Why the EU’s Digital Markets Act fails in the App Store context: network effects handily beat non-qualified FRAND

The DMA doesn’t lack buzzwords. “Contestable and fair markets in the digital sector” sounds great, as does “fair, reasonable, and non-discriminatory [FRAND] general conditions of access for business users to … software application stores.”

It’s just that buzzwords, in and of themselves, don’t decide cases. And Apple will go to court to defend its terms.

The DMA is what happens when politicians and public servants figure out a problem but lack the competence or (as the result of compromising) the collective will to put a workable solution in place. In the end it’s just a recipe for protracted litigation, not a game-changer in the marketplace.

The mobile app store situation is a colossal market failure. Software makers and service providers in an increasingly digital economy depend on two platforms—the Apple-Google duopoly—to reach customers. The answer is not to tell everyone to just develop for other platforms. You need to be on iOS and on Android, or in most contexts you just don’t exist.

Even governments experienced the extent to which this has gotten out of hand when they had to negotiate with Apple and Google what their national COVID tracking apps were allowed to do. The UK government had to capitulate with respect to a feature that could have prevented some infections and death on a voluntary basis.

Not only is it a market failure, but also a result of underenforcement. In the U.S. they weren’t prepared, for a long time, to go after national icons. The European Commission could at least have tried to tackle the problem through the enforcement of conventional antitrust law. It didn’t. The Spotify case, which is expected to end (after too many years) in a fairly useless decision this month, was a wasted opportunity. If the EC had taken swift and decisive action against Apple in that narrow context (where an app maker had to compete with Apple), and if it had defended the most important elements at least in a first round of appeal, it could then have moved on to pick another case, such as Epic Games’ 2021 complaint but preferably a year or two before. Instead, they were dragging their feet, moving in circles and not changing anything.

To be fair on the enforcers, it wouldn’t have been trivial under conventional antitrust law to satisfy the exacting requirements for an abuse of a dominant market position, starting with a single-brand market definition (iOS app distribution). Even a minor procedural mistake might have doomed the case. And it’s generally hard to tackle specific economic terms (app store commissions).

The “Coalition Against the 30%”: App Store complainants’ strategy is part of the problem

Complainants made a huge mistake because they were self-serving, not sufficiently strategic. They focused exclusively on the 30% cut. Their story came down to more money for them, less for Apple.

They should have additionally identified or created a situation where something objectively beneficial to end users (and ideally, for political and psychological reasons, small businesses or gig economy workers) would have been blocked by Apple (and possibly also Google) based on their partly very unreasonable app review guidelines. In that case, there would have been an actual exclusion (as opposed to just a margin) argument. Attacking the app review monopoly would have shown to regulators and courts that you need alternative app stores to compete with each other: you can’t force Apple to carry in their own store what they don’t like, but there must be another way of reaching Apple’s customers, which account for a very high percentage of the world’s richtest billion or two billion people.

The Coalition for App Fairness has consistently just been a “Coalition Against the 30%” (the organization as well as its key members). They are now standing before the ruins caused by their shortsightedness and initial impatience. Spotify’s EU complaint: a failure as we’ll see this month. Epic’s U.S. case against Apple: a failure after the exhaustion of all appeals, apart from an anti-anti-steering injunction under California Unfair Competition (not antitrust!) Law that games fray believes has the potential to cost Apple significant amounts of money (January 19, 2024 games fray article). Epic’s late 2023 trial win over Google in the U.S. is by far the best result, but it’s not final: Google’s chances on appeal (such as with respect to market definition) should not be underestimated (February 24, 2024 games fray article). Remedies have yet to be defined, and won’t enter into force in the near term.

There’s a high risk of them repeating, instead of recognizing, that mistake now. They’ll try to desperately enforce the current DMA, despite the easily predictable failure of that effort. In theory, it would be the fastest path; in practice, it’s going to be a waste of time, money and energy. What they would have to do is talk to the policy makers at the Commission, to MEPs and to the governments of EU member states about the insufficiency of the DMA, and push for change as soon as possible. They can try to pressure the Commission’s enforcers, but one can’t blame them for the limitations and deficiencies of the law. Apple’s new EU app rules lay bare the shortcomings of the DMA, long before the review clause kicks in. The Commission’s enforcers should be honest and tell everyone there’s not much they can do here. They can do something, or at least try something, but with the DMA they can’t bring about competition in iOS app distribution beyond adult content, crypto/NFT apps, and apps for smokers and vapers.

