Most brands only notice stock when a SKU hits zero. Paid social notices earlier. If you ignore that lag, you pay to send people into a bad experience — then wonder why the account “got harder”.
I will not tell you to never advertise on thin stock. Flash drops, pre-orders, and honest “last sizes” campaigns are real. What I will say: running prospecting hard into inventory you cannot replace is usually expensive education.
Thin stock is not dead stock
A hero size at four days of cover is a logistics problem. A colourway you regret buying is a clearance problem. Mixing them into one panic pause is how you leave money on the table and burn trust at the same time.
Write the definitions down. Arguing about them every Tuesday is a tax.
A rule of thumb (steal it, then adjust)
When projected cover on a SKU you are actively pushing falls under about two weeks at current velocity, change the ads plan the same day. Not “Friday”.
- Under ~7 days: stop scaling that SKU. Pull it from broad prospecting. Retargeting can stay if fulfilment still holds and the page does not lie.
- 7–14 days: cap spend, watch refunds and CS noise, prep a substitute if you have one in stock.
- Healthy cover: scale is allowed to be boring. Boring is underrated.
Made-to-order brands live elsewhere. True drops with hard end dates do too. The point is a deliberate threshold, not vibes after a bad night’s sleep.
What goes wrong if you do nothing
You keep buying traffic. Conversion softens. ROAS still looks “fine” because remarketing props it up. Refunds arrive. Support explains delays. Meta optimises toward whoever still clicks. You train the account on a worse experience.
That loop is quiet. It does not put up a red banner.
What to do without rebuilding the brand
- List cover days for SKUs in active ads (or the SKUs those ads actually sell).
- Match them to ad sets, product sets, or creatives that feature them hard.
- Cut the worst mismatches first. One clean cut beats a week of “monitoring”.
- Leave remarketing a little longer if you can still fulfil. Prospecting is where damage compounds.
- If you keep spend on low cover, say so on-site. Lying in the ad is worse than pausing.
A familiar pattern
Small apparel brand scales a bestseller after a good creative week. Cover looks fine Monday. Sizes gone by Thursday. Ads keep spending the weekend because nobody owns the bridge between Shopify inventory and Ads Manager. Monday’s “efficient” ROAS bought a week of apology emails and a messy learning reset.
The fix was not a smarter bid strategy. It was a morning check that treated cover as a media input.
Where operators lose time
The hard part is not the pause button. It is noticing the mismatch while answering Slack, rewriting a subject line, and wondering why Shopping looks soft.
If stock lives in Shopify and spend lives in Meta, someone has to be the bridge every morning. On a small team, that someone is you. Putting cover and spend in the same glance is how you avoid funding empty shelves. Wider stack: Shopify ads software, Meta for Shopify.
Questions people actually ask
Should I ever advertise last sizes?
Yes, carefully. Be honest in creative and on the page, use tighter audiences, cap spend. Prospecting into last sizes without saying so buys returns and bad comments.
Does this apply to Google Shopping?
Yes. Feeds lag. See Shopping and empty shelves.
Edge cases worth naming
Pre-order models can advertise hard on zero stock if the page is honest about ship dates and your cash cycle can take it. Bundles complicate cover math — count the constrained component, not the fantasy of infinite virtual kits. Marketplaces plus DTC split inventory: if ads sell DTC stock you already promised wholesale, you will learn that the hard way.
International sizing is another quiet killer. “In stock” as a product can still mean the UK 10 is gone while the ad model wears a 10. Feature the available sizes or stop the spend.
If you want stock, ads and the morning list in one place: that is what we built Ralph for — a growth operator for Shopify, with review before anything ships. Not magic. Fewer tabs. See Ralph · Docs