The Competition Appeal Tribunal granted a collective proceedings order on 5 August 2026 against six companies in the Google group, sending an advertiser damages claim valued by the claimant side at around £5 billion to trial on an opt-out basis. Every UK-domiciled business that paid for Google search advertising between 1 January 2011 and 15 April 2025 is now inside the class unless it removes itself.

The judgment, carrying the neutral citation [2026] CAT 65 and case number 1720/7/7/25, was handed down by a panel of the Honourable Mr Justice Meade as chair, sitting with John Davies and Robert Herga. It follows a hearing held at Salisbury Square House in London on 8 July 2026, and it resolves an application that had been sitting with the Tribunal since April 2025.

The claimant is Or Brook Class Representative Limited, a special purpose vehicle whose sole director and member is Dr Or Brook, a competition law academic. The proposed defendants are Alphabet Inc, Google LLC, Google Ireland Limited, Google UK Limited, Google Asia Pacific Pte Ltd and Google Commerce Limited. The Tribunal treated them collectively as a single undertaking throughout, noting that it was unnecessary for present purposes to distinguish between them.

Certification is not a finding of liability. It is the procedural gateway that decides whether a claim can be pursued collectively at all, and on what basis. What the Tribunal decided on 5 August 2026 is that Or Brook Class Representative Limited is suitable to act as class representative, that the claims are eligible for collective treatment, and, most consequentially, that the proceedings will run on an opt-out footing rather than requiring each advertiser to sign up individually.

That last point was the fight. According to the judgment, counsel for Google told the Tribunal that its objections to certification fell into four broad groups: the funding arrangements and the level of costs incurred and budgeted; certain amendments the class representative sought permission to make to its claim form; the class definition; and whether certification should be on an opt-out or opt-in basis. Google indicated that if the first three groups were satisfactorily addressed, it did not oppose opt-in certification. It opposed opt-out certification in any scenario.

The Tribunal inverted that ordering. Although Google had placed the opt-in question last and made its non-opposition conditional, the panel dealt with it first, describing it as the biggest issue and the one that took the most time and argument. The other three groups were resolved essentially in the class representative's favour, at least to the extent that none of them was treated as a bar to certification, though several will require further attention as the case progresses.

Who is in the class and what it covers

The proposed class definition, reproduced in the judgment and in the Tribunal's earlier notice of application, reads: all UK-domiciled Advertisers who, during the Relevant Period, paid for search advertising services provided by Google, whether they purchased directly or via a media agency.

The Relevant Period runs from 1 January 2011 to 15 April 2025. That is a span of just over fourteen years, closing on the date the collective proceedings claim form was dated. Media agencies themselves are excluded from the class, as is the class representative and anyone acting for or instructed by it. Natural persons who had died by the domicile date are excluded, as are legal persons that had been dissolved, struck off the register or otherwise ceased to exist by that date. The domicile date itself has not yet been fixed and will be determined by the Tribunal.

Google argued that the phrase "search advertising services" did not make clear what fell inside the definition. In its reply, the class representative clarified that the definition covers both text ads and product listing ads on Google's search engine results pages, although the definition itself was left unchanged.

The Tribunal did not regard the definition as problematic. According to the judgment, the fact that there might be arguments at the fringes does not affect that conclusion. One such fringe question was raised at the hearing: whether advertisements on YouTube would be covered. The class representative's answer, which the panel considered reasonable, was that video ads are not covered by the definition, and that factual inquiry or disclosure will be needed to establish whether text ads or product listing ads are served via YouTube. Google did not appear to have difficulty with the video ads proposition. A suggestion in Google's written response that the class definition ought to be tied to market definition was not pursued at the hearing.

The panel also noted that Google had been able to compile data about the proposed class for the purposes of its own expert evidence, a point that recurs later in the judgment when the Tribunal turns to how any damages award would actually be distributed.

The conduct alleged

The class representative alleges that Google holds dominant positions in a set of connected markets: the worldwide market for licensable smart mobile operating systems, the worldwide market for Android app stores, the various national markets for general search services, and the UK or wider market for search engine management services.

The alleged abuse comprises three principal elements, according to the judgment.

Android distribution and anti-fragmentation

First, in relation to the Android ecosystem, Google is alleged to have imposed conditions in its mobile application distribution agreements with device manufacturers that had the effect of tying Google Search to the Play Store, and Chrome to Google Search and the Play Store. The class representative also relies on anti-fragmentation agreements and related obligations said to have restricted the development of alternative versions of Android, referred to in the pleadings as Android Forks.

Beyond the contractual tying, the claim alleges that Google offered substantial financial incentives to original equipment manufacturers and mobile network operators, including through placement agreements, revenue-sharing agreements and mobile incentive agreements, in order to secure pre-installation and prominent placement of Google Search and Chrome and to discourage the installation of competing search services.