The DMA is called ex ante (“beforehand”) legislation but it’s simply too little (unless amended), too late

To understand the fundamental error that has led to this situation where the DMA enters into force but nothing significant changes in the mobile app market, one must take a step back.

The idea behind the DMA and similar bills in other jurisdictions was that traditional antitrust enforcement is too slow and especially too cumbersome to deal with the gatekeepers who control digital platforms. A regulator (or private plaintiff) has to prove in court that a dominant market position has been abused, which involves substantial hurdles to prove a single-brand aftermarket (without which there often is no case) as well as actual anticompetitive effects, and then has to defend the proportionality of any remedies. So the idea was to take a shortcut and impose remedy-like obligations just based on a gatekeeper position, regardless of wrongdoing.

An unsuitable term was coined for this: ex ante (“beforehand”) regulation, as opposed to ex post regulatory action when an abuse has occurred.

The term doesn’t actually make sense. The more appropriate way to describe it would have been “regulation on gatekeeper obligations.” Not only is ex ante a euphemism when there actually would be a case for past abuse (even if the European Commission never made a serious attempt to establish it) but the bigger problem is this:

When a market failure is dramatic and already extremely hard to repair, ex ante misleads people to think that the measure is in time. In reality, it comes too late unless there is a clear and workable plan for how to cure the market failure. The last part is what’s missing with respect to the DMA and app stores.

The question to ask (and which the DMA doesn’t satisfactorily answer) is: what would it take now, more than 15 years and millions of apps later (counting from the launch of Apple’s App Store and the Google Play Store for Android), to make those mobile app distribution markets competitive?

The app store part of the DMA is so weak that it would have been insufficient even if those rules had been in place and vigorously enforced since 2008 (let’s simplify this and ignore the fact that the app stores wouldn’t even have met the DMA’s gatekeeper thresholds back then). It would simply have been too weak to overcome the Power of Default: the fact that users will normally use the default app store. The DMA allows choosing a different default app store, a remedy that worked for browsers where people were interested in changing the default browser, but which won’t work for app stores that are not going to be able to compete seriously with the default App Store.

But by now, the network effects are way too strong. The DMA does mention network effects six times, even in Recital 2:

” Other such characteristics of core platform services are very strong network effects …”

As stated at the outset of this article, the DMA doesn’t lack buzzwords. It lacks reliably effective solutions.

At the height of a major eurozone sovereign debt crisis, then-European Central Bank president Mario Draghi was asked how far the ECB would go to stabilize the common currency. His answer was: “Whatever it takes.” Since then (2012), no major hedge fund has seriously tried to place a bet on a eurozone country’s bankruptcy. It may happen again, but since then it hasn’t.

The DMA does not say: “Whatever it takes.” It serves weak tea. If we ask ourselves what it would take to make mobile app distribution competitive, there are two radical options and a non-radical but (all going well) effective one:

• The most extreme measure, which games fray is not advocating at this stage and never may, would be a breakup. The EU could say that to operate in its market, Apple and Google must not have their own app stores, but those stores must be managed by totally independent entities, and there’d have to be competition in each field. In that case, the original Apple company that makes the iPhone would deal with the App Store spin-off on an arm’s-length basis, and if, for instance, Epic Games approached the EU with a proposal that is better for the iPhone business than what the formerly Apple-owned App Store company proposes, it would normally get the deal because the iPhone business would optimize its own gain. It is obvious that the United States, despite some politicians there (such as Senator Elizabeth Warren (D-Mass.) having similar ideas), would be furious.
• A similarly extrem measure would be to give the app developer economy leverage over the gatekeepers by exempting them from cartel law if they want to negotiate collectively. In that case, for instance, all major mobile game makers (Tencent, EA, Epic, Activision Blizzard King etc.) could enter into an agreement that unless Apple meets certain demands, they’re all going to stop offering their products on the iPhone. In that case, many gamers might switch to Android. Allowing such a cartel would completely fly in the face of the basic principles of cartel law. There would also be a risk of small developers being the ultimate victims.
• The only solution that can work without a breakup or a group boycott would be a regulatory environment in which new entrants (such as Epic and Microsoft) can make their own app stores and gain substantial traction. The last part is what the DMA simply doesn’t enable. Microsoft and Meta have been clear about it, and Epic is trying to offer a store because they hope to bring Fortnite back to the iPhone (even if only in the EU), but they also consider Apple’s new EU app rules terrible.

