Australian Competition and Consumer Commission v Google LLC
[2020] FCA 1563
•14 October 2020
Federal Court of Australia
Australian Competition and Consumer Commission v Google LLC [2020] FCA 1563
File number: NSD 1760 of 2019 Judgment of: THAWLEY J Date of judgment: 14 October 2020 Catchwords: PRACTICE AND PROCEDURE – application for discovery of 58 “categories” – whether categories were directly relevant within the meaning of r 20.14 of the Federal Court Rules 2011 (Cth) – consideration of delay in bringing application – discovery of some categories ordered Legislation: Federal Court Rules 2011 (Cth) rr 20.11, 20.13, 20.14, 20.15 Cases cited: Australian Competition and Consumer Commission v Hillside (Australia New Media) Pty Ltd t/as Bet365 [2015] FCA 1007
Baxter v British Airways PLC (1988) 82 ALR 298
Bridge Stockbrokers Ltd v Bridges (1984) 4 FCR 460
Dennis v Chambers Investment Planners Pty Ltd (2012) 201 FCR 321
Rhone-Poulenc Agrochimie SA v UIM Chemical Services Pty Ltd (1985) 10 FCR 567
Telstra Corp Ltd v Phone Directories Co Pty Ltd (2014) 316 ALR 590
Unilever Australia Ltd v Goodman Fielder Consumer Foods Pty Ltd [2009] FCA 1305
Division: General Division Registry: New South Wales National Practice Area: Commercial and Corporations Sub-area: Economic Regulator, Competition and Access Number of paragraphs: 64 Date of hearing: 14 October 2020 Counsel for the Applicant: K Richardson SC with A d’Arville Solicitor for the Applicant: Norton Rose Fulbright Australia Counsel for the Respondents: R Yezerski with S Hartford Davis Solicitor for the Respondents: Corrs Chambers Westgarth ORDERS
NSD 1760 of 2019 BETWEEN: AUSTRALIAN COMPETITION AND CONSUMER COMMISSION
Applicant
AND: GOOGLE LLC
First Respondent
GOOGLE AUSTRALIA PTY LTD
Second Respondent
order made by:
THAWLEY J
DATE OF ORDER:
15 October 2020
THE COURT ORDERS THAT:
In these orders:
(a)Discovery Application means the applicant’s interlocutory application dated 11 September 2020; and
(b)Category or Categories means the numbered categories set out in the annexure to the Discovery Application.
Discovery and Production
1.The respondents are to provide discovery of Categories 1 to 7 on or before 22 October 2020.
2.The respondents are to provide to the applicant copies of the documents referred to in Categories 9 to 24, 28, 29, 31, 32, 34 to 41 and 49 in their unsealed form (but which may contain existing redactions for privilege under Arizona and/or other United States law) by 20 October 2020.
3.The respondents are to provide discovery of Categories 9 to 24, 28, 29, 31, 32, 34 to 41 and 49 by 23 October 2020.
4.On or before 23 October 2020, the respondents are to provide to the applicant a document (or documents) identifying the name and job title (at the date of the communication or interview) of the persons:
(a)involved in the communications referred to in Categories 9 to 15, 17 to 24, 28, 29, 32, 34 to 41 and 49; and
(b)interviewed in the documents referred to in Categories 16 and 31.
5.The Discovery Application is otherwise dismissed.
6.The applicant is to inform the respondents, on or before 28 October 2020, which of the documents in the respondents’ discovery the applicant proposes to tender in the proceeding.
7.The respondents are to pay 50% of the applicant’s costs of the Discovery Application as agreed or assessed.
Timetabling
8.Paragraph 3 of the orders made on 4 September 2020 be varied so as to extend to 23 October 2020 the date for the respondents to file and serve any expert evidence upon which they seek to rely on the issue of liability.
9.Paragraph 11 of the orders made on 31 July 2020 be varied so as to extend to 6 November 2020 the date for the applicant to file and serve any expert evidence in reply.
10.Paragraph 12 of the orders made on 31 July 2020 be varied so as to extend to the week commencing 16 November 2020 the date for the parties to participate in a mediation before a mediator agreed between the parties.
11.The proceeding be listed for a case management hearing at 9.30am on 6 November 2020.
Note: Entry of orders is dealt with in Rule 39.32 of the Federal Court Rules 2011.