The pleaded theory is that these strands were mutually reinforcing. Tying arrangements and anti-fragmentation obligations meant that manufacturers wanting to ship devices with the Play Store were, in practical terms, also required to ship Google Search and Chrome. Restrictions on Android Forks are said to have prevented the emergence of alternative Android ecosystems capable of operating independently of Google's services, which in turn reinforced the central position of the Play Store.

Browser default payments

Second, the claim concerns agreements with browser developers under which Google Search was set as the default search engine in exchange for monetary incentives. Particular reliance is placed on Google's arrangements with Apple, though the class representative also refers to agreements with Mozilla and Opera.

The Tribunal's earlier notice of application put the alleged effect plainly: Google ensures that a very large proportion of devices and browsers are supplied so that Google Search is pre-installed, prominently placed or set as the default, meaning device users are overwhelmingly likely to use Google Search instead of any rival search engine. The consequence pleaded is that Google's own search platform becomes the only viable means of serving search advertisements to the vast majority of device users.

Search Ads 360 and multi-homing

Third, the claim challenges Google's conduct in relation to Search Ads 360, the advertising management platform used by agencies and large advertisers to run campaigns across different search providers. According to the judgment, it is alleged that Google introduced functionality for its own advertising services while refusing or delaying equivalent functionality for Microsoft's competing offering, restricting advertisers' ability to use multiple search and advertising platforms effectively. The pleaded effect is that users of SA360 are artificially biased towards advertising on Google.

The technical substance of that allegation is familiar territory for anyone who has managed enterprise search campaigns. SA360, formerly DoubleClick Search, sits above the individual ad platforms as a cross-engine layer, pulling reporting and applying portfolio bid strategies across Google Ads, Google Shopping and Microsoft Advertising. Feature parity between what the platform can do for Google inventory and what it can do for rival inventory determines, in practice, how much of a campaign an advertiser can realistically run outside Google. Search Ads 360 integrated Microsoft automated bidding in August 2024, several years after equivalent Google functionality was available, and the platform has continued to add partners, most recently bringing Koddi's commerce media network inside SA360 in June 2026.

The overall case is that these strands of conduct formed a single and continuous abuse of dominance aimed at securing and maintaining Google's position in general search and search advertising, allowing it to charge supra-competitive prices for search advertising and to reduce the value obtained from that advertising. The consequence pleaded is that advertisers paid more and achieved a lower return on advertising spend than they would have in a competitive market.

Follow-on and standalone elements

The claim is advanced on both a follow-on and a standalone basis. The follow-on component relies in part on the European Commission's decision of 18 July 2018 in case AT.40099, the Google Android decision, as subsequently reviewed by the General Court and the Court of Justice of the European Union. The class representative contends that certain aspects of the conduct alleged correspond to conduct already found unlawful in those proceedings, and that the tying arrangements and anti-fragmentation obligations were conclusively established there, while alleging that such conduct continued beyond the period the Commission considered.

That foundation firmed up shortly before certification. The Court of Justice dismissed Google's appeal in full on 2 July 2026, confirming a fine of 4.125 billion euros and rejecting the argument that an as-efficient-competitor test is a universal requirement under Article 102 of the Treaty on the Functioning of the European Union, with Alphabet remaining jointly liable for 1.52 billion euros. Or Brook referred to that judgment in a LinkedIn post published after certification, describing the present case as, in part, a follow-on action from the Google Android case that had been confirmed by the Court of Justice the previous month.

The standalone component covers elements of Google's conduct not established as unlawful by the Commission decision but which the class representative says form part of the same abuse, including conduct that continued and evolved after the period covered by the 2018 decision.

Supporting material cited in the claim includes the Competition and Markets Authority's online platforms and digital advertising market study of July 2019, its mobile ecosystems market study final report of 10 June 2022, its mobile browsers and cloud gaming report of 12 March 2025, and its notices under section 11(1) of the Digital Markets, Competition and Consumers Act 2024 commencing initial strategic market status investigations into Google. The claim also relies on the judgment of the District Court for the District of Columbia in United States v Google, case 20-cv-3010, dated 5 August 2024, which found that Google had illegally maintained a monopoly in general search services and general search text advertising. Google filed a 100-page opening brief on 22 May 2026 asking the D.C. Circuit to reverse that ruling in its entirety.

The central question: opt-out or opt-in

Under section 47B of the Competition Act 1998 and rules 77 to 79 of the Competition Appeal Tribunal Rules 2015, the Tribunal must be satisfied of two conditions before making a collective proceedings order: the authorisation condition, concerning whether the proposed representative can be authorised, and the eligibility condition, concerning whether the claims are suitable for collective treatment. The judgment sets out that framework by reference to the summary given in Spottiswoode v Airwave [2025] CAT 60, which was itself cited in Stasi v Microsoft [2026] CAT 34.

The choice between opt-in and opt-out is governed by rule 79(3), which permits the Tribunal to take into account all matters it thinks fit, including the strength of the claims and whether it is practicable for the proceedings to be brought on an opt-in basis, having regard to the estimated amount of damages individual class members may recover.