Alternative app stores won’t gain substantial traction based on the current DMA because in order to overcome the Power of Default, user habits that go back more than 15 years and especially all those network effects, those alternative app stores would have to provide huge benefits to both end users and app makers. Huge costs savings in cluded.

The more users an app store has, the more app makers will offer their apps, and the more apps it has, the more users will use it. Network effects. You don’t overcome them with a 5% discount. You need a breakthrough. The DMA is not a breakthrough.

The EU is better at identifying problems (and at using buzzwords) than at providing actually workable solutions. The DMA allows alternative app stores, and Apple allows them now as a result of the DMA. But those alternative app stores won’t make much sense for apps that Apple’s App Store is also prepared to carry:

• If the app makers offers one version of the app on Apple’s App Store and an alternative version on a rival store, and if Apple allows that at all (as opposed to just rejecting the submission or kicking the app out of its store), the only reason for which customers will use the alternative app store is that they get a benefit. There must be an incentive, and the only incentive that matters here is price. If Candy Crush gamers knew that they’d pay, say, 20% less for items if they download Candy Crush from Microsoft’s store, then that may, slowly but surely, result in some (even if limited) erosion of the App Store monopoly. But Apple’s terms are designed in such a way that a rival app store won’t be able, without huge losses, to offer users significantly (much less substantially) lower prices on the alternative store (January 28, 2024 games fray article). There would be “competition” in name only, as opposed to a competitive constraint based on substitutive impact.
• The other alternative is exclusive content. Here we have to distinguish between full and partial exclusivity, and between first-party and third-party content. The next subsection will explain why it won’t work.

If there was competition at some point, other aspects such as the design of an app store or discoverability could also play a role. But those aspects would matter only on a level playing field, not if the default App Store has millions of apps and rivals can only offer a small number of relevant apps.

Why exclusive content can’t solve the problem under Apple’s terms

Let’s start with the simplest scenario: full exclusivity and first-party content. Microsoft and (thanks to the acquisition of Activision Blizzard King) Candy Crush.

If Microsoft said that Candy Crush would from now on be available only via an Xbox Store, and did the same for some other compelling first-party content (Minecraft, Call of Duty Mobile, Hearthstone etc.) as well as possibly for some of its non-game apps (Teams, Outlook, Skype etc.) such a store could get traction. And it could then create an opportunity for third parties (we’ll get to that in a moment).

The upside for them would be that they can grow their store and, if the terms allowed it, operate Candy Crush more profitably. The downside would be that they would reach fewer users than on Apple’s App Store. It’s even worse: because of Apple’s EU rules, they would lose users. The version of Candy Crush that would not be distributed on Apple’s App Store anymore would be a new app, and users would have to uninstall the old one, then reinstall the new one from the Xbox Store, for which they’d first have to install that one despite scare screens and other impediments. And they’d have to fragment the user base as it would be only for the EU.

Even if those fragmentation and friction issues were resolved through successful enforcement of the current DMA (which is far from certain), app discovery would remain an issue in two ways:

• The Xbox Store frontend app wouldn’t be found on Apple’s App Store, so users would have to download it from the web.
• Candy Crush would no longer be found on Apple’s App Store, where it is currently easy to find. People search for it there, and people see it when they browse the catalog, such as in the puzzle games category.