REASONS FOR JUDGMENT
(revised from transcript)THAWLEY J:
By an interlocutory application dated 11 September 2020, the applicant (ACCC) seeks discovery under rr 20.13 and 20.14 of the Federal Court Rules 2011 (Cth) (FCR) of what it has described as 58 categories of documents. The word “category” is somewhat misleading as many of the so-called categories seek discovery of particular documents. The respondents (Google) consent to an order requiring discovery of categories 1 to 7. The ACCC no longer presses for discovery of categories 25, 26, 27, 30, 33 and 42 to 48. The trial is set to commence on 30 November 2020.
Rule 20.11 of the FCR provides:
20.11 Discovery must be for the just resolution of the proceeding
A party must not apply for an order for discovery unless the making of the order sought will facilitate the just resolution of the proceeding as quickly, inexpensively and efficiently as possible.
Rules 20.13 to 20.15 of the FCR include:
20.13 Application for discovery
(1)A party may apply to the Court for an order that another party to the proceeding give discovery.
(2)The application must state:
(a) whether the party is seeking standard discovery; or
(b) the proposed scope of the discovery.
…
20.14 Standard discovery
(1)If the Court orders a party to give standard discovery, the party must give discovery of documents:
(a)that are directly relevant to the issues raised by the pleadings or in the affidavits; and
(b)of which, after a reasonable search, the party is aware; and
(c)that are, or have been, in the party’s control.
(2)For paragraph (1)(a), the documents must meet at least one of the following criteria:
(a) the documents are those on which the party intends to rely;
(b) the documents adversely affect the party’s own case;
(c) the documents support another party’s case;
(d) the documents adversely affect another party’s case.
(3)For paragraph (1)(b), in making a reasonable search, a party may take into account the following:
(a) the nature and complexity of the proceeding;
(b) the number of documents involved;
(c) the ease and cost of retrieving a document;
(d) the significance of any document likely to be found;
(e) any other relevant matter.
(4)In this rule, a reference to an affidavit is a reference to:
(a)an affidavit accompanying an originating application; and
(b)an affidavit in response to the affidavit accompanying the originating application.
Note: Control is defined in the Dictionary.
20.15 Non‑standard and more extensive discovery
(1)A party seeking an order for discovery (other than standard discovery) must identify the following:
(a)any criteria mentioned in rules 20.14(1) and (2) that should not apply;
(b)any other criteria that should apply;
(c)whether the party seeks the use of categories of documents in the list of documents;
(d)whether discovery should be given in an electronic format;
(e)whether discovery should be given in accordance with a discovery plan.
(2)An application by a party under subrule (1) must be accompanied by the following:
(a)if categories of documents are sought—a list of the proposed categories; and
(b)if discovery is sought by an electronic format—the proposed format; and
(c)if a discovery plan is sought to be used—a draft of the discovery plan.
(3)An application by a party seeking more extensive discovery than is required under rule 20.14 must be accompanied by an affidavit stating why the order should be made.
(4)For this Division:
category of documents includes documents, or a bundle of documents, of the same or a similar type of character.
Note:A discovery plan is a plan that has regard to the issues in dispute and the likely number, nature and significance of the documents discoverable in relation to those issues—see the Court’s Practice Note CM6, ‘Electronic Technology in Litigation’.
The relevant principles are well known and need not be repeated. A useful discussion of r 20.14 and the phrase “directly relevant” may be found in the decision of Barker J in Dennis v Chambers Investment Planners Pty Ltd (2012) 201 FCR 321 at [18] to [33].
Although in form the application has been made under rr 20.13 and 20.14, it is in substance an application under r 20.15. Rule 20.14 provides for standard discovery. The categories of documents sought by the ACCC are not put forward as reflecting all the documents which would fall within an order for standard discovery. Rather, what is sought here is discovery of specifically identified documents and documents within certain categories.
BACKGROUND
The background so far as is relevant to this application may be briefly stated. On 13 or 14 August 2018, the Associated Press published an article (AP Article) which included:
So the company lets you “pause” a setting called Location History.
Google says that will prevent the company from remembering where you’ve been. Google’s support page on the subject states: “You can turn off Location History at any time. With Location History off, the places you go are no longer stored.”
That isn’t true. Even with Location History paused, some Google apps automatically store time-stamped location data without asking.
