First-party research
LiveThe Amazon Buy Box is not national
We check the same products from many locations. Every day. These are aggregate findings from that data. The numbers update on their own. Use them freely.
20.4%
split across locations
of products checked in 2+ locations on 2026-09-14
12 pts
win-rate spread by city
between Los Angeles and Seattle right now
18.2%
win where they are cheapest
of 44 products whose status differs by city
1,160,739
checks logged
17,746
positions in rotation
113
brands in sample
11.8/min
current check rate
Location changes the answer
A national Buy Box check gives you one answer. The question has several. On any given day, about a fifth of the products we track in more than one location show a different Buy Box status. It depends on where the shopper is standing.
Of products checked in two or more locations on a given day, the share that showed a different Buy Box status across those locations that same day. Today is excluded until its scrape completes.
View as table
| Day | Split rate |
|---|---|
| 2026-07-29 | 16.6% |
| 2026-07-30 | 17.6% |
| 2026-07-31 | 19.7% |
| 2026-08-01 | 20.6% |
| 2026-08-02 | 20.9% |
| 2026-08-03 | 16.4% |
| 2026-08-04 | 22.4% |
| 2026-08-05 | 21.8% |
| 2026-08-06 | 21.4% |
| 2026-08-07 | 21.7% |
| 2026-08-08 | 22.1% |
| 2026-08-09 | 23.0% |
| 2026-08-10 | 23.6% |
| 2026-08-11 | 23.3% |
| 2026-08-12 | 21.8% |
| 2026-08-13 | 21.6% |
| 2026-08-14 | 19.7% |
| 2026-08-15 | 19.4% |
| 2026-08-16 | 20.1% |
| 2026-08-17 | 21.2% |
| 2026-08-18 | 21.9% |
| 2026-08-19 | 21.3% |
| 2026-08-20 | 20.6% |
| 2026-08-21 | 20.3% |
| 2026-08-22 | 18.1% |
| 2026-08-23 | 20.3% |
| 2026-08-24 | 22.1% |
| 2026-08-25 | 21.7% |
| 2026-08-26 | 20.3% |
| 2026-08-27 | 20.1% |
| 2026-08-28 | 21.7% |
| 2026-08-29 | 19.6% |
| 2026-08-30 | 22.2% |
| 2026-08-31 | 21.4% |
| 2026-09-01 | 21.7% |
| 2026-09-02 | 22.6% |
| 2026-09-03 | 21.4% |
| 2026-09-04 | 20.4% |
| 2026-09-05 | 19.4% |
| 2026-09-06 | 20.8% |
| 2026-09-07 | 19.1% |
| 2026-09-08 | 20.5% |
| 2026-09-09 | 19.0% |
| 2026-09-10 | 19.1% |
| 2026-09-11 | 18.1% |
| 2026-09-12 | 17.3% |
| 2026-09-13 | 18.8% |
| 2026-09-14 | 20.4% |
The gap between cities is large
Win rates are not evenly spread. The best and worst city we track are 12 percentage points apart. Same catalogue. Same moment.
Share of tracked product-locations currently winning the Buy Box, by city. Bars start at zero. Cities backed by fewer than five brands are excluded. A single-brand city would publish one seller's performance, not an aggregate.
View as table
| City | Win rate | Brands |
|---|---|---|
| Los Angeles, CA | 78.4% | 18 |
| Dallas, TX | 76.8% | 15 |
| Atlanta, GA | 75.8% | 18 |
| Miami, FL | 74.2% | 19 |
| Chicago, IL | 72.7% | 23 |
| Denver, CO | 72.6% | 65 |
| New York, NY | 72.3% | 21 |
| San Francisco, CA | 69.5% | 17 |
| Austin, TX | 66.6% | 50 |
| Seattle, WA | 66.4% | 18 |
Price does not explain it
The obvious theory is that sellers win where they are cheapest. We looked at 44 products whose Buy Box status differs by city. Only 8 were winning in the city where their price was lowest. 22 were winning at a higher price than the city where they were losing. Another 14 showed no price difference at all.
The sample is small. 44 products is an indication, not a law. It is still enough to say that price alone does not decide the Buy Box.
Losing the Buy Box and having none at all are about equally common
Sellers watch for competitors taking the Buy Box. They should also watch for it vanishing. A suppressed listing shows no winner at all. Roughly as many of our tracked positions are suppressed as are lost to another seller. Both sit near 15%.