It’s hard to estimate a number, but there can be no doubt that there would be a very significant loss of users. A loss of a part of the user base could be offset by higher profitability, but would that really be the case? Not at a closer look. And there’s also the issue that as an app maker you really want to reach a broad audience (mind share) for strategic reasons. Therefore, the break-even point involves long-term considerations going beyond short-term margin vs. volume.

Apple would charge its Core Technology Fee (CTF) of 50 eurocents per install per year to the Xbox Store (for its frontend app) and especially to each app distributed through it, such as Candy Crush in this hypothetical scenario. It’s difficult to find more recent numbers, but it’s unlikely that things have changed much since 2013. Back then, only about 4% of all Candy Crush gamers ever paid a cent (Statista webpage). That means 1 user out of 25 is a paying customer. The number could be higher now if they’ve managed to optimize monetization and if gamers are more prepared to pay, but it could also be that the percentage was higher in 2013 than with the larger but more casual audience they reach now.

Apple’s CTF would be paid for 100% of all users that have the app installed. That number is even higher than the number of actual users, as people would keep Candy Crush on their phone potentially for years (and would keep getting more or less automatic updates) without either deleting or playing it. But even if we ignore that factor, it means we have to multiply the CTF by 25. That means the cost per paying user would be € 12.50 per year.

Operating an app store is relatively cheap if the volume is high, but there are some costs, starting with the development of the frontend app. Store operating costs would further reduce the profitability of making Candy Crush an Xbox Store exclusive.

It just wouldn’t be profitable, not even for a game like Candy Crush.

Partial exclusivity would mean that some content or features, or simply some goodies such as extra moves in Candy Crush, would be available only if you download it from the Xbox Store. It would take some development and quality assurance effort to make that work reliably, and the question is whether Apple would even tolerate it. There wouldn’t be a loss of users as long as discoverability on the App Store is not impaired. But the number of gamers who would install an additional store, delete Candy Crush and reinstall it from another store (which would require many to set up a different type of account to keep their game progress), would be limited. It would be an interesting audience, but a small one. Far from what you need to reach scale for a store. And why wouldn’t you actually want the version on the leading App Store to be the best you can offer?

For third-party content, the situation becomes more difficult. Other game makers would not have a long-term strategic interest in growing the Xbox Store (they would just generally have a strategic interest in competition among app stores). They would have to look at short-term effects. Unless Microsoft pays them for exclusive deals, they’re always going to prefer maximum reach and discoverability.

Also, Apple’s CTF will be even more onerous on apps that are not as profitable as Candy Crush. And the CTF makes it totally prohibitive for apps that don’t have in-app purchasing as Apple is still prepared to distribute such apps for free via its own App Store: the app maker just has to opt for the old (pre-DMA) terms, but alternative app stores can’t offer free distribution as Apple would charge the app maker the CTF.

For a third-party app maker that accepts to pay Apple’s CTF, there’s also the option of distributing via the App Store at the reduced (20%) rate. That leaves even less room to give end users and developers a better deal where app makers pass some of the benefits on to their customers.

DMA enforcement (barring an amendment) won’t tip the scales and Apple won’t make meaningful concessions

The above analysis was based on Apple’s terms, rules and technical setup as announced. Couldn’t those change? Couldn’t Apple come under public and/or regulatory and/or litigation pressure?

After all, they changed their rules for home-screen web apps in the EU, didn’t they? (March 1, 2024 games fray article)

Unfortunately, even the last move concerning web apps is not encouraging. It’s actually an anticompetitive move unless and until they allow browsers using alternative browser engines (as opposed to Apple’s own, which is optimized for the user experience advantage of native apps over web apps) to run home-screen web apps. There’s also a fundamental difference: in the web apps context, Apple had to take a feature away from users. There were people who couldn’t play xCloud (Microsoft’s cloud gaming service, for which there is no native app due to Apple’s past rules and lasting app tax) or use the Starbucks web app as conveniently as before. By contrast, alternative app stores are a new feature and the issues are not visible to end users until there are such stores, and even then they typically won’t understand the economics.

In fact, what Apple did about web apps shows that they aren’t really afraid of the DMA. It was just because of the DMA that they initially wanted to disable web apps in their own Safari browser as well (so as to not expose themselves to allegations of disadvantaging rival browser engines).

