This week we deleted five of our own apps from the App Store. Bike Story, Moto Story, RV Story, Kitty Story, Paws Story — gone. Not because they failed. Because they should never have been separate apps in the first place.

What’s left is two: Vehicle Story, which now covers cars, motorcycles, and RVs in one log, and Pet Story, which covers dogs and cats. This is the story of why we split them up, why that was a mistake, and what it cost to undo.

The instinct that felt right

When you read about App Store optimization, the advice points one way: be specific. An app called Bike Story ranks for “bike maintenance.” An app called RV Story owns “RV.” A tight, literal name with the keyword in it beats a broad name every time. So if you have a maintenance engine and four kinds of vehicle, the “obvious” move is four apps, each laser-aimed at its niche.

We took that seriously, because on paper it’s correct. Search is literal. A person typing “motorcycle service log” is more likely to tap Moto Story than a general Vehicle Story. We split the pet apps the same way — a dog person searches “dog,” a cat person searches “cat.” Seven listings, each pointed at exactly what someone might type.

What the niche math leaves out (this is the real post)

The ASO logic is real. It’s just dwarfed by everything it ignores. Here’s what we actually felt, once the apps were live:

  • Every app is a tax, forever. A live listing isn’t a one-time push. It’s its own metadata, its own screenshots, its own review cycle, its own privacy labels and age rating, its own support and privacy pages, its own update each time the shared code changes. Seven listings, two people. We weren’t running seven products; we were paying upkeep on seven copies of one product.
  • They were literally the same app. Car, Bike, Moto, and RV Story ran the same engine pointed at different datasets — we’d already written about that pattern in One Dataset, Two Products. The split was cosmetic. Every bug fix and feature had to be shipped four times to four review queues to reach what was, underneath, one piece of software.
  • Ratings don’t merge — they fragment. Social proof is the one thing a new app can’t fake, and splitting guarantees you start from zero in four places at once. Four apps with two reviews each look abandoned. One app with eight looks alive. Same eight people.
  • Real owners aren’t niches. A person with a car and a motorcycle needed two of our apps, two purchases, two logs. Households have a dog and a cat. The niche model made us charge our most engaged users twice for the same thing, and asked them to check two apps to see one picture.
  • We competed with ourselves. The cross-sell we imagined — “you have Car Story, try Moto Story” — never happened. What happened instead is that our own apps split the same demand and the same attention between them.

None of this showed up in the keyword tool. All of it showed up the moment we had to run the portfolio.

What merging actually cost

The App Store has no “merge apps” button, so undoing this was its own small grind, and worth being honest about:

  • You pick a survivor and fold the rest into it by hand. We kept the original Car Story listing — same App ID — renamed it Vehicle Story, and widened its data to cover bikes and RVs. Keeping the App ID matters: it carries the ratings and install history we’d earned. The other four got deleted outright.
  • The deleted listings took their keywords with them. Whatever “RV Story” had earned in search is just gone; Vehicle Story has to re-earn it under one roof.
  • Everything downstream had to follow. New names, new icons, rewritten descriptions and screenshots, and a pile of redirects so the old links and our own site pages don’t break. A rename is never just a rename.

We did the pet apps the same way: Pet Story is the old Pup Story listing, widened to cats, with the standalone cat and “paws” apps removed and their care schedules folded into one pet hub.

The lesson

Niche-per-app is optimization theory. Portfolio upkeep is the physics. For a two-person studio the scarce resource was never listings — it was attention, and every extra app spends it whether or not anyone installs the thing.

So the rule we’d give our past selves: build the engine once, then ship the fewest apps that cover the most ground. One vehicle app, not four. One pet app, not three. Let the single listing get good — good ratings, good screenshots, a real foothold in search — instead of spreading thin across niches that were only ever different rows in the same database.

It’s the same principle as the free web tools and the screenshot factory: do the structural work once, and don’t multiply the things you have to keep alive. We had to ship the wrong way to feel why.

More soon. We’ll keep these honest.