Amazon Ads Attribution Window Explained (7-Day vs 14-Day)

Amazon's attribution window is the period after an ad interaction during which a purchase is credited to that ad. For sellers, Sponsored Products uses a 7-day click window. Sponsored Brands and Sponsored Display use 14 days. Vendors see 14-day attribution for Sponsored Products as well.

The window explains one of the most common PPC mistakes: judging the last few days of data before all the sales have arrived. Here's how attribution works and how to make decisions that account for it.

Attribution windows by ad type

Amazon's attribution rules can change, and details vary by format. Check the metric definitions in the Ads console (hover over a column header) for the current rules on your account.

How a sale gets credited

1. A shopper clicks your Sponsored Products ad on Monday.
2. They leave, compare other options, and come back on Thursday to buy.
3. Because Thursday is within 7 days of the click, the sale is credited to your ad and to the date of the click (Monday) in your reporting.
4. If they had bought the following Tuesday, 8 days later, it wouldn't count toward the ad.

Why recent ACoS always looks worse

Spend is recorded immediately. Sales trickle in over the following days. So for any recent date, spend is complete and sales are partial.

Worked example

If you'd cut bids on Tuesday because Monday's ACoS was 55.6%, you'd have reacted to a number that ended up at 29.4%. The exact pattern depends on your product: impulse buys convert quickly, while considered purchases take longer.

Halo and brand sales

Amazon credits more than just the advertised item. If a shopper clicks your ad for a 12-inch skillet and buys your 10-inch skillet, that sale is attributed to the ad as an other-SKU (halo) sale. The Advertised product and Purchased product reports separate these, which helps when you're deciding which products to advertise.

• High halo share: the advertised product works as a doorway to your catalog. It may be worth advertising even if its own ACoS looks weak.
• Low halo share: the ad sells only that product. Judge it on its own numbers.

Comparing ad types fairly

Because Sponsored Brands and Sponsored Display get a 14-day window and Sponsored Products gets 7 days for sellers, a direct ACoS comparison slightly favors the longer window. Sponsored Display's view-through attribution on vCPM campaigns credits sales to people who saw but didn't click the ad, which also makes its numbers look different. Keep this in mind before moving budget between ad types based on ACoS alone. See Sponsored Products vs Sponsored Brands vs Sponsored Display.

Attribution vs data delay

Two separate things make recent data incomplete:

• Reporting delay: clicks, spend, and sales can take hours to appear in the console. Same-day numbers are not final.
• Attribution window: even after reporting catches up, sales keep being credited to past clicks for 7 or 14 days.

Both mean that 'yesterday' is the least reliable day in any report.

How to make decisions that account for attribution

1. Exclude the last 2 to 3 days from bid decisions. Use a range that ends three days ago, for example the 14 days ending three days ago.
2. Use longer windows for low-volume keywords. Thirty days or more gives enough clicks to judge.
3. Look at your own lag. Compare how much a day's sales grow between day 1 and day 7. If most sales arrive within 2 days, you can act faster; if they keep arriving, wait longer.
4. Don't judge a keyword on zero orders until it has enough clicks and the window has had time to close.
5. When comparing weeks, compare weeks that are both fully attributed.

A simple bid rule that respects attribution

Every week, look at the 14 days ending three days ago. For keywords with 15+ clicks: if ACoS is more than 20% above target, lower the bid 10% to 15%. If it's more than 20% below target and impressions are limited, raise it 10%. Leave everything else alone. This avoids reacting to incomplete days while still moving quickly enough.

Measure your own attribution lag

You can work out how quickly sales arrive for your products with a simple test. Pick a recent date. Note its attributed sales the next day, then three days later, then a week later. Do this for several dates and average the results.

In this example, a day's numbers are about 85% complete after three days, so a three-day buffer is reasonable. A product with a longer buying cycle, such as furniture or electronics, might only be 60% complete after three days, which means a longer buffer before acting. These percentages are illustrative; your own will differ.

Adjusting for lag in a pinch

If you must look at recent data, you can estimate the final figure by dividing current sales by the completion share for that age. Day-old sales of $110 at 55% complete suggest about $200 once fully attributed. Treat it as a rough estimate, not a basis for big bid changes.

Attribution and external traffic

Sponsored Ads attribution only covers Amazon ad interactions. If you drive traffic from Google, Meta, TikTok, or email, use Amazon Attribution (available to eligible sellers and vendors) to tag those links and see what they drive. Those sales won't appear in your Sponsored Ads ACoS, but they will appear in total sales, which is one reason to track TACoS too.

AdSpark's bid optimizer accounts for attribution lag when it evaluates recent performance, so bids aren't cut on days that are still collecting sales. For more on reading ACoS correctly, see what is ACoS, and for the mistakes this causes, common Amazon PPC mistakes.

FAQ