How to read a Keepa chart before you buy a book

Almost every Keepa guide on the internet was written for private-label sellers watching one ASIN they own. Media is a different problem. You are looking at a used book for eight seconds in a thrift store aisle, deciding yes or no, and moving on.

This is that version. It is the test we run on every unit that comes through our building.

What you are actually trying to learn

One question: has this specific item demonstrably sold in the last twelve months?

Not what it is listed at. Not what it sold for once in 2019. Not whether the cover looks valuable. Whether copies have been moving, recently, at a price worth your time.

Everything below is in service of answering that in a few seconds.

The lines that matter

Keepa's chart has a lot on it. For media sourcing, three lines do the work.

The sales rank line. Usually green, and the most important one. Rank is not a measure of value, it is a measure of recent sales activity relative to everything else in the category. Lower number means more activity.

The used price line. What used copies have actually been offered at over time.

The new price line. Less relevant for most used media, but it tells you if the title got reprinted, which matters.

Turn off everything else while you are learning. The chart is much easier to read with three lines than with nine.

The single most important concept: drops

The sales rank line is jagged. It rises steadily, then falls sharply, then rises again.

Every sharp drop is a sale.

When a copy sells, Amazon's rank for that item improves suddenly, then decays back over the following days as other items sell. That sawtooth pattern is the whole signal. You are not reading the rank number, you are counting the teeth.

A chart with many drops over the year is an item that sells regularly. A chart that is a smooth line climbing gently with no drops is an item nobody has bought.

This is why rank as a single number is misleading on media. An item can show a decent rank right now because one copy sold last week, while showing no other activity for eighteen months. The number looks fine. The chart tells you the truth.

Set your window to 365 days

This is the setting that matters and most people leave it on the default.

Set the chart to a one-year view. Not 30 days, not 90.

Thirty days on media is far too noisy. A book that sells four times a year might show nothing at all in a given month, and you would wrongly reject it. Conversely, a single seasonal spike can make a dead title look alive.

A year smooths both problems out and matches the actual question: does this thing move?

What you are looking for

Clear sales history in the last 365 days. Multiple visible drops spread across the year rather than clustered in one week. That is a sellable item.

A used price that has been stable or rising. Prices that trended steadily downward all year mean supply is increasing, and by the time your copy is live, the price may be lower still.

Consistency rather than volume. A book that sells six times a year at $30 is an excellent buy. It does not need to be a fast seller. It needs to be a real one.

What makes you put it back

No rank at all. The line is absent or flat across the whole window. This usually means the item has never sold on Amazon, or has not in so long that there is no data. It is the most common trap for beginners because these listings often carry high asking prices. A high price on an item that has never sold is not value, it is somebody else's wishful thinking. Put it back.

One drop, twelve months ago. Somebody bought a copy last spring. That is not a market.

Drops but at a collapsing price. The item sells, but each sale is lower than the last. You are buying into a falling market.

A recent new-price crash. A sudden drop in the new line often means a reprint or a new edition. The used market usually follows it down.

The traps specific to media

Edition and ISBN confusion. Textbooks especially. The chart you are looking at belongs to one specific ISBN, and the copy in your hand may be a different printing that maps to a different listing with completely different economics. Scan the actual barcode rather than searching the title.

Bundled or boxed listings. A box set listing has different data from the individual volumes. Make sure the chart matches the physical object.

Seasonality. Textbooks spike hard in January and August and go quiet between. A textbook chart read in June looks worse than the item actually is. Read the full year and note the pattern rather than the current moment.

Single-copy spikes. On obscure items, one collector paying a high price creates a dramatic-looking spike. One sale is not a market. Look for repetition.

The eight-second version

In an aisle, with a pile to get through:

Scan the barcode. Set or confirm the one-year window. Look at the rank line and count the drops. Several drops spread through the year, at a price that has held up, means buy. No drops, one drop, or a collapsing price means put it back.

Then check whether you can actually list it, because a perfect chart on a brand you are gated on is worth nothing to you.

Where this connects to us

That 365-day rule is not a guideline we invented for this article. It is our actual acceptance criterion. If an item shows clear sales history in the last year, we can sell it. Nothing stricter than that applies.

So if you are wondering whether something is worth sending us, you already have the test. Same chart, same window, same question.

And if the chart is good but you are gated on the brand, that is the exact situation consignment exists for. MaxX Access sells books, CDs, and DVDs on consignment, splitting net proceeds 50/50. If you can list it yourself, do, and keep everything. If you cannot, the chart still says the item has value, and the only thing standing between you and that value is which account it goes up on.

The habit worth building

Most sellers who fail did not fail at listing or pricing or software. They failed at this one check, and they failed at it several hundred times before the storage fees made it obvious.

Run the test on everything. Even when you are sure. Especially when the asking price is high, because that is exactly when the chart is most likely to be empty.

Related reading: what we accept, and free scanning software for Amazon.