Blog

Why founders and CFOs should treat platform dependency as a financial risk, not a tech problem.

A mobile studio can route millions a year through a business partner it has never spoken to, cannot telephone, and whose relationships are largely managed by AI. Ask most founders for the name of one human being on the other side and you will often get silence. In almost any other market that sentence would stop a credit committee cold. In mobile, it is simply just another day at the office.

We have a name for this in finance, and it is not “an app review issue.” It is counterparty risk, and its cousin, concentration risk. Both come with decades of vocabulary, limits and governance, because businesses have a long history of being quietly killed by a dependency they never priced. Mobile has been oddly exempt from that discipline — not because the risk is smaller, but because the industry learned to think like a growth-equity market, not a credit one. We model ROAS, payback and LTV to the decimal, and rarely ask the duller question a lender might ask first: what is the single relationship this whole cashflow depends on, and what happens if it breaks?

A word in the platforms’ defence – Apple, Google and Meta operate at a scale that is hard to comprehend, and at that scale they need systems, policies and automated enforcement – there is no realistic alternative. When someone is genuinely defrauding a platform, gaming its store, or running abusive creative, they should be shut down, and fast. This piece is not a complaint about any of that, but about the thousands of legitimate developers, trying to operate honestly and grow, who are subject to the same blunt machinery and, too often, treated as though they were the fraudsters with an automatic presumption of guilt before then having to prove their innocence.

That machinery shows up in two quite different ways, and it is worth considering each of them separately.

The distribution platforms: the innocent foul

On Apple and Google, the danger is usually an honest mistake. A policy you did not know had changed, a build that trips a rule, a payment flow that steps over a line, a piece of housekeeping that lapses. The consequence is then wildly out of proportion to the error.

I have never forgotten one example. A developer with monetisation running at around $20,000 a day had it switched off abruptly – not for anything they did to the app, but because they forgot to renew a $95 annual developer agreement. Ninety-five dollars against twenty thousand a day, gone because a routine renewal slipped (someone had left and reminder emails had slipped between the cracks). The gap between a working business and a dead one can be that small, that self-inflicted, and that avoidable. And when it happens, the only route back is the same opaque review queue as everyone else. An app can sit in limbo, or monetisation stay dark, for weeks.

The ad platforms: the automated ban

On Meta and the other ad networks, the danger wears a different face. Here the exposure is your acquisition engine, and it can be switched off in an afternoon because an AI classifier decided a creative sat outside policy. No human probably even looked at it. The appeal is answered by a template. And while the account is down, your growth simply stops. Ad spend, installs, and the cohorts your model was counting on, all frozen.

The recent Freecash episode – a large rewarded and offerwall player, reported as a $500M-plus network, abruptly deplatformed – is the warning shot. It shows that a single enforcement decision can take out an entire channel, and that the damage does not stay put. It propagates: disrupted UA changes the forecast, a changed forecast changes cash planning, and that reaches runway, hiring, debt capacity and valuation. Concentration is the transmission mechanism.

Enterprise money, consumer-grade support

The two faces share a root cause. In most B2B markets, even modest scale buys you relationship depth – named contacts, real escalation, someone with the authority and context to fix a problem before it becomes fatal. In the platform economy that bargain often does not hold, and the trend is running the wrong way. As the platforms have automated, the human relationship layer has been the first thing cut: partner and support teams thinned by layoffs, the gap backfilled with bots and automated enforcement. The result is sub-consumer support – faster, more opaque decisions, reviewed by fewer people, appealed into a queue that replies in boilerplate. I have yet to meet a developer who says it is getting better.

The platforms answer this in the language of compliance: we must enforce our terms consistently, at scale. Fair enough. But the overwhelming majority of developers are not lawyers parsing the latest revision of a policy written to protect the platform, not to guide them. The canyon between “we enforce our policies” and “we will help a good-faith operator fix an honest mistake” is exactly where livelihoods disappear.

Why it bites harder when growth is financed

All of this matters more when the growth is funded. Capital deployed into user acquisition – from cashflow, receivables, cohort financing or private credit – assumes the relationship holds: the app stays live, the ad account keeps running, the payouts arrive. That continuity is baked into the financing and almost never underwritten explicitly.

It collides with how lending works. A financier cannot sit inside your account watching daily compliance; the relationship is governed at arm’s length, through covenants. Covenants are blunt – they cannot tell a failing business from one that made an innocent mistake. So a misread policy or a suspended ad account can stall receivables and strain a covenant through no fault of intent, while the only party who can lift the block is a support queue. None of this is a reason for capital to retreat; good businesses remain eminently financeable. It is a reason to manage the risk on both sides, and the founders who get ahead of it are easier to back, not harder.

There is no clean way out

The instinct is to look for the exit. Regulators are circling – Epic in the US, competition proceedings in Australia, the DMA and its imitators – and alternative stores and payment rails are more credible than they were. But regulation opens doors while adding complexity, and alternative stores are a diversification tool with real cost, not a like-for-like replacement. For most studios, you cannot leave. So the task is not to pretend you can. It is to price the dependency and manage it.

What to actually do

1. Build a platform risk matrix. Map each platform against what it controls — distribution, revenue, payouts, acquisition, attribution, data — and beside it the events that could trip you, from a lapsed agreement to a rejected creative to an account suspension. Then size the concentration: what share of revenue, installs and cash each platform carries. You cannot manage an exposure you have never measured.
2. Make it someone’s job. Platform risk that sits between growth, product, legal and whoever happens to know someone at Meta is owned by no one. Give a senior person accountability for renewals, contacts, escalation paths and a clean record of every approval, warning and appeal.
3. Treat policy changes as a live compliance function. When a platform ships an update, someone should read it and test your app, creatives and monetisation against it within days — not discover the gap when a review bounces or an account is flagged. The terms change constantly; your review of them should not be annual.
4. Stress-test a shutdown in the forecast. Model a 14- or 30-day account suspension or a monetisation switch-off: does it trip a covenant, miss payroll, or force an emergency raise? The point is not to predict the event, but to know your sensitivity to it while there is still time to act. That is what makes this a finance question, not an IT one.

The bottom line

Something can always go wrong. The better question is whether, when it does, you understand the financial impact and have any realistic way to respond. For too long the industry’s answer has been to submit an appeal and wait – and to accept that a partner can take enterprise economics from your business while offering a chatbot when it turns you off.

The platforms are right to police fraud at scale. But the honest operators caught in the same net deserve better than boilerplate, and the studios that treat the platform relationship as something to be owned, modelled and cushioned rather than assumed will be steadier, and more financeable, for it. These relationships carry real livelihoods. They are too important, and too fragile, to be left to a dashboard and a bot.

Apply To Join