…
To stop Google from saving these location markers, the company says, users can turn off another setting, one that does not specifically reference location information. Called “Web and App Activity” and enabled by default, that setting stores a variety of information from Google apps and websites to your Google account.
When paused, it will prevent activity on any device from being saved to your account. But leaving “Web & App Activity” on and turning “Location History” off only prevents Google from adding your movements to the “timeline,” its visualization of your daily travels. It does not stop Google’s collection of other location markers.
The ACCC was aware of the AP Article from the commencement of its investigation which led to the present proceedings.
The ACCC was also aware, by at least 17 September 2018, that the Arizona Attorney General’s office had announced an investigation under state consumer protection law against Google in relation to its location tracking practices.
On 27 May 2020, proceedings were brought in the United States by the Attorney General of the State of Arizona against the first respondent, Google LLC. The originating process filed in the Superior Court of Arizona, being a “Complaint for Injunctive and Other Relief” (Complaint), is in evidence on this application in redacted form. The Complaint has been redacted to reflect claims of confidentiality. It was first received in this form by the ACCC’s solicitor on 24 August 2020.
Categories 8 to 58 relate to the Complaint, albeit they might also be said to relate to the AP Article:
·Category 8 seeks discovery of the Complaint without certain redactions;
·Categories 9 to 49 seek discovery of specific documents referred to in the Complaint;
·Categories 51 to 53 seek documents relating to matters referred to in the Complaint;
·Categories 54 to 58 seek data relating to matters referred to in the Complaint.
Categories 8 to 49, being the unredacted Complaint and the documents identified in the Complaint as exhibits, are readily available to be produced at presumably minimal expense. Discovery of categories 51 to 58 would require substantial inquiries, time and expense.
A central aspect of the ACCC’s case in the present proceedings, as set out in the Concise Statement, is that Google is alleged to have engaged in misleading and deceptive conduct because:
(1)between January 2017 and October 2018 Google represented to users that, if the Location History setting was turned “off”, Google would not obtain or use personal data about the user’s location: [27]-[29], [49] of the Concise Statement;
(2)in fact, where Location History was turned “off”, but the Web & App Activity setting was turned “on”, personal data about the user’s location continued to be obtained and retained by Google and used for its own purposes: [57] of the Concise Statement.
The Complaint is significantly broader in scope than the present proceedings. This may be seen from Part III of the Table of Contents to the Complaint, which provide:
III. FACTUAL ALLEGATIONS ........ ........ ........ ........ ........ ........ ........ ........ ........ .... 6
A. Google Engages in Acts and Practices In Connection With the Sale and Advertisement of Merchandise In And Affecting The State of Arizona 6
B. Overview of Google’s Many Location-Related Settings ........ ........ ...... 10
C. Google Admits Its Location-Related Settings Are a “Mess” That Mislead and Deceive ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ... 12
1. Google Misleads and Deceives Users Through Its Location History and Web & App Activity Settings ........ ........ ........ ........ . 14
2. Google Misleads Users Into Sharing Their Location Via Its Misleading WiFi Scanning and WiFi Connectivity Settings ...... 19
D. Google Uses Its Users’ Locations Even When Users Turn Off the Relevant Permissions ........ ........ ........ ........ ........ ........ ........ ........ ........ .... 21
1. Google Shares Location with Apps That Users Explicitly Forbid From Using Location ........ ........ ........ ........ ........ ........ ........ ........ .. 21
2. Google Collects Location Data Even When Users Turn Their Device Location Off ........ ........ ........ ........ ........ ........ ........ ........ ... 24
3. Google Serves Personalized Ads Based on User Location Even When Users Turn Off Personalization ........ ........ ........ ........ ........ 26
E. Google Automatically Changes the State of Permissions Without Notifying Users ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ..... 27
F. Google Changes the Android User Interface to Increase Location [redacted] at the Expense of User Choice and Consent ........ ........ ........ 29
G. Google Misleads and Deceives Users Regarding Its Deletion of Their Location Information ........ ........ ........ ........ ........ ........ ........ ........ ........ .... 33
H. Google Has Engaged In Willful Violations Of The Arizona Consumer Fraud Act ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ...... 34
Categories 1 to 7
As noted, it is agreed that discovery of these categories should be ordered.