Most of the argument turned on the Supreme Court's December 2025 decision in Evans v Barclays Bank Plc and others [2025] UKSC 48, which examined how that assessment should be conducted. The judgment quotes the Supreme Court at length. Among the passages set out, the Supreme Court stated that the assessment involved is not all or nothing, and that it is appropriate to view the strength of the claim on a sliding scale. On practicability, the Supreme Court said that the implication of rule 79(3)(b) is that if it is practicable for proceedings to be brought as opt-in proceedings, then generally speaking they should be.

Evans also addressed class composition. The Supreme Court said that where the Tribunal identifies groups of claimants with distinct profiles, it should consider the practicability of an opt-in claim for each group separately and then stand back and make an overall assessment of the balance of justice. In the passage Google leaned on hardest, the Supreme Court endorsed the Tribunal majority's conclusion in that case that a sub-class whose total claims were a tiny fraction of the aggregate should not alter the practicability conclusion, quoting the phrase "That would be to allow the tail to wag the dog".

Google argued that the long tail of smaller claims in the present proceedings should not be permitted to wag the dog of the higher-value claims held by a far smaller percentage of class members, and that large entities making up the bulk of the claim by value should not be allowed to bolster their position by bundling themselves with class members whose claims make up a tiny fraction of the total.

The class representative pointed to three post-Evans decisions, Stasi, Rodger v Alphabet [2026] CAT 49 and Ennis v Apple [2026] CAT 55, in each of which opt-in or partially opt-in certification was sought by defendants and rejected. It also cited Rodger v Alphabet [2025] CAT 45 and Spottiswoode v Airwave as examples of opt-out certification granted where the class was far smaller than here, at 2,200 in the former and between 400 and 2,000 in the latter.

Google's counter-examples pointed the other way: the Trucks litigation, where the Tribunal certified an opt-in claim with 11,000 claimants over a proposed opt-out alternative, and the Dieselgate group litigation order, where 1.6 million claimants participate.

Notification and distribution

Central to Google's position was the argument that the class representative had not shown it would be impracticable to notify a very large class, and had not engaged adequately with how an aggregate award would be distributed.

The evidence the Tribunal weighed included the parties' statements of case, a witness statement from Mr Teague of the class representative's solicitors, and a supplemental report of 10 June 2026 from Epiq, the claims handler retained by the class representative. A letter from Google's solicitors dated 29 June 2026, setting out what information Google says it does and does not hold about individual advertisers, arrived after both the Teague statement and the Epiq report.

The Tribunal was pointed about that sequencing. According to the judgment, the panel could see no reason why Google could not have provided the letter earlier, said it was material to assessing Google's criticisms that the class representative had been unable to deal with the matters raised in it, and stated that it was unfair for Google to pick away at the class representative's materials in those circumstances. The panel described Mr Teague's evidence as cogent and reliable.

Google identified what it said was a tension in the class representative's position: asserting that contacting class members would be difficult because Google may not be able to identify them, while simultaneously asserting that distribution ought not to be difficult because many class members are current Google account holders. The Tribunal agreed there was something of a tension, but found it less profound than Google claimed, and treated the position as clarified by the time of oral submissions.

The clarified position is the analytical core of the ruling. Contacting a class member to persuade it to opt in and contacting the same class member to hand it money are, according to the class representative, radically different exercises. The Tribunal agreed. It reasoned that class members can be expected to be far less attracted by a communication requiring them to consider joining a legal action with uncertain and distant prospects, and to commit themselves to further action, than by a communication at the distribution stage that simply requires them to accept money, or, easier still, to accept what would amount to a refund from Google if they remain account holders.

The Gutmann benchmark

Google relied heavily on Gutmann v First MTR South Western Trains Limited [2025] CAT 72, so far the only collective proceedings case to have reached distribution, where take-up was below 1 per cent. Epiq was the claims handler there too. Google's arithmetic was direct: on a 1 per cent take-up, an aggregate award of £5 billion would produce roughly £50 million actually distributed, a figure comparable to the total projected costs of the proceedings.

The Tribunal accepted that take-up in Gutmann had been extremely low but declined to treat it as a benchmark, finding the distribution exercise there totally different, particularly because of the absence of a good direct means of individual communication, which the panel expects to be present in this case.

Epiq's supplemental report set out points of difference between a distribution exercise based on defendant-held data where the defendant has a continuing relationship with class members, and a broader public-facing campaign of the Gutmann type. Those points, which the Tribunal described as highly material, persuasive and consistent with common sense, cover direct and credible communication rather than general advertisements or websites, the likelihood that class members will be accurately identified, the greater authority class members will attach to such communications, and a lower-friction claims experience in which class members have less to do to demonstrate they are genuine and to document their usage of the services.