Apple has a clear priority: to hold onto the App Store monopoly for as long as possible. They don’t like it, but can live with the DMA creating an opportunity for “adult content” or non-fungible token (crypto) transactions. They may or may not be OK with allowing Epic to bring back Fortnite, which they kicked out because of Epic’s deliberate breach of rules it considered unlawful but which the courts of law upheld. What Apple doesn’t want is that alternative app stores gain traction in the mass market. They don’t want alternative app stores to emerge where end users might go to discover apps: they want that to happen only on their own App Store.

The remainder of this article explains why the current DMA was poorly designed with respect to the economic terms. It’s not impossible that some minor adjustment would result from competition enforcement, but the problem is structural and not going to be solved unless the DMA is modified in response to Apple’s belligerent stance.

FRAND, without more, is just an indication of weak resolve, lack of understanding, or both

All the DMA says is that Apple’s App Store terms must be FRAND. In fact, it doesn’t even say explicitly that the terms Apple offers to the operators of alternative app stores, or to app makers that distribute apps via alternative app stores, must be FRAND. But let’s assume, for now, that the FRAND obligation is comprehensive.

When a legislature or a regulatory agency doesn’t really know what they can do or what they want to do, they’ll say “FRAND” to duck the trickiest and most resource-intensive questions, but it’s not a magic word like “abracadabra”. Unless FRAND is fleshed out, it’s merely an abstract concept. It’s one of those words that make everyone hope that their view will prevail, but often that’s just wrong.

There is only one context in which terms are regularly determined by courts to be (or not to be) FRAND: standard-essential patent (SEP) licenses. Apple as the world’s largest SEP licensee is far more familiar with the practical implications of FRAND than the open app store lobby. In fact, Apple has been fighting over FRAND questions, in and out of court, ever since it negotiated the first cellular SEP licenses for the iPhone a decade and a half ago. Its rivals are, by and large, new to the subject. None of them has been party to a judicial FRAND determination anytime during the last 10 years.

The European Commission doesn’t have any FRAND expertise either. Never in its history has the EC taken enforcement action on the basis of someone allegedly charging more than a FRAND rate for something, which would have required the EC to determine a FRAND rate in the first place. There have been cases where the alleged behavior was somewhat related to FRAND, but in those cases the focus was on structural and behavioral aspects as opposed to the rates themselves.

The three key determinants: target, parameters, comparables

To understand why the DMA favors Apple on FRAND, let’s start with the three key determinants that can be more favorable to one party or the other.

Target

FRAND is not just a point. It’s always a range. If the most logical FRAND rate for a certain item (say, a patent license needed to build 5G-capable smartphones) was $4.50, a court of law usually won’t hold someone liable for charging $5.00. It’s within striking distance. But if $5.00 is still acceptable, why not $5.50? Or $6.00? There will be different views on how much of a deviation is acceptable.

When parties negotiate arbitration agreements, there is almost always a dispute over how to define the target. The right holder wants to get paid as much as possible and will prefer an instruction for the arbitration panel that comes down to saying: “Please determine the highest rate I can charge without acting unfairly, unreasonably or discriminatorily.” The other party would prefer the other end of the spectrum: “Please determine the lowest price I must pay.” Typically, they’ll agree on something else.

The DMA says in its Recital 62 that there “should” be an alternative dispute resolution mechanism (which means arbitration) in the EU. The fact that it’s merely a recital, not a statute, and the auxiliary verb “should” make it a weak rule. The most important juncture, however, is going to be when the EC has to defend a non-compliance fine in court. In that situation, the question is going to be whether Apple overcharges so clearly as to warrant fines (and the heftier the fines, the clearer the overcharging has to be). Courts can be very conservative in such a situation.

The target (whether the middle, the high end or something above the upper end of the FRAND range) is going to be neutral at best and favorable to Apple at worst.

Parameters

The most important deficiency is that the DMA doesn’t limit what Apple may charge for. The statute (Art. 6(12)) and the recital (Recital 62) both say “access“.

