The smallest thing we ever shipped has 20,000 users
In 2017 we published a tiny Android app that did one thing: write an amount in words. It has outlasted and out-reached projects ten times its size — and it taught us something about what people actually adopt.
Every cheque written in Algeria carries the same small obstacle. You fill in the figure, and then you have to write it again in words. Same for invoices, contracts and most official paperwork. The number in words is the one that counts legally, and it is the one people get wrong.
In 2017 we shipped an Android app that did nothing except that. Type a figure, get the amount correctly written out. No account, no dashboard, no upsell. It is still running today, it now works on Android, iOS, web and Windows, and more than 20,000 people use it.
Why writing a number out is harder than it looks
The problem sounds trivial until you write the rules down. French is full of traps: seventy is soixante-dix, eighty is quatre-vingts — with an s, unless another number follows it, in which case the s disappears. Cent follows the same rule. Mille never takes one. Get any of these wrong on a cheque and the bank can refuse it.
Arabic is harder again. Numbers agree in gender with what they count, and the agreement inverts in the three-to-ten range. There are dual forms. The case ending shifts with the grammatical position. A rule engine that handles French correctly will still produce wrong Arabic.
English is the easy one, which is exactly why it is a poor test of whether your implementation is right.
One job, done correctly
The whole product is the correctness. There is no clever interface to design around it and no feature roadmap that makes it more valuable — an amount is either written correctly or it is not.
That constraint is what made it spread. Nobody recommended it because it was impressive. Accountants told other accountants, offices told other offices, and it moved through exactly the people who hit the problem several times a day.
What it taught us
We build large systems — points of sale running across dozens of branches, platforms handling national-scale volume. This app is the smallest thing we have ever published, and by user count it reaches more people than most of them.
The lesson we kept: adoption follows a real, repeated, boring problem far more reliably than it follows scope. A tool that removes thirty seconds of friction from a task someone performs twenty times a day earns its place without any persuasion.
It also stayed useful because we kept it alive. Starting Android-only was right for 2017; people then asked for it at their desk, on iPhone, in a browser. Following users onto the platforms they were already on cost far less than building something new would have.