Google's more granular objections were given short shrift. According to the judgment, the panel considered that Google majored on a number of points reflecting a nit-picking, negative approach, and which in some cases attacked arguments the class representative was not making. Google noted that it might not hold payment details for some class members, and that some might no longer be account holders. The Tribunal found those points did nothing of substance to suggest that direct contact with the majority of class members at the distribution stage would be impossible.

The panel went further on burden. It stated that it considered it inherently improbable that a company like Google, in the business in which it operates, cannot identify and contact its advertisers if it wants to, even where advertising is placed through agencies, and said that in practical terms the burden on this issue had shifted to Google to show difficulties existed.

One exchange at the hearing illustrates the gap. The Tribunal identified as an important though far from decisive question whether the unique identifiers used when media agencies place advertisements with Google map back to individual advertisers, since a distribution exercise would be considerably easier if they did. According to the judgment, Google was not able to answer.

On working through agencies, the Tribunal was careful to say what the class representative had not proposed. It does not mean handing agencies many millions of pounds and hoping the money is passed on, a proposition the panel accepted would be hopeless. It means securing agency assistance in an information-gathering and analysis task so that contact can be made with their clients, something the Tribunal considered agencies would obviously be motivated to help with, and which Google itself had postulated when arguing that an opt-in exercise would be practicable.

A fallback of ordering account credits to Google advertisers for unclaimed aggregate damages, discussed in BT v Le Patourel [2022] EWCA Civ 593, was described as somewhat speculative and uncertain, with some potential for that scenario, but not necessary to the decision.

The conclusion was that effective, direct distribution to specifically identified individual class members of an aggregate award far in excess of 1 per cent may reasonably be expected, and that effective recovery direct to class members would be much greater than the expected costs, large though those are, without the panel putting a specific number on the ratio.

Class profile

The judgment records that it was common ground that the great majority of the claim value is accounted for by a small percentage of class members, with a tiny percentage holding claims in the millions of pounds, while many class members have very small claims. Specific figures were withheld as commercially confidential.

The Tribunal attached importance to a middle tier: a material number of claims of middling size, in the tens of thousands of pounds, accounting for a significant proportion of the claimed damages. Google did not argue that the distribution of claim values allowed a clean subdivision into sub-classes, and did not contend for a hybrid arrangement with part opt-in and part opt-out.

On the practicalities of building an opt-in book across a class the judgment puts at about 880,000, the Tribunal accepted evidence that the exercise would be long and difficult. Making contact is only the start. Businesses would want to understand and ask questions about their liability for costs, their exposure to giving disclosure, and their relationships with the lawyers handling the case. The panel found that many class members would simply not join, that a significant proportion of the possible aggregate claim would be lost, and that the impact would be felt materially among the middle-sized claims.

Duration was the second effect. The Tribunal declined to put a specific figure on how long a book-building exercise would take but said it could be well over a year, perhaps two. That delay would be likely to push back the parallel Stopford proceedings, given the need to manage common issues efficiently, and the panel said it would be unfair for that to happen. Counsel for Google submitted that Stopford would not be at a complete halt because disclosure directions on common issues are being progressed, which the Tribunal accepted as partial amelioration only.

Two of the class representative's arguments received cooler treatment. On the submission that rearranging funding and insurance from an opt-out to an opt-in structure would be time-consuming and uncertain, the panel found it somewhat presumptuous for the class representative to have arranged the whole proceedings on an opt-out basis and then argue that changing it would be hard, while noting the overall decision would be the same either way.

On the submission that some class members would not opt in for fear of damaging existing commercial relationships with Google, the panel found no evidence that Google might retaliate against advertisers who joined, and said it was fair for Google to react as it did to the suggestion. But it accepted that there would be a general disinclination to enter into a dispute with a key commercial partner, supported by regulatory materials and by common sense. Again, the Tribunal said the outcome would have been the same without that factor.

Other factors and the cost-benefit balance

Three further factors were weighed. First, the Tribunal found these are not weak claims, observing that the simple fact they are largely follow-on demonstrates as much, and noting that Google did not really try to argue otherwise. That factor weighs in favour of opt-out, in contrast to the position in Evans.

Second, since the Stopford proceedings are opt-out and will be decided absent settlement, there would be a risk of inconsistency if the present proceedings were opt-in.

Third, opt-out increases the size of the claims and Google's overall exposure. The class representative argued that Google is so big and powerful that it need not care. The Tribunal rejected that framing, noting the same could have been said in Evans, and assessed the additional pressure as a modest difference that is not unfair in all the circumstances, adding that to the extent the claims are meritorious there are sound policy reasons why recovery should be maximised.

Having rejected the 1 per cent take-up scenario, the panel concluded that the prospective value of aggregate damages is much higher than the costs, large though they are, and that the cost-benefit analysis falls well in favour of certification on an opt-out basis.