Category 8
Category 8 seeks a copy of the Complaint with the redactions removed from paragraphs 42 to 70. Paragraphs 42 to 49 of the Complaint appear under the heading “Google Admits its Location-Related Settings Are a ‘Mess’ that Mislead and Deceive” and paragraphs 50 to 70 appear under the heading “Google Misleads and Deceives Users Through Its Location History and Web & App Activity Settings”. Relevant parts of the redacted Complaint form Annexure A to these reasons.
The ACCC submitted that, whilst the Complaint could not be tendered as evidence against Google in these proceedings, at least to prove the facts asserted in the Complaint, an unredacted version of the document is necessary so as to assist to understand the import of the various documents referred to therein in circumstances where the parties will not have available to them all documents which were available to the Arizona Attorney-General, or at least the draftsperson of the Complaint.
In my view, to the extent the documents in categories 9 to 49 should be discovered, it is those documents which might have sufficient direct relevance to warrant an order that they be discovered. The description of those documents by, or the perceived significance of the documents to, the person drafting the Complaint is not of sufficient direct relevance to warrant an order for discovery of category 8.
Categories 9 to 49
Categories 9 to 49 are the various documents which are referred to in [42] to [70] and [77] of the Complaint. As mentioned, the ACCC no longer presses for discovery of categories 25, 26, 27, 30, 33 and 42 to 48.
Categories 9 to 24 and 34 to 41
The ACCC submitted that the relevance of categories 9 to 24 and 34 to 41 lay in the opinions or “admissions” by Google employees about the confusing nature of the Location History settings.
The ACCC submitted that, to the extent that the view of Google employees was that aspects of what they told users, as described in the AP Article, was misleading or confusing, that is a matter which, although not determinative, supported the ACCC’s case and which was, therefore, directly relevant.
The ACCC’s submissions were made by using category 40 as an example. Category 40 sought production of Exhibit 224 to the Complaint. That Exhibit was an email chain between Google employees containing various observations about the AP Article. It included a comment from one Google employee as follows:
Although I know how it works and what the difference between “Location” and “Location History” is, I did not know Web and App activity had anything to do with location. Also seems like we are not very good at explaining this to users.
Similar comments were also made by other employees.
The ACCC accepted that the question of whether the material Google presented to users was misleading was ultimately one for the Court, referring to Rhone-Poulenc Agrochimie SA v UIM Chemical Services Pty Ltd (1985) 10 FCR 567 at 585-586; Baxter v British Airways PLC (1988) 82 ALR 298 at 303. The ACCC distinguished the evidence in Exhibit 224 (category 40) from evidence which might be given by users. This was evidence of employees, which might constitute admissions and the weight of which might vary depending upon the relevant employee’s position and perhaps experience.
Leaving category 40 aside, the ACCC submitted:
(1)The documents referred to in categories 9 to 24 were set out in a section of the Complaint containing an allegation that “for years, Google has known that the user experience they designed misleads and deceives users”. It was submitted that, in most instances, it was not possible to identify the precise difficulty to which the document was directed. This was in part a function of the redactions. It was submitted it was reasonable to infer that these matters related to location settings, the primary matter to which these proceedings are directed.
(2)The documents referred to in categories 34 to 41 were documents which were said by the Complaint to “confirm the conclusion of the AP story” or to contain admissions by Google employees in relation to disclosures about Web & App Activity or Location History.
Google submitted that the subjective views of a handful of individual employees were not probative of whether the alleged representations were misleading in the relevant sense, referring to Telstra Corp Ltd v Phone Directories Co Pty Ltd (2014) 316 ALR 590 at [564]. This submission was emphasised in particular in relation to categories 12, 13 and 38 to 41.
At [564] to [566] of Telstra, Murphy J addressed the utility of evidence of users. His Honour concluded that, in the particular case before him, the evidence was of little utility. His Honour did not conclude that such evidence could never be probative:
The viva voce evidence has little utility
[564] Even if I accept the viva voce evidence that directory users and/or directory advertisers were confused as to the origin of the PDC directories, the isolated examples put forward by Telstra are of little assistance in relation to whether a significant proportion of the two classes were misled by the respondents’ conduct: National Exchange [National Exchange Pty Ltd v Australian Securities and Investments Commission (2004) 61 IPR 420] at [23] per Dowsett J. The evidence shows that Telstra looked hard for instances of directory users and advertisers who were misled or confused as to the origin of the PDC directories, including by making “trap” telephone calls. However, it adduced firsthand evidence of very few and comparatively few second-hand accounts. The examples given stand in contrast to the many millions of impugned directories published by the respondents since 2005 and the tens of millions of sales contracts that must have been made between advertisers and PDC.
