The first thing anyone notices about the App Store is the thirty percent, and it's the wrong number to worry about. Thirty percent is a price. It's written down, and you can work out in advance whether the other seventy adds up to a business. What isn't written down anywhere is the rest of the deal: the shelf, the search results, the payment relationship and the right to be installed at all belong to someone else, and they're lent to you on terms that can be rewritten without asking you.
None of this is a complaint. The alternative, building your own distribution from nothing, costs so much more that almost nobody starting out should choose it. It's just worth being honest that what looks like owning a product is often renting a channel, and that the rent isn't really charged as a percentage. It's charged as the chance that the terms change.
What dependence actually costs
Suppose in any given year there is some chance that the channel you depend on does something that materially hurts: a ranking change, a policy that outlaws a feature you built on, a review process that suddenly reads your category differently, a competing first-party app shipped in the operating system. Call that probability \(q\). The chance of getting through \(t\) years untouched is:
A small yearly risk stops being small once you run it for a decade.
The numbers are unkind. At a ten percent chance each year, ten years of depending on a channel leaves you roughly a thirty-five percent chance of coming through untouched. At twenty percent, which is realistic for anything sitting close to what the platform wants to do itself, it's about eleven. Nothing has to actually go wrong in any particular year. It only has to be possible every year.
Two things follow. First, this risk grows with time, so it changes what kind of product is worth building, not just how you market it. A product that needs ten years to pay off is a very different bet from the same product that pays back in months. Second, \(q\) isn't fixed. It goes up the closer you sit to what the platform wants to do itself, and down the more boring and self-contained your product is. Being uninteresting to your landlord is a real form of insurance.
You also pay in permission
The part that costs most is rarely the money. It's that you have to keep asking permission, and how long the asking takes isn't up to you. A release can sit in review while your launch date goes past. A feature can be approved for two years and then not be. The right response isn't outrage, it's a list: know which parts of your product exist at someone else's discretion, and don't let that list grow without noticing.
So we try to keep a few things ours whatever any platform decides. We keep the customer relationship, and a way to reach people that doesn't go through a store. We keep the domain, so an address we've published for years keeps working. We keep the data in a form that can be moved. A product built to run at almost no cost per user is easier to hold onto this way, because less of it was living on someone else's machine to start with. None of that grows the business on its own. What it does mean is that a bad month is just a bad month rather than the end.
What we actually think
Borrowed distribution is worth borrowing. It's the reason one person can put something in front of millions of people without a sales team, and writing it off as a trap is the kind of purity that gets you no users at all. The mistake isn't using the channel. The mistake is building as though you own it rather than rent it, and then being surprised when the terms change.
In practice that means a few dull rules. Assume the channel will change at least once in any ten years, and check that the product still makes sense if it does. Don't build anything central on a permission that a policy grants and a policy can take away. Keep the direct relationship with customers, even when it's smaller and slower than the borrowed one, because it's the part that survives.
The way to think about it is that borrowed distribution buys you time, and time is exactly what you should spend on building something you own. Companies that forget the second half of that sentence do very well for a while, and then find that everything they built was standing on ground they rented by the year.