Whatnot hands every seller the same thing each week: a CSV with 34 columns that records every order, fee, buyer and show. Here is what two years of them hold, using sample data: 104 weekly reports, October 2024 to September 2026. Read them the way an accountant would, every row, every fee, every buyer, every show, and they say more about the business than the app ever will.
The shape of two years, in the sample data
- About $248,000 in item sales across 104 weeks: 13,000 paid orders, 13,500 items.
- About $217,000 paid out. The gap, about $31,000, is Whatnot’s commission and processing, an effective rate of 12.6% of sales. Budget with that number, not “8%.”
- 196 live-show sessions, 48 of them raid trains.
- Year two ran 28% ahead of year one in item sales, on 9% more shows. The average order climbed from about $17 to about $21. The growth came from selling better, not just more often.
The Whatnot app is built for selling live. The weekly CSV is where the business view lives, and all of it is in there.
Lesson 1: the fee rate is not 8%
The headline commission is 8%, but processing is 2.9% of the buyer’s total including shipping plus $0.30, so cheap items with paid shipping run a much higher effective rate. At a $17 average order the $0.30 alone is 1.8% of the sale; at $21 it drops to 1.4%. The fee rate you feel sits closer to 13% than 8%, and it is worst in the weeks you sold small. Raising the average order was worth more than any single show.
Lesson 2: giveaways are a marketing spend. Measure them like one.
1,120 giveaways went to 740 different buyers over the two years. The direct cost was small, under $800 in shipping labels, but the real question was whether recipients came back. Track each recipient’s spend before and after their first giveaway and the answer is plain: 305 of the 740 (41%) bought again, and as a group they spent about 1.8 times as much after their giveaway as before. The other 59% took the item and never bought again. Whatnot draws the winners at random, so that is not a list to aim at. It is a measure of what a giveaway is worth to the shop, and it is correlation, not proof: regulars who win would often have bought anyway.
Lesson 3: “regulars” are a smaller group than it feels like on stream
Live selling feels like a crowd. The report says otherwise. Of 2,050 buyers, the top 10% produced 49% of revenue, and the buyers who came back in two or more different weeks, about a third of everyone, produced 70% of it. In year two that group grew faster than the buyer list did, which is the healthiest number in the whole set. Knowing your top twenty by name, and what they bought last, changes how you run a show more than any lighting upgrade.
Lesson 4: the report keeps time in UTC (convert it before judging your hours)
Timestamps are UTC. Every evening show looks like it happened the next morning until you convert. Once you do, the peak is obvious: the biggest revenue hours were late evening, roughly 8 to 11 PM local time, and Sunday evenings outsold every weekday. Check yours. Moving a show costs nothing.
Lesson 5: preloaded BINs make your shows look longer than they were
Buy-it-now items on a preloaded show page get attributed to that show even when they sell hours before you go live. In the sample data, one show that ran under four hours read as more than nine until the run of sales while live was separated from the stragglers. If you judge shows by sales per hour, this one matters.
Lesson 6: check that gross, fees, and payout reconcile
The payout column in the CSV is already net of fees. Subtract the fees again, as plenty of spreadsheets do, and you understate profit by about 13% of sales. Whatever tool or spreadsheet you use, check that gross minus fees equals payout. If it does not, one of the definitions is wrong.
What we would tell a new seller
- Keep every weekly CSV. It is your ledger.
- Enter cost of goods when you list. Everything downstream depends on it.
- Budget fees at 13%, not 8%.
- Know what your giveaways are worth. Whatnot draws the winners; you choose the type and how often.
- Look at your shows in local time, and by the hours you were actually live.
Upload one weekly report and MARGN.PLUS runs every one of these checks for you.
MARGN.PLUS is independent and not affiliated with Whatnot, Inc. Whatnot is a trademark of Whatnot, Inc.
