When is the best time to buy on Amazon?

Compare a price with itself

Every useful judgement about a price is a comparison with the same item at the same shop over time. The median of the last 90 days is the honest centre of that comparison: half the time it cost more, half the time less.

Two numbers around it decide the rest. The 90-day low tells you what is achievable. The gap between the low and the median tells you what waiting is worth: if they are $6 apart, waiting for the floor is not a plan, it is a hobby.

How long you have to decide

The other half of the question is never asked by a price tracker: not "is this cheap" but "how long will it stay cheap". A drop that lasts eleven hours and a drop that lasts three weeks are the same shape on a chart and call for opposite behaviour — buy now, or finish thinking about it first.

It is answerable, because the record here is written only when a price moves: the gap between one observation and the next is exactly how long the earlier price stood. So each product page reports how long past drops at that retailer actually lasted, as a median over completed ones, and how long the current one has been running.

Read it as a deadline, not a countdown. A drop whose past versions lasted a fortnight is one you can sleep on. A drop whose past versions lasted a day is one where waiting costs you the price. The stretch still in progress is shown separately and deliberately kept out of that median — counting an unfinished drop as if it had ended would make every drop look shorter than it is, which is the one direction a number like this must never be wrong in.

What our data cannot tell you

A price history tells you what happened, not what will. We do not forecast, and we will not print a prediction dressed as a fact — nobody can see a retailer’s margin plan or their inventory position.

It is also silent on stock, on return policy, and on whether a cheaper seller is one you should hand a card number to. Those decide purchases too, and no chart contains them.