The new system that forgot the old one's job

A residential tower in Doha used to manage visitors with a paper book at the front desk. It was clumsy. The handwriting was illegible, the guard had to ring each flat, and nobody could find anything a week later.

So the building replaced it with an app. Residents must now register every visitor in advance, with a full name, an identity number and a time window. The visitor receives a code by message, which often arrives late. The guard, unable to read the code from the visitor's phone, rings the flat anyway.

It costs more, does less, and everyone involved now has one more password.

This is a common way for systems to change. The old one has grown awkward through years of patches, so somebody buys a replacement. But the question they ask is how to make the existing compromise neater, not why the thing existed in the first place. The new system faithfully digitises every flaw of the old one and adds a few of its own.

Start from the beginning and the job is small. The building wanted to know who was inside, and residents wanted their guests to get upstairs quickly. That is all. A resident could share one link and be done. A regular visitor could simply be recognised. The cleaner who comes every Tuesday need not be registered afresh every week of the year.

Every good design question comes back to two things: who is this for, and what is it for. Replacements tend to skip both, because the old system already answered them, badly, and the new one inherits the answers without noticing.

The moment something is being replaced is the rare moment it can be rethought. Most of the time we spend it polishing.