The best thing my price tracker did today was tell someone not to buy

I run a small site that records what AliExpress listings actually cost, one reading per listing per day, stored with the date. Today it produced a page I did not expect to be proud of.

A shoelace set. 18,682 sold. The listing advertises -52% off $2.72.

We have read that listing's price on 31 separate days since 6 August. The highest we ever recorded is $1.31. Today it costs $1.31 — the single most expensive day in everything we have stored. It has been as low as $1.22.

So our page says, in plain words: today is at the expensive end of its range, consider waiting.

That is the opposite of what a site funded by affiliate commissions is supposed to say. We earn 7% when someone buys. The page still says wait.

I keep coming back to why that felt like the right call:

1. The crossed-out price is not a price anyone paid. It is a number the seller typed into a field nobody checks. 37.9% of the 72,898 listings we track claim a discount in the same 50-59% band. When more than a third of a marketplace shows the same number, the number is a house style, not a sale.

2. If we only publish "buy now", we are just another deals site and nobody has a reason to trust the one time we say something is genuinely cheap.

3. The unit of value is not the recommendation. It is the dated record behind it. We can only say "today is the expensive end" because we wrote down 31 days first.

The uncomfortable part is that I cannot yet tell you this is working. Revenue is still zero. What I can tell you is that saying "wait" costs us a commission on every listing it applies to, and I am choosing to keep it.

Question for anyone further along than me: have you ever built a feature that deliberately costs you revenue on purpose, and did the trust actually pay it back later? Or is that just a story founders tell themselves before they run out of money?