variation to item 11 (5 x $1000). I see no basis to apply any variation to the Expenses provided for in Part 2 of the Schedule which reflects transport and accommodation allowances as firstly it is not clear to me to what extent such costs were incurred, and secondly, I consider the actions of Factor are fairly dealt with via the variation to the scale of items 13 and 14.

134. It is worth making clear that there I do not consider there to be evidence of “intentional” or “calculated” misconduct by the Factor. However, I do consider that Factor has fallen quite short of demonstrating appropriate and timely action in preparing a reasonably clearly defined case for Xero to answer and fallen similarly short of satisfactorily engaging with relevant deadlines and steps for the conduct of the hearing. The result of this finding is a “rough and ready” proportion of the indemnity of costs being awarded against Factor. Such a proportion of indemnity is appropriate given Factor taking a course of action which leads to wasted time, effort and expense in opposition proceedings. Efficient conduct of opposition proceedings is clearly of interest to the parties, the Commissioner and the general public.

135. Thus, in summary, I award costs according to Schedule 8 against Factor. I apply a variation to the scale of costs against Factor to the extent of:
· Five times item 11 of Schedule 8 Part 1 ($5000)
· Full cost for two days of attendance for item 13 ($1800) and item 14 ($2700) of Schedule 8 Part 1.

Dr N. R. Madsen
Deputy Commissioner of Patents