We do not claim one is commoner than the other. The two are within a point. The ordering reverses depending on how you measure it. A snapshot of current status puts suppression ahead. An average over the last seven days of checks puts loss ahead. A gap that small is not a finding. We would rather say so than pick the cut that reads better.
Current status of every tracked product-location. Suppressed means Amazon shows no Buy Box at all. A further 0.7% is unknown. It is omitted from the bar.
View as table
| Status | Positions | Share |
|---|---|---|
| owned | 10,115 | 72.3% |
| suppressed | 2,330 | 16.6% |
| lost | 1,451 | 10.4% |
| unknown | 104 | 0.7% |
What changes on a product page
We snapshot product pages daily and diff them. The most common change is not the one most sellers watch for.
Content tracking began 19 July 2026, so this series is deliberately short. Variations appearing and disappearing dominate everything else. They run at roughly five times the rate of title edits.
- variations removed
- variations added
- title changed
- description changed
- bullet points changed
- Other
View as table
| Change type | Total in window |
|---|---|
| variations removed | 9,641 |
| variations added | 7,768 |
| title changed | 1,767 |
| description changed | 469 |
| bullet points changed | 361 |
| Other | 1,082 |
Method and limits
- Buy Box checks logged
- 1,160,739
- Tracking since
- 2026-01-27
- Distinct ASINs
- 10,772
- Product-location pairs
- 17,746
- Brands in sample
- 113
- Competitor offers captured
- 146,103
What a location is. Each product is tracked once per city. Each city uses one representative postcode. So this measures variation between cities. It does not measure variation between postcodes inside one city. That is why we say “location” and not “ZIP”.
Which cities appear. A city is shown only if at least five distinct brands are tracked there. A rate built from one brand describes that seller, not the market. Publishing it would be neither useful nor fair to them.
Whose data this is. These are aggregate statistics across our whole customer base. No individual product, brand, seller or price appears on this page. None is available through it.
What this sample is not. It is the catalogue our customers chose to monitor. These are products they already suspected were worth watching. It is not a random sample of Amazon. A seller with no Buy Box problems is under-represented here. Three brands account for roughly a third of tracked positions.
Today is always partial. Every series here drops the current day. Its scrape has not finished. Including it would put a false dip on the right edge of every chart, permanently.
Why the charts start on 29 July 2026. Until 28 July our scraper had a bug. It treated a listing with missing delivery data as suppressed. It is not. The bug inflated our suppressed counts. In the location-split series it produced two spikes above 60%. Those spikes said something about our software, not about Amazon. We could have plotted them with a caveat. Instead we start the series on the day the fix landed. The charts are shorter. They are also honest.
Our sample as it grew. The plateau through June and July is a throughput ceiling in our own scraping. We found it and raised it in early August. It is a fact about our infrastructure, not about Amazon. The final partial week is excluded.
View as table
| Week beginning | Checks |
|---|---|
| 2026-01-26 | 6,439 |
| 2026-02-02 | 9,395 |
| 2026-02-09 | 10,423 |
| 2026-02-16 | 10,749 |
| 2026-02-23 | 8,643 |
| 2026-03-02 | 9,725 |
| 2026-03-09 | 10,397 |
| 2026-03-16 | 11,964 |
| 2026-03-23 | 20,484 |
| 2026-03-30 | 24,381 |
| 2026-04-06 | 17,553 |
| 2026-04-13 | 26,606 |
| 2026-04-20 | 28,534 |
| 2026-04-27 | 29,214 |
| 2026-05-04 | 45,658 |
| 2026-05-11 | 31,626 |
| 2026-05-18 | 32,044 |
| 2026-05-25 | 31,929 |
| 2026-06-01 | 32,215 |
| 2026-06-08 | 32,769 |
| 2026-06-15 | 32,517 |
| 2026-06-22 | 32,843 |
| 2026-06-29 | 33,115 |
| 2026-07-06 | 32,952 |
| 2026-07-13 | 33,720 |
| 2026-07-20 | 33,329 |
| 2026-07-27 | 36,215 |
| 2026-08-03 | 44,478 |
| 2026-08-10 | 54,892 |
| 2026-08-17 | 58,294 |
| 2026-08-24 | 73,554 |
| 2026-08-31 | 117,364 |
| 2026-09-07 | 119,299 |
Check your own products
Run a free Buy Box check across several locations and see whether your own catalogue splits.
Run a free checkFigures regenerate automatically. Free to cite with attribution to Buy Box Checker.