How TokConnect estimates TikTok Shop sales

What the numbers are

Every revenue and unit figure on these pages is an estimate. The source reads TikTok's public sold counter for each listing and multiplies the change by the price it records. That price can sit below a listing's list price when discounts apply, so a page's average price is revenue divided by units, not a list price. The source cannot see refunds, cancelled orders, what a brand keeps after fees or its own accounting, and no brand has approved or endorsed a page. The source gives no error range, so treat any figure as an order of magnitude with a direction, not an audited total.

Brand names appear only to say whose sales the numbers describe. We link to no logos and hot-link none.

Windows, dates and refresh

Pages use two windows: the last 30 days and the last 7 complete days. Every page shows Updated (the day we pulled the data) and Data to (the last complete day the source reports, which runs a day or two behind). Days are the source's calendar days; it does not state a time zone. Brand pages refresh weekly and studies monthly. If a pull fails we keep the previous data rather than show a blank, and a page whose data is more than 14 days old fails our publishing check.

Growth compares the 7 most recent complete days with the 7 complete days before them, computed from the daily series on the page, and is published only when the earlier 7 days made at least $10K.

Days the source reports low

Some days read far below normal across most brands at once. For each brand we divide a day's revenue by that brand's median day; when 25% or more of the profiled brands read under half of their own median on a day, we treat the day as incomplete in the source, not as a sales dip. In the current data that is: 30 Sep and 1 Oct. Those days are drawn as a hatched, dashed segment, are left out of every peak, growth and typical-day figure, and are never filled in by estimate. Totals stay as the source reports them, so a 30-day total can read a little low.

Growth against the market

Each brand's change is shown next to the market's change over the same days. The market index is the median, across profiled brands, of day revenue divided by each brand's median day. Days when 25% or more of brands sold above 1.5 times their median are platform-wide surge days (8 to 11 October and 15 to 17 October in the current data); a window that includes them is described as partly a surge. We use a verdict word such as growing, slowing or cooling only when a brand differs from the market by 20 points or more and the difference holds on at least 5 of the 7 days. Otherwise we state the two numbers and nothing more. A single day is called a spike only when it is at least twice the median of the other days and the next best day is under 70% of it; three or more days near the peak are called a plateau. We do not publish weekday patterns: 30 days gives four samples of each weekday.

For the last 7 complete days (13 to 19 October) the market index was 0% against the 7 before (6 to 12 October).

Which brands get a page

A brand needs all of these: the source files it as a brand and not a retailer, at least $250,000 of estimated revenue in 30 days, sales on at least 20 of those 30 days, at least 3 products with sales and no single product above 90% of its revenue. All storefronts of one brand are combined: where shop names share a brand name (VEVOR Garden Store and VEVOR Tools Store are VEVOR), revenue and units are summed, daily series merged by day and products combined, and the page says it combines N storefronts. 8 of the 88 profiled brands combine more than one storefront. A brand that fails a floor is not published; we keep a log of why.

Ranks, medians and peers

Rank compares a brand's estimated 30-day revenue with every brand in the shop pool: 1,337 shops at $250,000 or more, of which 1,139 are filed by the source as brands and 198 as retailers; after combining storefronts that leaves 1,070 brands. Category rank uses the source's shop classification, which is by the shop's main category and not by what each listing sells; we state a category rank only when the category holds 10 or more tracked brands. Medians are medians across the other tracked brands (so the figure differs slightly by page, because the brand itself is left out). The pool is rebuilt monthly from 29 search characters across 6 revenue bands.

Products, price flags and coverage

Product tables list each listing by estimated revenue and cover the top 100 listings of a brand; every chart on listings says so. The count of listings with sales is the source's own count where it gives one, and "at least N" where it does not. The source flags listings whose revenue per unit falls far outside the product's listed price range, usually a placeholder price on TikTok, and it can fail in either direction. We run that check for every brand: the brand name is searched, and any best seller that search does not return is searched by its own title. When the flagged listings carry 10% or more of a brand's source total, the headline revenue, units, average price, listing shares and rank use the figure without them, labelled as such, with the source total and the flagged amount stated once beside it (20 of the 88 profiled brands are shown this way, marked † in the directory). Below 10% the headline keeps the source total and the flagged part is noted in the method line. Other brands' figures in a rank are as the source reports them. Commission is the public affiliate rate where the source lists one. "Listed" is the listing date TikTok shows.

What the labels mean

Affiliates are independent creators who sell a brand for a commission. Brand accounts are accounts whose name or handle is the brand's own name. Linked host accounts are the other accounts the source files under the shop's own promotion; they are counted, not named. Shop tab and mall is revenue from the shop tab, search and the TikTok Shop mall, not from a creator or the brand's accounts. Video, LIVE and shop tab is a second split of the same revenue by format, so it overlaps the source split: affiliates also sell through LIVE. Shares of the source and channel splits are shares of the source total. The top-N affiliate bands cover only part of a brand's affiliate revenue (each page states how much).

How we mask creators

Individual affiliate creators are private people. We never publish a creator's name, handle, bio, contact details or id next to sales estimates, and we remove personal names and creator collaborations from product titles. Creators appear only as aggregates: a rank, a follower band and a share of revenue. Accounts that carry the brand's own name are named because they are the brand. Files we publish are checked for handles and email addresses before every release.

Studies and pools

Studies are computed from stored pools by fixed functions, so the same data gives the same numbers. The product pool holds products with at least $25,000 of estimated 30-day revenue gathered with category and keyword searches, grouped by the category ids on each row, with price-flagged rows removed. Totals are sums over these tracked products, not the size of a market.

Corrections

If a number is wrong, or you represent a brand and see something that is not accurate, email dj@tokconnect.com with the page and the figure. We check it against the source, correct the page and note the correction here. Corrections published so far: none.