[565] As French J said in State Government Insurance Corp v Government Insurance Offıce (NSW) (1991) 28 FCR 511; 101 ALR 259 at 278; 21 IPR 65 at 84 (State Government Insurance):
Generally speaking … evidence from consumers that they have been misled by the impugned conduct is of limited utility. It has no statistical significance and the court cannot draw inferences from it that any section or fraction of the population will have similar reactions. But if the inference is open, independently of such testimonial evidence, that the conduct is misleading or deceptive or likely to mislead or deceive, then it may be that the evidence of consumers that they have been misled can strengthen that inference. [Emphasis added in Telstra.]
[566] As Franki J said in Taco Bell [Taco Co of Australia Inc v Taco Bell Pty Ltd (1982) 42 ALR 177] at 202:
[E]vidence that some person has in fact formed an erroneous conclusion is admissible and may be persuasive but is not essential. Such evidence does not itself conclusively establish that conduct is misleading or deceptive or likely to mislead or deceive. The court must determine that question for itself. The test is objective. [Emphasis added in Telstra, citations omitted.]
The relevant inquiry is by reference to a hypothetical ordinary or reasonable member of the class, the test is objective, and proof of actual deception is no substitute for the court’s exercise of independent judgment.
In any event, these categories are not directed to the evidence of users, but to the evidence of employees of Google, some of whom might have had relevant specialised knowledge or experience. This might affect the weight to be given to the evidence.
Google also submitted that the statements were not admissions by Google and would be inadmissible hearsay. Admissibility is a quite different question to the relevance of documents. Discovery is not only ordered where it can be shown on the application for an order for discovery that the document would be admissible in evidence. Whether a particular discovered document is admissible for a particular purpose arises when it is sought to tender the document in evidence. Whether the material sought on discovery contains statements which are admissible as admissions is a question which can only be addressed if and when the document containing the asserted admission is in fact before the Court. I would accept that, on the facts in a particular case, the question of admissibility might not be irrelevant to whether the discretion should be exercised to order discovery. The present case is not one where that issue has significance.
Google submitted in relation to category 9, being Exhibit 56 to the Complaint, that it could not be directly relevant because it was dated October 2014 which predated the conduct relevant to these proceedings which was alleged to have occurred between 1 January 2017 and 29 October 2019. The mere fact that the document pre-dated the relevant conduct does not establish that the document is not “directly relevant”. Without wishing to suggest that it is the case, if the document might tend to establish that Google intended to achieve a particular result by its design that material might be directly relevant to whether the intended result was achieved, perhaps particularly if the relevant employees had specialised knowledge or experience – see: Bridge Stockbrokers Ltd v Bridges (1984) 4 FCR 460 at 474-5; Rhone-Poulenc at 488; Unilever Australia Ltd v Goodman Fielder Consumer Foods Pty Ltd [2009] FCA 1305 at [43] and [45].
The point is perhaps illustrated by reference to the quote in the Complaint at [44] where reference is made to Exhibit 207 (category 12) which states:
The current UI [user interface] feels like it is designed to make things possible, yet difficult enough that people won’t figure it out.
If documents established that settings relevant to this case were designed with an intention to the effect suggested, such material might, on the authorities referred to, be probative of whether the resulting conduct was in fact misleading or deceptive. I do not regard it as significant or decisive that this precise issue is not identified in the Concise Statement as a contention advanced by the ACCC. The issue is relevant on that facts as identified in the Concise Statement.
Google’s objections to categories 10, 11, 15 to 20, 23 and 24, 32 and 34 to 37 relied on a submission that, in the Complaint where the relevant document was referred to, what was said about the document had been wholly redacted with the result that the ACCC could not establish that the document was “directly relevant” within the meaning of r 20.14(2).
Where a specific document not already in the possession of a party is sought by way of discovery, a conclusion of “direct relevance” is necessarily drawn without the document being before the Court. The assessment is made from the surrounding circumstances. In the present case, it is enough to observe that the headings of the Complaint give rise to a reasonable inference of sufficient “direct relevance” to warrant ordering discovery. The fact that some of the material which might indicate the content of the document in more detail has been redacted does not mean that an inference of relevance cannot be drawn from the Complaint as a whole.