The assessment paragraph is unequivocal. According to the judgment, the balance is very clearly in favour of opt-out certification, opt-in would carry many practical problems, would be slow, would have an undesirable and unfair impact on Stopford, and would unjustifiably reduce the size of the overall claim. The panel added that there is almost nothing to be said in favour of opt-in by contrast, and that in so far as there is a presumption in favour of opt-in, it is comfortably displaced on the facts.

Money: funding, costs and governance

Certification cleared, but the Tribunal spent a substantial portion of the judgment on cost discipline, and the language is markedly less accommodating than in the opt-out analysis.

The class representative's rate

Google submitted that Dr Brook's hourly rate of £250 is excessive, relying on the expectation in Waterside Class Limited v Mowi ASA [2026] CAT 32 that governance arrangements should be proportionate and aligned with public sector benchmarks.

The Tribunal noted it had heard relatively little argument on the broader context, and that the decision of the Tribunal in case 1759/7/7/25 JLP A&A v Apple Inc. and Amazon, with judgment expected later this year, may give detailed consideration to the issue. It made clear its decision is not intended to set a pattern, precedent or significant data point.

Three factors moved the panel. Dr Brook is willing to cap what she charges for any one day at £750 even if she works more than three hours; she does not charge in the way a solicitor would, frequently not billing for individual tasks that take only a short time; and she has given information about monthly totals billed that the Tribunal described as by no means excessive. The effective hourly rate will therefore be considerably lower than £250, particularly for periods when she works whole days, though the panel said it is not readily possible to calculate the effective rate with precision.

Balancing that against her contribution, the Tribunal found she makes major, careful, detailed contributions to the shaping, drafting, management and governance of the claim, and is not a mere figurehead. It concluded the rate is not undue or excessive, while acknowledging that later and wider-ranging decisions may set policy under which the rate could be said, with hindsight, to be on the high side.

The consultative panel

Google devoted less attention to the consultative panel but pointed to its size and to the fact that two members, Mr Kaye and Ms Prevezer KC, charge £1,000 an hour.

The Tribunal recognised that consultative panel members may be regarded as outside professional advisers, so the public sector comparison does not apply to them in the same way. It nevertheless described the rates as extremely high and said ongoing scrutiny of whether the panel and its members are being used appropriately and economically is important. According to the judgment, not every matter needs the involvement of all of them, and use of the more expensive members should be carefully constrained to matters where they, and only they, are specifically necessary.

The panel called this a significant and specific concern, required careful records of what the consultative panel does, how much time each member spends on each matter and why, and said usage will be reviewed on an ongoing basis. It did not treat the issue as a reason to withhold certification.

According to the claim's own website, the consultative panel comprises Sue Prevezer KC, Mark McLaren, Dr Lara Stoimenova and Ann Pope. The judgment records that Mr Kaye became chairperson of that panel under the carriage settlement, a change the published About Us page had not reflected as of the date of certification.

Overlap with the consumer claim

The present claim overlaps significantly with case 1606/7/7/23, Nikki Stopford v Alphabet Inc and others, which concerns substantially the same underlying conduct and was certified on an opt-out basis on 22 November 2024. The essential difference is the class: Stopford claims on behalf of consumers, on a theory that increased advertising costs are passed on to them, while the Brook proceedings claim on behalf of the advertisers who paid those costs directly.

Counsel appeared for the Stopford class representative at the certification hearing but made no submissions, written or oral.

The class representative identified savings of about £2 million to £3 million to be had from cooperation with Stopford. The Tribunal called that very disappointingly small in the context of budgeted costs of up to almost £30 million for this claim alone. Told that the estimates had been made conservatively so as not to raise hopes of savings that did not materialise, the panel said the efficiencies ought to be much greater, and criticised the method: the approach embeds the original budget and then looks for savings, whereas the class representative should instead identify the minimum work necessary for issues unique to this claim and assume everything else is common.

The Tribunal signalled that it intends to make joint case management directions in both sets of proceedings that explicitly limit the resources each side may devote to issues. Other than on pass-on, where the two classes have a direct conflict, the panel said it ought to be possible for there to be only one team of lawyers and, where relevant, one expert on any part of the case.

Dr Brook raised the concern that if joint issues were left to Stopford and that case settled, the advertiser claim would be left stranded. The Tribunal endorsed Google's proposed solution: joint instruction of a single expert and legal team for common issues, on terms that if one class drops away the instruction on behalf of the other continues. The panel described that as a reasonably common arrangement and saw no reason why it would not work.

A joint case management conference will be held to give effect to that structure, and the Tribunal described it as of critical importance, and as an opportunity to exercise its continuing supervisory role over the costs of the proceedings as a whole.

Costs to date

Costs incurred before certification exceed £6.4 million. The Tribunal called this extremely high and said it is not easy to see that it can be justifiable. It accepted that the case is very complicated and that much analytical work has been front-loaded, and it accepted that direct comparison with cases such as Waterside is not meaningful, since the claims are very different. Google had pointed out that pre-certification costs of £2.5 million in Waterside were described by the Tribunal there as "inexplicably high", against more than £6 million here.