Those who wanted to demonopolize mobile app distribution (whether politicians or stakeholders) should never have accepted that approach. They should have told everyone “thanks, but no thanks!” Allowing Apple to charge for access to customers defeats the purpose of the DMA.

Access to customers means Apple can make a variety of arguments, and can particularly point to “the most prevalent rate” (as Epic Games CEO Tim Sweeney truthfully testified in U.S. court) being 30%. Apple can argue on the basis of the value of access to its customer base, charge for its services (such as notarization of apps based on a review of limited scope) and make all sorts of intellectual property arguments-

In order for the DMA to serve its purpose of making those markets contestable and fair, the ability to charge would have had to be limited to an intellectual property license, and even that one would have had to be defined rather restrictively (application programming interfaces).

Recital 62 is not really helpful. The following sentence makes valid points, but it’s meaningless when a FRAND determination has to be made:

“Pricing or other general access conditions should be considered unfair if they lead to an imbalance of rights and obligations imposed on business users or confer an advantage on the gatekeeper which is disproportionate to the service provided by the gatekeeper to business users or lead to a disadvantage for business users in providing the same or similar services as the gatekeeper.”

There is no objective standard for an “imbalance” or for a “disproportionate” advantage. The passage following that sentence then talks about how to arrive at a FRAND rate, but all it provides is a list of obvious comparables:

“The following benchmarks can serve as a yardstick to determine the fairness of general access conditions: prices charged or conditions imposed for the same or similar services by other providers of software application stores; prices charged or conditions imposed by the provider of the software application store for different related or similar services or to different types of end users; prices charged or conditions imposed by the provider of the software application store for the same service in different geographic regions; prices charged or conditions imposed by the provider of the software application store for the same service the gatekeeper provides to itself.”

The above actually favors Apple and takes us straight to the next subsection.

Comparables

Even without that recital (which by definition is not as strong as an actual statute), it would have been clear that any FRAND determination in this context would rely on other market rates. That’s the way it always works, across the various jurisdictions that have set FRAND rates so far.

Basically, courts will either rely largely on one comparable or they will identify a set of comparables they consider relevant and look at a (potentially weighted) average.

If you want to bring a rate down with a FRAND argument, by far the most (and often the only) helpful point of reference is that the same right holder charges others a lot less than it demands from you. That just helped Apple in a recent UK case, though the patent holder appealed. It won’t help anyone against Apple in an App Store context, however. Apple has been very consistent.

The second-best thing, trailing by a distance, is that you can show what others charge for something comparable. One problem with those comparables is that there will always be some way to differentiate offering A from offering B in a technology market.

With different Apple fees being at issue, we have to look at comparables in separate sections now.

Existing comparables

Comparable terms to Apple’s standard App Store commission of 30%

In response to the DMA, Apple offers new terms. But it still offers app makers access to its own App Store on the basis of the old 30% rate (let us ignore for the purposes of this analysis the small business rate of 15%, which applies only to small developers that account for a minuscule portion of App Store revenues).

As mentioned further above, 30% is “the most prevalent rate.” Apple can point to Android, to the three major video game console makers and even to PC stores, where Steam charges 30%, albeit a headline rate that some large players may manage to renegotiate. Microsoft also used to charge that rate, then lowered it. And there’s Epic, but with sworn testimony that Epic’s store generates losses, courts won’t give that rate much weight.

For the avoidance of doubt, games fray does not consider the 30% rate fair and reasonable. But it’s difficult to prove if the focus is on comparables. Judges will be very hesitant to deem such a standard rate to be non-FRAND, and to uphold fines. The comparables would be avoided only if the DMA said that the gatekeeper can’t charge for access to customers. It doesn’t say that, and the use of the term “access” unfortunately suggests that access to customers can indeed be charged for.

Comparable terms to Apple’s Core Technology Fee (CTF) of €0.50 per install per year

The CTF makes the analysis more complex because it comes into play in three different contexts:

• For the installation of alternative app store frontends and of the apps distributed via them, Apple charges the CTF but no commissions.
• App makers who distribute via the App Store and prefer that arrangement over the old terms get a reduced commission (from 30% to 17% (commission) + 3% (payment processing) = 20%, or if you’re in the Small Business Program, from 15% to 10% (commission) + 3% (payment processing) = 13%).
• App makers who want to use alternative payment processing services have to pay the CTF as well. They pay the 17% of 10% commission, but not the 3% payment processing fee.

