CodeLabProsCodeLabPros — homeMenu

Writing

The complaint you hear costs a morning. The one you do not costs the contract.

Complaints arrive ranked by how annoying something is, never by what it costs. We counted 274,676 posts from business owners to see how far apart those two orders run.

September 2026 · 3 min read

We pointed a research engine we built at 274,676 posts from business owners talking about their own companies, and counted what they complain about.

Doing things by hand: 6,714 posts. One customer being able to see another customer's data: 189.

Thirty-six to one, and it is backwards.

Posts about doing things by hand

6,714  ·  each mark is roughly 190 posts

Posts about one customer seeing another’s data

189

The volume is inverted relative to the consequence. Manual work is irritating every day and survivable forever; the other one happens once.

Manual work is irritating every single day and survivable forever. A business can run for a decade with somebody re-keying orders every morning, and many do. The other one happens once. It costs a contract, a reference, sometimes a certification, and occasionally the company.

Why the count comes out this way

Owners report the pain they feel. Data entry is felt continuously and by the person writing the post. A boundary failure is felt once, by somebody else, usually after it is too late to prevent. The people it happens to are not, on the whole, posting about it in public afterwards.

So the corpus is not a ranking of what matters. It is a ranking of what is noticeable, and the difference between those two is worth knowing about if you build software for a living.

What this means for what we get asked to buildNobody has ever arrived and asked us to make sure one customer cannot see another's data. It does not appear in briefs, it does not demo, and no customer has chosen a product because of it. It is also the thing that has to be true before any of the features matter, which is why we build it whether or not it was requested. We would sooner say that here than discover we disagreed about it in month three.

The count is a keyword measurement over a preserved corpus, not a semantic one. The exact patterns and the script are re-runnable, and we would rather hand those over than defend a number by repeating it.

Read nextWhat business owners actually describe when they describe a problemThe most common thing in 274,676 posts was not a technology request. It was a person doing something a computer should be doing. Almost nobody calls that a software problem.All writing