Written by: Nicole Ogonowska, IT Growth Manager, Digital Colliers
If you're still running your growth reviews around blended CAC and first-order ROAS, you're playing a game the winners quietly stopped playing about eighteen months ago. The math broke. Not slowly. It broke around the time paid social costs stopped resetting between quarters and returns started eating margin on the products that were supposed to be your acquisition workhorses.
The LinkedIn post said your competitors stopped solving CAC. This is the longer version of why.
The CAC curve isn't cyclical anymore
DTC customer acquisition costs are up roughly 40% since 2023, and Meta CPMs kept climbing through 2024 and 2025. That's not a bad quarter. That's the new floor.
At the same time, UK eCommerce is growing at single digits year over year. So you're paying more to reach a market that's barely expanding. If your plan for 2026 is a bigger paid budget and better creative, you're planning to lose margin more efficiently than last year.
The operators I keep seeing pull ahead have accepted something uncomfortable: CAC is a market price now, not a lever. You don't optimise a market price. You build economics that can absorb it.
First-order economics were always a fiction
Here's the part nobody wanted to say out loud in 2021. First-order contribution margin has never been the real number. It just used to be close enough to the real number that you could ignore the gap.
The gap is huge now, for three reasons stacked on top of each other:
Returns are running around 19-20% of gross online sales, and 25-40% in apparel depending on category. That's not a rounding error against a thin first-order margin.
Around 30% of SKUs at a typical multi-channel brand actually lose money per order once you subtract ad spend, returns, and fulfilment. Most finance teams don't know which 30%.
Support cost per order is a real line item that almost nobody tracks against the acquisition channel that produced the order.
If your P&L only sees the sale, you can't see any of this. You just see revenue that doesn't turn into cash.
What the retention math actually looks like
The brands winning this cycle connected three data sources their competitors still keep in separate tabs: purchase history, support tickets, and fulfilment or returns data. When you join those, the picture gets brutal and useful at the same time.
You stop asking "what did we spend to acquire this customer" and start asking:
Did they come back inside 60 days.
What did the second order cost to serve, net of returns and support contacts.
Which acquisition source produces customers whose second order is actually profitable.
That last one is where most of the money is hiding. Two channels can look identical on first-order CAC and differ by 3x on 90-day contribution margin. If you're only measuring the front of the funnel, you're funding the worse channel and starving the better one.
The metrics the 2026 winners actually track
The dashboards look different. Fewer vanity numbers, more operational ones. The pattern I keep seeing:
60-day repeat rate by acquisition cohort. Not blended. By channel, by first product, by discount depth.
Contribution margin per repeat order. After returns, after support, after fulfilment. This is the number that pays the rent.
Return-adjusted product margin at SKU level. So you can actually see which of your 30% loss-making SKUs to fix, reprice, or kill.
Support contacts per order by SKU and by channel. A cheap product with three support tickets is not a cheap product.
None of this is exotic. The data exists in Shopify, Gorgias or Zendesk, and your 3PL. The reason most teams don't have it is that joining those systems reliably is engineering work, and it doesn't feel as urgent as the next campaign.
The left-behind risk is quieter than you think
The brands falling behind in 2026 aren't going to have a dramatic quarter. They'll have eight quiet quarters. Paid budget creeping up, contribution margin creeping down, cash conversion getting worse, and no clear place to point.
Meanwhile the competitor who spent six months wiring their purchase, support, and fulfilment data together is now making channel decisions on 90-day margin instead of first-click ROAS. They can outbid you on the customers who repeat and let you have the ones who don't.
That's the game now. It's not about solving CAC. It's about knowing which customers are worth the market price and which ones aren't. If your stack can't tell you that by cohort, by SKU, and by channel, that's the work for the next two quarters.
Sources
This article was originally published on the Digital Colliers Blog. Digital Colliers helps DACH and UK companies implement AI — see our AI consulting services or contact us.
Top comments (0)