How to read a price history
What the shape of a price chart tells you, what it leaves out, why the 90-day window matters, and how to tell a real floor from a one-day blip.
By Ahsan Khan, Founder. Published . 5-minute read.
Summary
A price chart is evidence, not a verdict. Read it for shape rather than for the lowest point: a flat line is the real price, a staircase is a permanent repricing, a sawtooth is a rhythm you can wait for, and a single downward spike is usually not an opportunity at all.
Our charts show what we observed, per retailer, timestamped. They do not show what we did not sample, whether the item was in stock at that price, or what a coupon did at checkout. Knowing those gaps is most of the skill. We window claims at 90 days because that is what the law puts around the word "was".
Read the shape first
- A plateau is the real price. A long flat stretch is what the retailer intends to charge, and what any discount should be measured against.
- A staircase is a repricing. A step down that never comes back usually means a successor launched or the model year turned over, and everything above the step has stopped being relevant.
- A sawtooth is a promotional rhythm. Regular dips to about the same level mean the item goes on sale predictably, so you can wait for the trough instead of paying at the peak.
- An upward spike is often not a price at all. Sharp rises on a marketplace listing usually mean the main seller ran out and a third-party seller filled the gap.
- A lone downward spike is the one people misread. An observation far below its neighbors is more often an error, a mispriced listing or a few hours of stock than the start of a trend.
What the chart is not telling you
- Sampling gaps. A history is a series of observations, not a recording. A flash sale that opened and closed between two samples leaves no trace anywhere.
- Availability. A low on an out-of-stock listing was never a price you could have paid, so a low that coincides with a stock-out deserves suspicion.
- Who was selling. On marketplaces, first-party and third-party listings differ widely and swap places without warning.
- Condition. New, open-box and refurbished stock are different products at different prices. We keep them apart; a chart that mixes them means nothing.
- Checkout-time reductions. Clipped coupons, promotion codes and instant rebates change what you pay without changing the listed price the chart records.
- Financed pricing. A monthly installment or an activate-today price is not a retail price, so we show those as promotions and keep them out of the score.
Why the window is 90 days
The 90-day window is not an arbitrary chart setting. It is the period California's former-price statute puts around a "was" claim, so using it everywhere means our authenticity test and our historical percentile answer the question the law asks. It is also about one promotional cycle for most categories: long enough to contain an event, short enough that the model year has usually not turned over inside it.
Where today's price sits inside that window is worth 25% of the FlexiScore on its own. At the window's low it scores near the top, at the high near the bottom, and it is computed only over our own observations, never over a retailer's claim about its past.
Median versus lowest
The median is the price you can expect to pay: half the observations sit above it, half below, and one strange day cannot move it. The lowest is the best case, often a single day and sometimes a mistake. Using the wrong one is the commonest way to misread a chart.
- Judge whether an offer is good against the median. Meaningfully below it is a genuine improvement on the normal state of affairs.
- Judge urgency against the lowest. Close to the floor, waiting has little upside; far above it, there is room to fall again.
- Read the gap between them as volatility. A narrow gap means a stable price and a predictable purchase; a wide gap means the item goes on sale hard and you should wait for the trough.
Telling a floor from a blip
A floor is a level the price returns to. It has been visited more than once, it held for more than a day or two, and it usually appears at more than one retailer, because competitors follow each other down. A blip is unrepeated, brief, confined to one seller, and often followed by a stock-out.
One case fools almost everyone. Just after a permanent step down, every day is a new low, so a lowest-in-90-days claim is trivially true and says nothing about whether today is a good day to buy. Compare against the days since the step, not the whole window.
- Has the price been at this level before, and did it hold for more than a couple of days?
- Does a second retailer show the same move within about a week?
- Is there a cause you can name: a launch, a model-year changeover, a calendar event?
- Did the listing stay in stock at the low, or did the low arrive as it ran out?
- Is the low on the same condition, configuration and seller as the offer you are considering?
Three or more yes answers and it is a floor. One or none and it is an artifact of sampling.
Questions shoppers ask
Why does your history start when it starts, instead of going back years?
Because it is our own record. We begin a product's history the day we start tracking it and do not import a third party's series to make a chart look longer. A history we did not collect is one we cannot stand behind, which is why every claim is qualified as being from our own tracking.
Why is your lowest price sometimes lower than anything I can find today?
Because it happened. A low can belong to a promotion that has ended, a marketplace seller who has gone, or a listing that sold out within hours. That is why we distinguish the median from the lowest, and why a lowest-price claim carries its window and observation count on the page.
Two retailers show very different prices for the same product. Which is wrong?
Usually neither. Check condition first, then configuration, then who is actually selling it, then whether one price assumes a membership or a store card. Genuine differences for the identical item tend to close within days.
Does a low price on the chart mean the deal is real?
Not by itself. A low price against an inflated reference price is still an inflated reference price, and a low that arrives at a stock-out is not purchasable. The history tells you where today sits; the authenticity test tells you whether the advertised saving means anything.
What this is built on
- config/score-weights.yaml — The window length, the observation floors and the percentile weight quoted here are read from it at build time.
- California Business & Professions Code 17501 — The source of the three-month window we standardized on.
Putting this to use
- How we score a deal — the published weights, the verdict thresholds, and the former-price test this guide refers to.
- How we make money — affiliate links, and the architectural reason a commission cannot move a verdict.
- All guides and explainers
Bylined and human-reviewed, not generated per product: 2% of machine-written sections on this site are sampled for review, and 100% of guides are. Editorial policy and corrections.