The CTF doesn’t apply to the first million installs. For small app makers there’s a risk of a sudden spike in download resulting in huge costs. But it’s possible that a number of low-volume app makers will opt for the CTF, and Apple will use that as an argument in any FRAND litigation that the market accepts the CTF.

For the scenarios in which the CTF as well as a commission are charged, Apple explains in a Q&A that it expects to be compensated for a wide range of things:

“Apple’s traditional business model has reflected the value of all the technologies, tools, and resources that make it possible for developers to build and share apps with Apple users. That includes distribution, discovery, and promotion on the App Store; its secure payment processing and In-App Purchase commerce system; a range of developer tools, including our more than 250,000 APIs; the value of Apple’s operating systems and intellectual property; and much more.”

About the CTF, Apple merely says that it “reflects the value of Apple’s proprietary tools and technologies, protected by intellectual property.” But there is no fair and reasonable intellectual property valuation here: it’s just Apple deciding that it wants that fee, period.

There’s also the problem that Apple defines the range of IP-protected technologies very broadly, apparently including iOS as well. But end users pay for iOS already by buying an iPhone. That exhausts Apple’s rights, but Apple essentially seeks to double-dip from developers.

In order to reach a realistic IP valuation, Apple would have to face a situation in which it would have to enforce its rights in court and prove that there even is any valuable IP involved. Courts could not allow Apple to double-dip: they’d focus strictly on what IP the developers use that end users haven’t already licensed.

Due to interoperability privileges, the value of API-related copyrights is more than questionable. Apple said in the litigation with Epic that it holds a couple hundred patents on App Store functionality, but it’s doubtful that alternative app stores would actually infringe any valid Apple patent. Are there patents that app makers would infringe just by writing apps for iOS? Probably not.

The trickier part involves the developer tools. Those are protected by copyright and there may be some valid and relevant patents involved as well. The developer tools are used at design time; they are not needed at run time (when someone uses the app). But Apple would make arguments on that basis, too.

The proposals games fray outlined in its February 20, 2024 article would create a situation in which Apple would have to prove and (in the event of a disagreement on terms) enforce valid IP, which is just what Apple itself has consistently said the owners of standard-essential patents should do. The set of proposals would also address the question of developer tools by forcing Apple to either raise its prices worldwide or leave developers alone with respect to the distribution of apps in the EU. If one wanted to be on an even safer side, an amended DMA could also clarify that no distribution-based charges for developer tools are allowed.

Under the current DMA, there is one obvious FRAND issue with Apple’s CTF: free apps get free distribution only on Apple’s own App Store. Apple has an exception for non-profits and governments, but for many important apps made by companies, €0.50 per install per year is prohibitive. Even for an app like Teams it wouldn’t make sense as most people use it for free.

Meta’s criticism of Apple’s fees as failing to create an opportunity for them didn’t go into details, but it’s quite obvious that the CTF alone makes it prohibitive for Meta to go through any iOS app store than Apple’s: apps like WhatsApp, Facebook and Instagram have huge numbers of installs, but generate relatively little revenue per install per year.

There might be complementary arguments to support a case that tells Apple not to charge per install.

But what else can the Commission or private litigants do about the CTF?

It’s difficult. Exceedingly difficult. The Commission has never in its history engaged in IP valuation.

There was a settlement with Microsoft a long time ago that involved network protocols, and the Commission accepted that Microsoft announced a certain charge. There is no indication anyone ever took a license on those terms, nor of Microsoft actually enforcing its IP in that context. But the Commission never challenged an IP license fee for being allegedly supra-FRAND.

Those who believe that the Commission can do it now are going to see that it’s not going to work. But it could take a long time until the Commission affirmatively rules out dealing with that issue, or indicates a position on FRAND that complainants would consider unhelpful.

There is no comparable to the CTF that would help enforcers.

