Our website said our app could not do what it does

How a marketing page ends up wrong about its own product

The copy was written from the App Store description, the website's existing pages, and memory. All three were out of date relative to the app, and each one reinforced the others, so the error looked well-supported.

This is the failure mode of writing about your own product from secondary sources. Store listings are updated at submission time, marketing pages are updated when someone remembers, and the app changes every sprint. The gap widens quietly.

What was actually wrong

Two claims, repeated across roughly a dozen places. First, that the app does not produce expense reports — it produces simple expense, tax and client reimbursement reports as Excel, CSV, PDF summary or PDF with receipt images. Second, that scanning and organising are free — the app is free to use, but the AI extraction is metered.

The second one is the more embarrassing, because it appeared in the "Is it free?" answer on the product page. That is the single sentence a visitor is most likely to read before deciding.

Both were corrected the same day, along with a product page that described only capture and organisation and omitted the reports feature entirely — which is the app's main reason to exist.

The cost was not just accuracy

One of our guides targets small-business expense reports. For as long as the error stood, that page argued against our own product on the exact query it was written to win. We were paying for content that talked people out of the thing it was meant to introduce them to.

The same mistake also kept a whole feature out of every listing and directory description we had drafted, which would have compounded the error across other people's sites rather than only our own.

The rule we adopted

Capability and pricing copy now comes from the source, not the listing. Where the paywall sits is a line of code or a database rule, and it is quicker to read that than to argue about what we remember.

For our receipt app the answer was a single SQL function defining what a free user can do. Five minutes of reading replaced a year of accumulated assumption.

We also now check release state through Apple's lookup endpoint rather than trusting our own site — which immediately turned up a second problem: an app that had launched the day before while the website still called it "in development".

Why write this down

Our guides make a point of saying where our app is the wrong choice, because pages that only sell are not worth citing. That standard is meaningless if we quietly correct our mistakes and publish only the polished result.

If you write about your own product, the useful question is not whether your copy is persuasive. It is when anyone last checked that it was true.