You look at something online in the morning. You come back that evening and the price is different. Nothing about the product changed, nobody decided it was worth more, and yet there it is.

This happens constantly, and the explanation is more mundane than most people assume.

Nobody is setting that price by hand

Large online retailers run automated repricing systems. These watch competitors, stock levels, demand signals and conversion rates, and adjust prices without a person in the loop.

As the reference material on dynamic pricing puts it, online retailers in particular adjust prices according to competitors, time, traffic, conversion rates and sales goals, and Amazon changes prices frequently throughout the day.

The systems are largely reacting to each other. One seller moves, competing repricers notice within minutes, and the number on your screen is wherever that loop happened to be when your page loaded.

Grocery chains using electronic shelf labels can push this further, changing a displayed price several times per minute.

The scale is what makes it invisible

This only works because of volume. The Census Bureau put first quarter 2026 e-commerce sales at $326.7 billion, 16.9% of total retail, up 9.8% year over year while retail overall grew 3.9%.

At that scale nobody is approving individual price changes. There is no room for a person in the process.

Three things that actually help

Look more than once. Checking an item across a few days costs nothing and tells you whether the current price is high, low, or typical for that listing. A single look tells you nothing at all.

Check who is selling it. On many listings the buy button belongs to one of several sellers, each with a different price and different shipping. The price you see is that seller's offer rather than the product's price, and other offers on the same listing are frequently cheaper.

Treat urgency cues carefully. Countdown timers and low-stock warnings are frequently generated by the same system that sets the price. Sometimes they reflect real scarcity. Often they reflect a setting.

The sale price question

The one people ask most: are those discounts real?

Sometimes. The honest answer is that a percentage-off claim is measured against a reference price the seller chose, and that reference is not always a price the item genuinely sold at for any meaningful period.

This is why price history is the only comparison worth making. If an item shows as forty percent off but the history shows it sitting at the sale price for most of the year, the discount is describing the reference number rather than the deal. If the history shows it genuinely trading higher for months, the discount is real.

Same claim, two completely different situations, and you cannot tell them apart from the product page alone.

What the tracking sites are doing

Price history sites and browser extensions are not guessing. They are collecting listing data repeatedly and comparing snapshots over time.

That collection is a real engineering job, which is why most of them buy it rather than build it. Amazon Scraper API is one of the services underneath: a REST interface returning roughly 55 structured fields per product across 20 marketplaces, including price, currency, rating, review count, stock status and which seller holds the buy box.

The individual reading is worthless. Thousands of readings over weeks are what turn "the price seems to jump around" into an actual chart.

The useful mental model

A listed price is not a property of the product. It is the current output of an ongoing negotiation between automated systems, and you caught it at one moment.

That framing will not get you a discount by itself. It will stop you feeling cheated when a price moves after you buy, and it makes the case for checking twice fairly obvious. The businesses on the other side of that screen have been measuring this for years, using exactly this kind of tooling, and there is no particular reason shoppers should not understand the game being played.