A price on a large marketplace can change several times between morning and evening. The movement is produced by automated systems responding to conditions rather than by any promotional decision.
Repricing is automated for most sellers
Sellers of any size use software that monitors competing offers and adjusts their own price within set boundaries, often within minutes of a change appearing.
Each system follows rules about minimum margin and how aggressively to undercut, so a single adjustment triggers a chain of responses across several sellers.
Nobody decides the resulting price directly. It emerges from rules meeting each other repeatedly throughout the day.
Inventory position feeds the rules
A seller holding more stock than planned instructs the system to price lower, while one close to selling out lets the price drift upward to slow demand.
Storage fees at fulfilment centres reinforce this, since holding stock past certain periods costs progressively more and pushes sellers to clear it.
The same item can therefore be cheap one week and noticeably dearer the next without anything changing about the product itself.
Demand signals arrive continuously
Marketplaces observe how many people view a listing, how many add it to a basket and how quickly it converts, all of which indicate current demand.
Rising interest supports a higher price, and systems tuned to that signal will test upward until conversion slows.
This is why an item featured in a popular article or video often becomes more expensive rather than less over the following days.
Fees are part of the calculation
A seller's floor price is set by product cost plus referral fees, fulfilment charges and shipping, and those fees change with weight, dimensions and category.
A change to fee structure moves the floor for every affected seller simultaneously, producing a category-wide shift that looks coordinated but is not.
Understanding the floor explains why some items never discount below an apparently arbitrary level. Bulky low-value goods have the highest floors relative to their price, because shipping and storage consume most of the margin before any discount is considered.
What a shopper can do with this
Because the swings are frequent and mechanical, a single observed price says little about whether an item is currently cheap.
Price history tools show the recent range, which turns an isolated number into a position within a distribution and makes a discount claim checkable.
Items with a wide historical range reward patience, while those with a narrow one are unlikely to move much regardless of when the order is placed.