The ACCC bore the onus of establishing that discovery should be ordered and this included establishing that the documents were sufficiently “directly relevant”. The ACCC could only point to the context in which the documents were referred to and the weight of its evidence must be assessed understanding that it did not have access to the unredacted documents or what was said about them where such material has been redacted in the Complaint.
As to context, Google submitted that the categories could not be shown to be directly relevant having regard to the fact that the present case was more narrow than that pleaded in the Complaint. I accept that the Complaint is more expansive but the particular documents sought (being categories 9 to 24 and 34 to 41) are only a subset of the documents referred to in the Complaint and are referred to in the Complaint in a context which satisfies me that it is sufficiently likely that the documents are directly relevant such as to make them candidates for discovery assuming discovery were otherwise appropriate. In my view, the direct relevance of the documents has been sufficiently established. Google’s submission in this respect also related to the other documents sought within categories 9 to 49 (namely 28, 29, 31, 32 and 49) and I reach the same conclusion although I address those categories further below.
In relation to categories 34 to 36, Google made an additional submission. First, Google noted that these documents appeared in [59] of the Complaint where it was pleaded that “[t]estimony from Google employees and Google’s internal documents confirm the conclusion of the AP story”. It was submitted that such documents were not relevant to any fact in issue because the Statement of Agreed Facts and Screenshot Bundle together describe how the relevant settings operated in the period in question. I accept that how the relevant settings operated is a fact in issue. So too is whether Google represented to users that if Location History was turned “off”, Google would not obtain or use personal data about the user’s location. Whilst that is ultimately an assessment to be made objectively by the Court, it does not follow, as explained earlier, that such an assessment cannot rationally be informed by material of the kind sought in categories 34 to 36.
In relation to category 35, Google submitted that a redacted copy of the document (Exhibit 221) made it apparent that the document was not “directly relevant”. It was not stated that a review of the document in its unredacted form had been undertaken or that such a review had indicated that the document was not “directly relevant”. Having regard to the context in which the document is referred to I am satisfied that the likelihood of relevance is sufficient to warrant an order for discovery.
Google also submitted that discovery of the maintained categories within 9 to 49 should be refused on the basis of the ACCC’s delay in seeking these documents and in light of the impending trial. Further, it was submitted that, if discovery were ordered, the hearing date might need to be vacated to permit Google to adduce further evidence if it chooses to do so.
The application for discovery has been made late. However, that must be viewed in context. The ACCC’s solicitor was provided with a redacted copy of the Complaint on 24 August 2020. He made inquiries to obtain publicly available copies of the exhibits two days later. This yielded partially redacted copies of some of the exhibits. On 2 September 2020, he wrote to Google’s solicitors requesting copies of certain documents. This request was declined by letter dated 9 September 2020. Further communications occurred with respect to discovery. I also should refer to the fact that the parties have been cooperating to produce a Statement of Agreed Facts and a Screenshot Bundle, and to some extent this will have affected the parties’ views about the appropriate timing of an application for discovery.
I do not consider that the delay in seeking discovery requires refusing an order for discovery of the documents in categories 9 to 24 and 34 to 41, which are readily available. As to the proposition that the hearing date might need to be vacated if discovery of these documents is ordered, that would depend on a number of matters including whether, having inspected the documents, the ACCC wished to rely on the documents in the proceeding. The answer to that question would depend on a number of matters including the perceived probative value of the material. Given that minimal time is required for discovery and inspection of these documents and that the expense of doing so is also not great, in my view discovery should be ordered. I recognise that the consequence may be that Google may consider it necessary to adduce evidence to put any material relied upon by the ACCC in context or to establish that it does not have the significance which the ACCC contends and that this may result in the vacation of the hearing date. That would be unfortunate but in the particular circumstances of this case I consider discovery should be ordered despite that being a real risk. The matter is of significant public importance. The hearing date was fixed early in the proceedings. The matter is not so urgent that it must be heard this year.
Categories 28, 29, 31 and 32
The ACCC used category 28 to illustrate the relevance of the documents in categories 28, 29, 31 and 32. Category 28 sought discovery of Exhibit 20 to the Complaint.
Exhibit 20 was an email chain between Google employees following the publication of the AP Article. One email in the chain stated:
Your name came up today during our Monday morning “Oh Shit” meeting in relation to this story. Both comms and policy are looking for an update on where we are in terms of fixing “location history” fixes and having one single place to turn off instead of 3 …
The ACCC submitted that this email and other similar ones described what appeared to be a crisis meeting held by Google employees in respect of the AP Article and described the actions being taken to “fix” Location History.
