1. Monthly visits
We start from public traffic rankings of each store's domain. A rank turns into monthly visits with a curve calibrated on stores whose traffic is known. Up to rank 2,000 visits follow a power law, about 38,266,551 visits a month at that rank. Beyond it visits fall in proportion to the rank. Visits are the input with the most error, so treat traffic as a range.
2. Conversion rate
Orders are visits times a conversion rate. We start from a base rate of 2% and adjust it by the store's typical price, because cheap products convert more often than expensive ones:
| Typical price | Factor | Conversion |
|---|---|---|
| Under $25 | 1.4x | 3% |
| $25 to $75 | 1.1x | 2% |
| $75 to $200 | 0.75x | 2% |
| Over $200 | 0.4x | 0.8% |
The result is kept between 0.3% and 5%. Category and store maturity adjust it a little more: older stores with many reviews convert better than week-old tests.
3. Revenue
Revenue is orders times the average order value. We take the store's typical product price and multiply it by 1.2 items per order, a common average for small online stores. Revenue is shown in US dollars and in the store's own currency.
4. Which products sold
Stores rank their own best sellers. We split a store's orders across its products by that order with a power curve, so the top product takes the biggest share. The curve is steeper for small stores, where one product usually carries the business:
| Products in the store | Exponent | Share of the top product |
|---|---|---|
| 1 to 5 | 1.8 | 64% |
| 6 to 50 | 1.3 | 34% |
| 51 to 500 | 1 | 15% |
| More than 500 | 0.8 | 6% |
5. Tracked stores
When someone tracks a store we check it every hour. Changes in stock levels and best-seller order between checks refine the hourly picture and show which products moved. Tracked stores therefore have the most detailed and most reliable numbers.
6. Ads
Reach comes from the ad networks' public transparency data where it exists, mostly for ads shown in the EU. Elsewhere we model it from how long an ad runs and how many ad sets push it. Spend is reach times a typical cost per thousand views for the country and network. The AI score combines run time, reach growth and ad sets.
What the numbers are good for
Use them to compare stores, rank products, find winners and spot trends. Do not use them as a store's real accounts. A store with heavy repeat customers, a big email list or offline sales can sell much more than its traffic suggests.