Sales up $7,100 a week since early October
Sales for the week to Saturday 17 October, explained
Our total sales (Amazon, website and TikTok Shop) jumped $4,700 in the week to Saturday 10 October, then another $2,400 in the week to Saturday 17 October, to $28,700.
That's $7,100 a week more than at the start of October, and our best week since the two launch weeks of NMN-99 and SPMD-99 in mid-September.
Total sales each week, all products, by where they sold
What drove it
- Week to 10 October: Amazon's Prime Big Deal Days (Tue 6 and Wed 7 Oct).Our Amazon sales rose almost 80%, from $7,500 to $13,400 (BROQ alone rose 73%).Prime Day discounts brought BROQ's average Amazon price down to about $69 a pack, from $77, so more people bought.The week before had been Amazon's weakest in months, so part of it was Amazon getting back to normal.
- Week to 17 October: the ads.No Amazon sale event and almost no Amazon ads, yet Amazon held at $13,700, about $1,800 above its average week in August and September.TikTok Shop sales nearly quadrupled, from about $1,300 in the week to 3 October to $5,100, as we spent more on TikTok ads.
Where the extra $7,100 came from
| Where the extra weekly sales came from (11 to 17 Oct, against 27 Sep to 3 Oct) | Change |
|---|---|
| Amazon getting back to its average week, after its weakest week in months | about +$4,400 |
| TikTok ads: about +$3,800 on TikTok Shop (exact) and about +$1,100 on Amazon from people who saw the ads | about +$4,900 |
| Meta ads: about +$2,700 on our website (website sales that came from a click on a Meta ad) and about +$400 on Amazon | about +$3,100 |
| Payday (the 15th fell in the 11 to 17 Oct week; those weeks run about 3% higher on Amazon) | about +$300 |
| Fewer repeat and monthly subscription orders on our website than in the 27 Sep to 3 Oct week (they bunch up around the 1st of the month) | about −$5,600 |
| Total | +$7,100 |
TikTok Shop's own +$3,800 is counted exactly. The other figures are Claude's best estimate, worked out from our weekly sales since June 2024.
The TikTok row is one week against another. The brief uses the 66-week average instead: about $1.00 more on Amazon and our website per $1 of TikTok Shop sales.
Fewer monthly orders on our website, far more first-time buyers (website and TikTok Shop)
Our website took about $2,900 less than in the 27 Sep to 3 Oct week, because monthly subscription orders bunch up around the 1st.
55 people bought for the first time on our website or TikTok Shop in the 7 days to Friday 16 October. That's more than triple the 16 the week before.
Are the ads doing more than we give them credit for?
No. Claude's figures for what the TikTok and Meta ads earn already include about $2,500 a week of extra Amazon sales, worked out from past weeks with and without ads.
In the week to 17 October, Amazon showed about $1,500 of it: $1,800 above its average week, less $300 from payday, with no Amazon sale and almost no Amazon ads.
So we're not undercounting the ads; if anything, we may be a touch generous.
How we could measure what Meta really brings in on Amazon
- Switch Meta ads off in about a quarter to a third of US states for 4 weeks.The states are picked to match the rest on past sales. Then we compare website plus Amazon sales by the state each order ships to.TikTok runs everywhere, so its ups and downs hit both groups equally and cancel out.
- It costs nothing extra.It pauses about $440 to $580 a week of Meta spend, and gives up few sales, because Meta's own website sales don't cover its cost yet.After about 5 to 6 weeks (the 4 weeks, plus 1 to 2 for sales to catch up with the ads) it tells us whether Meta brings in about 1.25 times its own website sales in all (what we count now) or about 2 times (enough for Meta to keep its budget). It gives a range, not an exact number.6 of Claude's 7 Meta measurement experts recommend it. It needs your and Nick's OK, because it changes where the ads run: the question in the brief.
- Two free extras we could start right away.A "how did you hear about us?" question after website checkout, and a daily look at Amazon searches for our brand names next to Meta spend.Meta's own "lift study" isn't worth it at our spend and can't see Amazon sales.