The panel said it does not need a direct comparator to have a well-founded concern that the costs may be unjustifiable, noting that the case has not reached disclosure and yet about £2 million has already been spent on experts. As with the other cost issues, it declined to treat this as a reason to withhold certification, saying the important thing will be to control costs going forward alongside the deduplication exercise with Stopford.

The carriage dispute and its price

The judgment devotes a full section to how the claim came to be the only advertiser claim in the field, and the Tribunal treats the settlement that produced that outcome as a live cost question rather than a closed chapter.

A separate proposed collective claim on behalf of advertisers, brought by Mr Roger Kaye KC, was filed on 27 May 2025, after the Brook claim. The overlap produced a carriage dispute over which proposed class representative should be authorised. That dispute was heard on 6 and 7 October 2025. The Tribunal reserved its decision but encouraged the parties to consider a consensual resolution. The chair's remarks at the time, quoted in the certification judgment, included: "We think it is unlikely that we will reach a decision that this is a clean decision all one way."

By letter dated 24 October 2025 the teams told the Tribunal that discussions had taken place and that both sides considered there to be a realistic prospect of agreement. On 10 November 2025 an agreement in principle was reached. In broad terms, the Brook proceedings would continue as the sole set of proposed collective proceedings, the Kaye proceedings would be stayed, and members of the Kaye legal and expert teams would take agreed roles within the Brook proceedings. Mr Kaye would become chairperson of the consultative panel. Revised funding and insurance arrangements would be implemented. By order made on 7 April 2026 and drawn on 8 April 2026, the Tribunal stayed the Kaye proceedings, providing that they may be withdrawn if the Brook proceedings are certified.

The financial consequences drew the Tribunal's attention. Both Mr Kaye's costs and the class representative's costs of the carriage dispute are included within the Total Investment Amount, for which the funder is entitled to be fully reimbursed and which is used to calculate the funder's return as a multiple of that amount, up to a maximum of six times. Those payments to the funder were to rank ahead of the proposed class. According to the judgment, that means the settlement of the carriage dispute may cost the proposed class a substantial amount of money. Separately, Mr Kaye's legal team may be entitled to success fees for their work on the Kaye proceedings.

Google asked the Tribunal to decide three things immediately: that the carriage dispute costs are not recoverable against Google; that the funder is not entitled to any return on them out of a damages award; and that Mr Kaye's legal team should not be entitled to success fees.

The Tribunal ruled on the first, holding it clear that Google ought not to be liable for the costs of a dispute between two other parties in which it had no interest and did not participate. No submission to the contrary was made.

The second and third were deferred to a later stage of the proceedings, with the panel noting they are significant issues that will merit proper consideration then. The class representative's position on the record is that there is no rule that carriage dispute costs can never be recovered, relying on Hunter and Hammond v Amazon [2024] CAT 68, and that Dr Brook has negotiated the former Kaye team's success fees down to match hers for post-consolidation work, with the historical higher fees ranking only after the proposed class.

Google's broader argument was that the episode showed the class representative had not had sufficient regard to the interests of the class, and that certification should be refused on that basis. The Tribunal rejected it. It found that the argument does not take adequate account of the complicated situation at the time, or of the fact that the Tribunal itself had requested the parties to try to settle. The panel accepted Dr Brook's evidence that she considered the claim assembled under her leadership superior to the Kaye claim, that she had objective grounds to think so, and that she was confronted with a situation in which the Kaye claim might prevail on carriage absent settlement.

The panel also recorded a point of confusion. Counsel for the class representative said during oral submissions that there was pressure to reach a resolution because it was thought the Tribunal might criticise the parties' conduct. According to the judgment, the panel did not really understand this, and explained that its earlier indication had been given because the decision was close and some points had fallen away, not as a signal of impending criticism. The Tribunal questioned whether, even if criticism was thought possible, it was necessarily appropriate to make an agreement that increased the burden on the proposed class by a potentially significant amount.

Continuing loss and limitation

Two further procedural issues were resolved in outline and left for later formalisation.

On continuing loss, Google objected to proposed amendments claiming class members' losses with an open-ended end date. The Tribunal did not need to decide the underlying theoretical question because the class representative accepted it will limit relief so as to claim loss only from the issue of the claim form through to a defined moment before trial, which it suggested may be the last case management conference or the pre-trial review. That approach follows Kent v Apple [2026] CAT 1. Only proposed class members who fit the class definition and had a crystallised loss at the date of the final amended claim form will be included on certification.

After the July hearing and while the judgment was in preparation, another panel gave a decision in Kent v Apple [2026] CAT 57 concerning the permissible period for claiming losses, which Google drew to the Tribunal's and the class representative's attention by letters of 27 July 2026. Google reads that decision as permitting losses to be claimed up to the date of judgment but not beyond. The panel concluded nothing relevant had changed for its purposes.