Google, at least so far, doesn’t charge a CTF equivalent. Alternative app stores on Android can so far operate without any charges by Google.

But legally, Apple’s terms are not proven to be non-FRAND just because the only other comparable player gives something away. Similarly, no one could demand that Nokia gives them a free 5G patent license should some other 5G patent holder opt to open-source its IP.

The only other install fee that is known at this point is the one by Unity, which makes a popular game engine and recently introduced such a fee for the first time. Unity exempts low-volume (up to 1 million installs) and low-revenue apps (Unity Q&A). Depending on what kind of Unity subscription a game maker has and also on the geographic target market, different runtime fees apply (September 22, 2023 Ars Technica article). Unlike Apple’s flat CTF, they’re degressive, ranging from $0.15 at the top to $0.01 at the bottom. Also, Unity offers developers to for a 2.5% gross revenue share instead of install fees, and it doesn’t charge the install fee every year (only the initial engagement results in a charge).

Apple was obviously aware of Unity’s fees. Apple’s CTF wasn’t inspired by Unity’s recent price change, as the idea of charging per download even came up in connection with the Epic Games v. Apple litigation. But Apple has Unity’s example to point to when defending the structure of an install fee.

The EC won’t be able to use Unity’s price matrix as evidence that Apple’s CTF is supra-FRAND. There would actually be a strong argument that Unity’s fee should be higher as end users don’t pay for the huge amount of IP that is distributed as part of Unity-based apps (the entire game engine, which is like a specialized operating system except that for certain features it relies on the underlying operating system). But there’s no basis on which one can claim that Apple’s higher install fees are supra-FRAND just based on a comparison with Unity’s.

Simply put, the FRAND problem is that when it comes to the 30% (or anything derived from it), Apple can argue that it is “the most prevalent rate” (which may be enough in its own right to defend that rate in court), but when it comes to the CTF, there are no data points out there that enable the EC to build a solid FRAND argument.

Theoretically one could imagine a FRAND argument based on the costs related to the IP in question, plus a reasonable profitability. But that would amount to price regulation as Apple can argue that the market value of the IP in question is simply a huge multiple of the related R&D expenses. There is no reason to assume that the EU judiciary will support such an approach.

And, again, Apple may even dispute that the DMA imposes any FRAND licensing obligation that relates to its CTF in connection with alternative app stores.

Could the DMA have set, or could an amended DMA set, specific rates or caps?

The EU had no qualms over disallowing certain roaming charges when people travel within the EU and use their phone in another member state. Some charges were allowed, but with caps. So there is precedent for price-setting legislation in the EU.

The DMA could have said, and an amendment DMA could still say, that certain charges are not allowed, such as per-install fees. It could even say that no charges of any kind must be levied on alternative app stores and the apps distributed through them, and could close the potential loophole involving developer tools. In the end, Apple would still have the opportunity to make a lot of money, but it would have to charge users, not alternative app stores and the app developers using them. Users who pay a lot of money for an iPhone should have the right to get apps from any source they choose without Apple taxing that source beyond what developers also pay in other jurisdictions for using the developer tools.

An alternative would be what games fray has previously proposed and which is simply in line with Apple’s own position on SEPs. In that case, it would be left to the courts, but Apple would likely not get much from alternative app stores and from app makers distributing apps outside the App Store.

Sadly, enforcement of the current DMA won’t solve the problem anytime soon, and even in a realistic best-case scenario, the monopoly won’t be broken because it takes a lot to overcome the Power of Default and network effects. A lot more than the current DMA.

It will take some time before there will be momentum behind the idea of amending the DMA. But games fray isn’t necessarily wrong only because it considers the failure of the DMA 1.0 (at least in the app store context) predictable before others realize it. Others may first have to see a multi-year litigation campaign fail miserably.

Absent a meaningful amendment, the aggregate app distribution volume of all alternative iOS app stores combined won’t even come close to 5% of total app distribution volume in five years’ time. It’s probably not even going to be 1% in ten years’ time (absent an amendment), but that’s a more difficult prediction. Not even close to 5% in five years’ time, however, is a very safe assumption. And that market share is going to be the measure of success or failure for the DMA’s app store part.