New to Brand means the purchaser had not bought anything from your brand on Amazon in the previous twelve months.
Not new to the product. Not new to the category. New to you, on a rolling one-year definition, measured by Amazon across your whole brand rather than the ASIN they bought.
That definition drives every NTB figure in TrackIQ, and getting it slightly wrong leads to real budget mistakes.
Prefer to watch? Sixty-eight seconds on the definition, and what it changes about budget.
What follows from the definition
A returning customer buying a product they have never tried is not NTB. They already know your brand. The metric is about acquiring people, not about product discovery.
A customer who last bought thirteen months ago counts as NTB again. The window rolls, so genuinely lapsed customers re-enter the count. That is reasonable — winning someone back after a year does cost acquisition effort.
It is brand-wide. Buy your shampoo in March and your conditioner in June, and only March was new-to-brand. This is why NTB by product tells you which product does the acquiring, not which sells most.
Only Amazon purchases count. Someone who buys from your own website every month is new-to-brand the first time they buy on Amazon.
Why it is the number worth optimising
ACoS asks was this sale efficient? NTB asks did this sale buy me a customer? Those come apart constantly.
A campaign on your own brand name usually looks superb on ACoS and produces almost no new customers — it harvests people who were already coming. A broad campaign against a generic query often looks mediocre on ACoS and brings in people who have never heard of you.
Judged on ACoS alone, you cut the second one. That is the mistake NTB exists to prevent, and it is why the metric matters more than its reputation as an upper-funnel vanity number.
Where it shows up
| Where | What it tells you |
|---|---|
| NTB ASINs | which products bring in new customers |
| NTB Purchases | which campaigns bring in new customers |
| DSP reporting | new-to-brand purchases from display |
The first two are the pair worth reading together, because they answer different budget questions: which product to protect, and which campaign to fund.
Gateway products
Some products acquire; others are bought by people who already know you. NTB share by product tells you which is which.
A gateway product — high NTB share — earns different treatment from a high-revenue product. It is the front door. Running it out of stock, or cutting its advertising because its own ACoS is unremarkable, costs you customers who would have gone on to buy other things.
That is the practical use of the whole metric: it changes what you protect.
Reading NTB share correctly
Three cautions, each of which produces a wrong conclusion if missed.
NTB share falls as a product matures, and that is fine. Once most buyers in your category know you, a lower proportion of purchases can be new. Falling share on rising volume is not a problem.
Do not add months. AMC reports monthly, and each month is a distinct cohort of people. Somebody new in March may buy again in April — adding the months double-counts them. Compare months; never sum them.
A high NTB share on tiny volume is noise. Three purchasers, all new, is 100% and means nothing.
NTB purchasers, purchases and sales
Three related figures that answer different questions:
| Figure | Counts |
|---|---|
| NTB purchasers | people |
| NTB purchases | orders |
| NTB sales | money |
For acquisition questions you want purchasers, because that is the count of humans you gained. Sales figures are skewed by basket size.
Where it comes from
NTB is calculated by Amazon and delivered through AMC, which publishes monthly. So a brand-new account has no NTB data until Amazon's next delivery — that is expected, not broken.
See Why AMC Data Is Monthly and Sometimes Empty.
Related
- Understanding AMC Cloud — the dashboards NTB appears on.
- DSP Metrics Explained — NTB purchases in display reporting.
- Aria Prompt Library — questions 33 and 34 ask these directly.