Written by: Michał Sobieraj, Operations Manager, Digital Colliers
You bought 4 months of stock for a SKU your Meta account decided not to sell. That sentence lands because most DTC operators have lived it. The purchasing team ran their forecast in one tool. The growth team reallocated budget in another. Neither meeting had the other's numbers on the table.
The result is cash locked in a warehouse while your ad account quietly funnels spend to three or four hero SKUs. Nobody made a bad decision in isolation. The decisions just never met.
How the two sides decouple
Inventory planning usually runs on a rolling 8 to 16 week horizon. It looks at sell-through, lead times, MOQs, and last season's curve. It assumes demand is roughly a property of the product.
Growth spend runs on a 24 to 72 hour horizon. Meta's delivery system decides which SKUs to push based on ROAS, creative fatigue, and audience response. It assumes demand is a property of what the algorithm can profitably buy right now.
Those two clocks don't sync. And they've gotten worse recently. Customer acquisition cost across DTC has climbed roughly 40% since 2023, and Meta CPMs kept rising through 2024 and 2025. When paid gets more expensive, the ad account concentrates spend more aggressively on winners. The tail of SKUs you bought inventory for gets even less air.
The cash-lockup math
Here's the shape of the damage. Say you committed to 16 weeks of cover on a SKU at a landed cost of £8 a unit, 5,000 units on the water. That's £40,000 sitting still. If the ad account decides that SKU isn't a winner, your actual sell-through drops to maybe 30 weeks of cover. You've now got roughly £18,000 that won't recycle into new inventory for half a year.
Multiply across a catalog. Profitero's work suggests roughly 30% of SKUs at a typical multi-channel brand lose money per order once you net out returns and ad spend. In apparel it's uglier, because UK return rates run 25 to 40% by category and online returns overall sit around 19 to 20% of gross sales. A SKU the algorithm won't push, that also gets returned at category rate, is a cash fire.
And the backdrop matters. UK eCommerce grew about 3% in 2024. Single-digit growth is the new baseline. You can't outgrow a working capital mistake anymore.
How the closing-the-loop teams do it
The operators I see handling this well aren't running fancier models. They're just forcing the two sides to look at one number together. A few patterns keep showing up:
- A weekly SKU-level P&L that includes ad-attributed revenue, blended CAC, return rate, and weeks of cover. One row per SKU. Merchandising and growth both sign off.
- A rule that any SKU with more than 20 weeks of cover and declining ad delivery share gets flagged before the next PO cycle, not after.
- Creative and offer testing budget explicitly earmarked for the long tail, so the algorithm gets a fair look at SKUs the buying team already committed cash to.
- A returns feedback loop into the buying decision, not just the finance close. If a SKU returns at 35%, the buyer knows before reordering, not the quarter after.
None of this is exotic. It's mostly a data plumbing problem. Getting Shopify, Meta, the 3PL, and the returns platform into one shared view is what most brands haven't done yet.
What to alert on
If you can't rebuild the whole stack this quarter, wire up alerts for the four states that cost you the most money:
- SKU with more than 12 weeks of cover and less than 5% of ad account spend in the last 14 days.
- SKU with rising return rate quarter over quarter and an open PO.
- Ad set spending above threshold on a SKU with less than 4 weeks of cover.
- Category-level ad spend growing while category sell-through is flat.
Each of those is a specific piece of cash walking out the door. The cost of inaction isn't a bad quarter. It's a slow bleed where the growth team hits ROAS, the buying team hits fill rate, finance hits the forecast, and the business still runs out of working capital. Around 95% of enterprise AI projects never reach production, and most of the ones that do skip this exact seam. The unglamorous work of joining two calendars is where the money actually is.