It was submitted that the fact that such a meeting was held and that such fixes were deemed necessary supported the ACCC’s case that the messages which Google provided to users at the time were misleading.
Google submitted that, to the extent documents indicated improvements could be made, or different processes were implemented, this was not probative of whether the relevant conduct was, at the time, misleading or deceptive. In my view, that may or may not be so. The question is plainly whether the conduct at the time was misleading or deceptive. It does not follow as an absolute proposition that evidence of later events is necessarily irrelevant to that question. It all depends on the particular facts.
Google submitted that, to the extent documents indicated that employees of Google sought to make more explicit disclosures after the AP Article was published, that is not evidence from which the Court could infer that its earlier disclosures were misleading, referring to Australian Competition and Consumer Commission v Hillside (Australia New Media) Pty Ltd t/as Bet365 [2015] FCA 1007 at [122].
Hillside is not authority for an absolute proposition in the terms put in Google’s written submissions. What Beach J stated was:
[122] The ACCC has contended that the presentation of these “Key Offer Terms“ in the current OPEN ACCOUNT OFFER, in the central part of the opening webpage in white text against a black background, stands in contrast to the manner of presentation of those terms as displayed in the preceding DEPOSIT BONUS Offer and FREE BETS Offer. That may be so, but the relevance of the current presentation mode to the issues that I have to decide is questionable. All that it shows is what could have been done earlier. It does not by any necessary implication demonstrate that what was done earlier was misleading or deceptive.
What inference is appropriately to be drawn depends upon an assessment of the particular document in light of all of the surrounding circumstances. It is obvious that proof of a change, or even an acknowledged improvement, in conduct does not inexorably lead to the conclusion that what occurred earlier was misleading or deceptive. However, whether subsequent conduct is probative of whether the pre-existing state of affairs was misleading or deceptive depends on the nature of the subsequent conduct and the nature of the conduct contended to be misleading or deceptive and the precise inter-relationship between the two.
In my view, these categories should be discovered. I note that Google’s submissions in relation to delay and prejudice to the hearing date were also made in relation to these categories. I have earlier stated my conclusion in this regard so far as concerns the specific documents identified in the Complaint, namely those in categories 9 to 49.
Category 49
Category 49 sought an unredacted copy of Exhibit 236. Exhibit 236 referred to the proportion of signed-in search users who have Web & App Activity turned on. It then contains a redacted section before a reference to Location History. The ACCC submitted that this redacted section was likely to refer to the interplay between Web & App Activity and Location History and, for that reason, was likely to advance one parties’ case and, accordingly, be directly relevant within the meaning of r 20.14(1)(a).
The ACCC also referred to the fact that it had served an expert report from Professor Slonim, a behavioural economist, which contained various observations about the likelihood of users changing the Location History and Web & App Activity settings from their default state. It was submitted that the data (see categories 54 to 58 below) and discussion in this email was likely to be relevant to that issue.
The critical part of the document which was redacted was redacted on the ground of “privilege” which I infer to be a claim in the US of legal privilege. The fact that an equivalent claim might be made in Australia is not of itself a basis not to order discovery of it.
In my view, this category should be discovered but not because of any direct relevance with the report of Professor Slonim. Again, I note that Google’s submissions in relation to delay and prejudice to the hearing date were also made in relation to this category. I have earlier stated my conclusion in this regard so far as concerns all of the documents in categories 9 to 49.
Category 50
Category 50 was:
50 Documents which identify the name and job title (at the date of the communication or interview) of the persons:
(a) involved in the communications referred to in paragraphs 9-15, 17-26, 28-30, 32-41, 43-47 and 49 above; and
(b) interviewed in the documents referred to in paragraphs 16, 27, 31, 42, and 48 above.
As I understood the position of the parties, it was agreed that Google should be ordered to provide a single document recording the job titles at a particular point in time of each of the persons mentioned in category 50, assuming that was relevant by reason of other orders for discovery.
Categories 51 to 53
As noted earlier, after the publication of the AP Article, a group of Google employees held a crisis meeting, which included a discussion of how to “fix” Location History and, according to the Complaint, how to simplify Google’s location settings.
