Insights from auditing the trailing 12-months data of 150 real DTC brands.
Findings from Kynship's dataset of 150 DTC brand P&Ls, including 72 with a complete trailing twelve months of data.
This is not survey data and it is not self-reported benchmark data. It is line-item P&L and forecast workbook data from brands that handed over their actuals.
This is the one most people get wrong. Split the portfolio into thirds by fixed cost load and the contribution margin barely moves. The profit margin does.
The unit works. The company does not. This is the least discussed finding in the set, because the entire industry conversation is about CAC, ROAS, and creative.
Across the 57 brands with cohort data, repeat rate has no statistically significant relationship with profit. The correlation is -0.17. That should be surprising given how retention gets sold. Here is why it happens.
Take only the high-retention brands and split them by cost of delivery.
The mechanism: high-retention brands run a 1.2 aMER against 2.4 for everyone else. They knowingly halve front-end acquisition efficiency because they trust the repeat rate to pay it back.
With a healthy margin, that works. With a broken one, every repeat purchase compounds a loss instead of a profit.
Winner brands run a lower aMER but a far higher repeat rate. They are buying new customers less efficiently than everyone else and still coming out ahead, because the back end pays for the front end.
This runs directly against the industry's obsession with acquisition efficiency as the scoreboard metric. But read it alongside finding 02: the high repeat rate is only an advantage when the margin underneath it can carry a worse aMER.
Winning brands put a much larger share of new-customer revenue through Amazon. The read is that multi-channel presence compounds the acquisition flywheel rather than eating into it.
Treat this as correlation only. It could just as easily reflect that larger, more operationally mature brands both win and have Amazon operations.
We will run the same analysis on your numbers and tell you whether the problem is the unit or the overhead.
Book a P&L audit 30 MINUTES, NO PITCH DECKSource: Kynship proForma cross-brand analysis. Trailing twelve-month actuals pulled from each brand's forecast workbook.
72 brands cleared a complete trailing twelve months of data, representing roughly $745M in revenue. 57 of those carried cohort-level data sufficient for retention analysis. Group figures are medians taken independently, so cost lines do not sum exactly to the profit margin of each group.
No brand is named and no figure is attributable to an individual brand.