On limitation, Google pleaded that aspects of the claims are time barred. The position differs between the standalone and follow-on elements: the Relevant Period is reduced by eight years for the standalone aspects and by three and a half years for the follow-on claim, as the class representative has accepted, subject to a point about claims governed by Scots law. The class representative argues that under Scots law prescription does not begin to run until the offending conduct has ended, meaning there is no period in respect of which no class member has a valid claim. Google disputes that analysis and the parties agreed it should be argued at the next case management conference.

The consequence is material to the headline figure. According to the judgment, it is clear that the total claim value is likely to be reduced appreciably from the position at the time of the claim form and the class representative's initial evidence, though the approach depends on the Scots law point. The Tribunal also noted an uncertainty over the degree of pass-on, with the class representative applying what is described as a conservative 50 per cent assumption, and said it had borne the likely impact of limitation in mind when conducting the cost-benefit analysis.

The class representative has 14 days from the judgment to record in writing its positions on class definition, continuing loss and limitation as taken at the hearing. New points may not be raised, and Google may comment strictly on whether the letter properly embodies what was said.

Numbers that do not line up

Three figures for the size of the affected population appear across the documents, and they do not agree.

The Tribunal's notice of application, published on the direction of the Tribunal, states that the class representative estimates approximately 250,000 advertisers fall within the proposed class. The claim's public FAQ page repeats that figure, saying approximately 250,000 organisations that bought search ads from Google are potentially affected. The certification judgment, by contrast, refers to a class of the sort of size involved here as being about 880,000 when discussing the practical difficulties of a book-building exercise.

The judgment does not explain the gap, and neither does the notice. One plausible reading is that the earlier figure reflects an estimate made when the claim form was filed in April 2025, before Google's own data work, and that the later number reflects a revised view of the cumulative population of advertisers across a fourteen-year period rather than a point-in-time count. That reading is inference, not something either document states.

A separate reference point sits alongside both. According to CMA documents cited when the regulator designated Google with Strategic Market Status, between 200,000 and 300,000 unique entities used Google's search advertising in the UK during 2024. A single-year count in that range is consistent with a much larger cumulative total across 2011 to 2025, since advertisers enter and leave the market continuously. Neither the claim materials nor the judgment reconciles the figures explicitly, and the discrepancy will matter to any distribution exercise, since per-member recovery depends directly on the denominator.

A smaller inconsistency runs through the claimant's own published materials. The FAQ page describes Dr Brook as an Associate Professor of Competition Law and Policy at the School of Law, University of Leeds. The About Us page and the Tribunal's notice both describe her as an Associate Professor of Competition Law at the University of Oxford, with the notice adding that she is a Fellow and Tutor at Pembroke College. The judgment does not address her affiliation.

The UK regulatory backdrop

The claim proceeds against a domestic regulatory position that has hardened over the same period.

The CMA launched its first strategic market status investigation under the Digital Markets, Competition and Consumers Act 2024 on 14 January 2025, examining Google Search. It designated Google with Strategic Market Status on 30 September 2025, finding substantial and entrenched market power in general search services and search advertising. The designation runs for an initial five years and covers Google Search accessed through any medium, including AI Overviews and AI Mode, alongside Google Ads and Search Ads 360 when they provide search advertising.

The economics the regulator recorded track the period covered by the claim. According to the CMA, Google accounts for more than 90 per cent of all general search queries in the UK, and UK search advertising spending nearly doubled from £8 billion in 2019 to £15 billion in 2023. Google's public response, published on 10 October 2025 by Oliver Bethell, its senior director of competition, put Google Search's contribution to UK economic activity at £118 billion in 2023 and warned about the pace and scope of new rules.

The first binding obligation under the designation, the Publisher Conduct Requirement, was imposed on 4 June 2026, with substantive control obligations coming into force on 3 December 2026. That instrument concerns publisher content in generative AI products rather than advertising pricing, but it establishes that the designation carries operative consequences rather than remaining a label.

The Tribunal's own catalogue of Google-related proceedings has grown alongside. Beyond Stopford and the now-stayed Kaye claim, the CMA issued a statement of objections in its separate ad tech investigation on 6 September 2024, and the European Commission's ad tech decision, fining Google 2.95 billion euros, was published in redacted form in January 2026.

What certification changes for advertisers and agencies

The immediate practical effect of an opt-out order is that inclusion is automatic. Any UK-domiciled advertiser that paid for Google search advertising during the Relevant Period, whether it bought directly or through an agency, is a class member unless it takes positive steps to remove itself by a date the Tribunal will set. Class members will be bound by any judgment on the common issues and will not be able to bring individual claims raising the same issues.

Media agencies occupy a distinctive position. They are expressly excluded from the class, since the claim is for the advertisers whose money was spent rather than for the intermediaries who placed the buys. At the same time, the distribution model the Tribunal accepted assumes agency cooperation in identifying and reaching clients whose spend passed through agency accounts. The unanswered question about whether the unique identifiers used when agencies place advertisements with Google map back to individual advertisers is not a technicality; it determines how much of the class can be reached from Google's own records without a manual reconstruction exercise.