Categories 51 to 53 were:
51 In respect of the meeting referred to at paragraph 54 of the Complaint:
(a) All documents provided to participants either at or prior to that meeting;
(b) All documents recording what was said during the meeting;
(c) All communications between any combination of persons who attended that meeting between 13 August 2018 and 31 October 2018 in relation to:
(i) Location History;
(ii) Web & App Activity; or
(iii) The AP Story entitled “Google tracks your movements, like it or not” dated 13 August 2018.
52 Documents which refer to:
(a) the AP Story entitled “Google tracks your movements, like it or not” dated 13 August 2018; or
(b) the meeting referred to at paragraph 54 of the Complaint; or
(c) changes (or potential changes) to any of the screens contained in the Screenshot Bundle, including Google’s privacy policies,
and which were, at any time between 13 August 2018 – 31 October 2018 provided to:
(d) the board (or equivalent decision-making body) of Google LLC; or
(e) the persons or decision-making body responsible for determining the content of the screens contained in the Screenshot Bundle.
53 Any memorandum, report, analysis or other document created in the period 13 August 2018 – 31 October 2018 in respect of the following matters:
(a) the matters raised by the AP Story entitled “Google tracks your movements, like it or not”;
(b) the behaviour of users in respect of their “Web & App Activity” or “Location History” settings; and
(c) changes (or potential changes) to any of the screens contained in the Screenshot Bundle, including Google’s privacy policies,
produced by or provided to any person or group responsible for investigating and considering the matters referred to in sub-paragraphs (a) to (c).
In summary, the ACCC sought documents which recorded the content and outcome of the meeting, including Google’s investigations and actions taken thereafter. It was submitted that the documents could be expected to contain:
(a)direct admissions by relevant employees as to the misleading nature of Google’s communications;
(b)a statement of what went wrong and what was done to fix the problem, which itself would amount to an admission of what went wrong;
(c)relevant data (available at the time) about the number or proportions of users who had the relevant settings “on” or “off”; and
(d)relevant data (available at the time) about the way that users navigated their way to the relevant settings.
In my view, discovery of these categories should not be ordered at this point. It is not fair to require Google to engage in what is likely to be a time consuming and costly exercise this close to the hearing date. Further, Google submitted and I accept that if these categories were ordered, the hearing would inevitably have to be vacated. That is not the inevitable outcome of ordering discovery of the categories in respect of which I have ordered discovery.
Categories 54 to 58
Categories 54 to 58 arise from category 49 (Exhibit 236). The ACCC submitted that the redacted email being Exhibit 236 gives rise to an inference that there is available to Google a ready source of data as to how many users have Web & App Activity turned on at any point in time and that the same data is available in respect of Location History.
The ACCC submitted that this data was relevant to the report of its expert, Professor Slonim. I understood this to be the main reason it was contended to be relevant. It was also said that the data was relevant because, although not determinative, the data might assist the determination of the objective question before the Court about whether the screenshots were misleading or deceptive in the manner contended in the Concise Statement.
Google submitted that the ACCC had offered no compelling explanation for why the data sought was relevant or why the data was not sought earlier. Google emphasised that Professor Slonim stated in his report that, while obtaining certain data would be informative, that data was not necessary for him to express an opinion in relation to any question he had been asked, as his opinions would nevertheless hold. Google submitted, perhaps more importantly, that in any event, the data requested by categories 54 to 58 were not of the same nature as that identified by Professor Slonim as potentially useful.
I am not satisfied that the data is of sufficient relevance to warrant an order for discovery. I am not satisfied that the data sought is data which falls within the kind of data which Professor Slonim stated would be useful to review. Accordingly, I do not order discovery of categories 54 to 58.
CONCLUSION
In summary, orders should be made requiring Google to give discovery of categories 1 to 7, 9 to 24, 28, 29, 31, 32, 34 to 41 and 49. Subject to hearing further from the parties, an order should be made giving effect to the agreement of the parties in relation to category 50. A timetable should be made requiring the ACCC to file or otherwise notify Google of any of the discovered documents it intends to rely upon in order, amongst other things, to provide Google a proper opportunity of considering whether it wishes to rely upon further material in light of the ACCC’s position.
I certify that the preceding sixty-four (64) numbered paragraphs are a true copy of the Reasons for Judgment of the Honourable Justice Thawley. Associate:
Dated: 27 October 2020
ANNEXURE A
1
14
1