The common issues to be determined are set out in the Tribunal's notice: the definition of the relevant markets; whether Google held and continues to hold a dominant position on those markets; whether Google abused and continues to abuse that position, including through the alleged exclusionary conduct; whether any abuse caused class members loss and damage; the quantification of any aggregate award; and the basis, rate and duration of interest.

No money is available now, and there is no guarantee any will become available. The claim must be won at trial or settled before any distribution question becomes live. According to the claim's published materials, the funding is provided by Burford Capital, which has committed up to £40 million payable in tranches, with the class representative also holding an indemnity against adverse costs, meaning class members face no financial exposure. The claim's FAQ estimates the case will take two to three years, which it describes as standard for collective actions.

For advertisers watching several jurisdictions at once, the UK proceedings now sit inside a wider pattern of private claims following public enforcement. In the United States, Keller Postman announced a mass arbitration campaign on 11 May 2026 targeting alleged advertising overcharges on behalf of thousands of US businesses, following law firms recruiting advertiser plaintiffs from October 2025 and a New York court granting partial summary judgment on 27 October 2025 that gave preclusive effect to earlier monopolisation findings. Competing ad tech firms have filed their own follow-on actions, among them OpenX in August 2025, Raptive in October 2025 and Teads on 3 August 2026. Google separately updated its Google Ads terms of service on 1 July 2026, removing the arbitration requirement in unspecified markets.

The distinguishing feature of the UK claim is what it targets. The US ad tech cases concern open-web display intermediation. The Brook proceedings concern the price of advertising on Google's own search results pages, which is the single largest line item in most UK performance marketing budgets. In a LinkedIn post published after the ruling, Or Brook described the claim as the first case of its kind in the UK seeking redress for the direct harm caused to businesses forced to pay heightened prices for advertising space on Google search results pages.

What happens next

The Tribunal granted the collective proceedings order on an opt-out basis, found both the authorisation and eligibility conditions met, and ruled that Google is not liable for the costs of the carriage dispute in any event.

Several matters return to the Tribunal. The class representative must file its formalising letter within 14 days. The Scots law limitation point will be argued at the next case management conference, after which quantum calculations will have to be redone and directions given for that work. A joint case management conference covering both these proceedings and Stopford will address deduplication of legal and expert teams, the use and cost of the consultative panel, and the resource limits the Tribunal has said it firmly intends to impose. The questions of whether the funder may earn a return on carriage dispute costs, and whether the former Kaye team may take success fees on pre-consolidation work, were deferred rather than resolved.

The domicile date, the opt-out deadline and the mechanics of notification all remain to be set. Until then, the practical position for UK advertisers that bought Google search ads between January 2011 and April 2025 is that they are already in.

Timeline

Summary

Who: Or Brook Class Representative Limited, a special purpose vehicle whose sole director and member is Dr Or Brook, acting as class representative for UK-domiciled advertisers, against Alphabet Inc, Google LLC, Google Ireland Limited, Google UK Limited, Google Asia Pacific Pte Ltd and Google Commerce Limited. The Tribunal panel comprised Mr Justice Meade as chair, John Davies and Robert Herga. Kieron Beal KC and Ali Al-Karim appeared for the class representative, instructed by Geradin Partners Limited and KP Law Limited; Sarah Abram KC, Alfred Artley and Edmund Eustace appeared for Google, instructed by Simmons & Simmons LLP.

What: The Competition Appeal Tribunal granted a collective proceedings order on an opt-out basis in case 1720/7/7/25, allowing an aggregate damages claim estimated by the class representative's expert economist at around £5 billion to proceed to trial. The Tribunal also ruled that Google is not liable for the costs of the earlier carriage dispute, deferred questions about the funder's return on those costs and about success fees, and required the class representative to formalise its positions on class definition, continuing loss and limitation within 14 days.

When: The judgment is dated 5 August 2026, following a hearing on 8 July 2026 and an application dated 15 April 2025. The class covers advertising purchased between 1 January 2011 and 15 April 2025.

Where: The Competition Appeal Tribunal, Salisbury Square House, 8 Salisbury Square, London EC4Y 8AP, sitting as a Tribunal in England and Wales. The class is limited to advertisers domiciled in the United Kingdom on a domicile date still to be determined.

Why: The claim alleges that Google abused dominant positions across licensable smart mobile operating systems, Android app stores, general search services and search engine management services, foreclosing competition in general search and search advertising and allowing it to charge supra-competitive prices for search ads while reducing their value. The Tribunal found the claims are not weak, that an opt-in exercise across a class of about 880,000 would be slow, would lose a significant portion of the aggregate claim and would delay the parallel consumer proceedings, and that direct distribution using Google's own records could reach far more than the 1 per cent take-up Google projected.