By Steve Merrill · October 26, 2026
Why are Meta ad costs climbing while returns flatten?
Because the auction is more crowded and the targeting got worse at the same time. More brands are bidding for the same eyeballs, and Apple's privacy changes gutted the signal Meta used to find buyers. You pay more to reach people, and Meta knows less about who those people are.
I watched this happen in slow motion.
I ran a clothing company for 15 years. Grew it to about $10 million a year before I exited. And for a long stretch, I refused to run Facebook ads. I thought I was being disciplined. I was being stubborn.
Competitors who went all in on paid social pulled ahead of me. So I caught up. I ended up spending more than $5 million of my own money on Facebook ads over the years. I learned the channel cold.
Here's the thing. The channel I spent all that money mastering doesn't work the way it used to.
What actually broke the Meta ads math?
Two things stacked on top of each other.
First, Apple's App Tracking Transparency. When it rolled out in 2021, most iPhone users opted out of tracking. Meta told investors the change would cost the company around $10 billion in 2022 revenue (CNBC). That is not a rounding error. That is the signal that powered Meta's targeting getting cut off at the knees.
Second, more competition in the same auction. Every DTC brand learned the same playbook. When everyone runs the same prospecting and retargeting motion, CPMs rise and nobody wins. eMarketer has tracked rising customer acquisition costs across DTC for years, and the trend line hasn't reversed (eMarketer).
So your cost to acquire a customer goes up. Your return on ad spend goes flat. You run harder to stay in the same place.
Same story. Different year.
What is AI discovery and why does it matter for DTC?
AI discovery is when a shopper asks ChatGPT, Perplexity, or Google's AI mode for a product recommendation and the model names specific brands. No ad auction. No CPM. The customer gets pointed at you because the model decided your product is a good answer.
This is small today. It is not small tomorrow.
Adobe Analytics reported that traffic from generative AI sources to US retail sites grew more than 1,200% year over year heading into the 2024 holiday season (Adobe). Off a tiny base, sure. But that is the kind of curve you want to be early on, not late.
And the behavior is sticky. When someone asks an assistant "what's the best merino base layer for cold-weather running," they treat the answer like advice from a friend. They don't scroll past it the way they scroll past an ad.
Because an ad interrupts. A recommendation gets trusted.
Why is AI discovery a hedge against Meta ad fatigue?
The economics run in the opposite direction.
Paid social gets more expensive as more brands pile in. AI discovery rewards the brands that are easiest for a model to understand and cite. One channel punishes you for competition. The other rewards you for clarity.
I'm not telling you to turn off Meta. Meta still moves real volume, and ripping it out would be dumb. What I'm telling you is to stop betting your whole acquisition engine on a channel that charges you more every quarter for the same result.
Build a second path. One you own. One that doesn't live and die by an auction.
How do Shopify stores actually get picked up by AI assistants?
AI models recommend products they can read and verify. That comes down to three things.
Clean structured product data. Your feed and your Product schema tell the model what you sell, what it costs, and whether it's in stock. Shopify's own guidance points merchants toward structured data and solid product information so their catalog is machine-readable (Shopify). Messy data gets skipped.
Third-party proof. Models lean on reviews, comparison articles, and roundups to decide which brand is a safe recommendation. If nobody credible mentions you, the model has no reason to.
Direct answers on your own site. Pages that answer the real questions buyers ask give the model clean language to quote. That is the whole game behind answer engine optimization, and it is a different muscle than keyword SEO.
Most stores fail the first one. Their feed is a mess, so the model can't trust them.
What should a DTC operator do about this right now?
Start by auditing what an AI model can actually see when it looks at your store. Pull up ChatGPT and Perplexity. Ask them to recommend products in your category. See if you show up. Most brands don't, and finding that out takes ten minutes.
Then fix the plumbing. Tighten your product feed. Add Product and FAQ schema. Make sure your best pages answer the questions buyers actually type into an assistant.
I made the mistake of ignoring a new channel once already. I sat out Facebook ads while competitors ate my lunch, and I paid for that lesson in real money. I'm not sitting this one out. Neither should you.
The window is open because almost nobody is doing this well yet. That is exactly why it's worth your time.
Where WRKNG Digital comes in
We build the AI discovery channel for Shopify and DTC brands. Clean feeds, structured data, answer-ready content, the full setup that makes your store easy for ChatGPT, Perplexity, and Google AI to recommend.
We don't promise specific results. We tell you exactly what our team will do and then we do it.
See how agentic commerce acquisition works for your store: wrkngdigital.com/agentic-commerce-landing-page.
FAQ
What is Meta ads fatigue for DTC brands?
It's the point where rising CPMs and shrinking audiences push your customer acquisition cost up while your return on ad spend stays flat or drops. You spend more every quarter to get the same number of orders.
Is AI discovery a real acquisition channel yet?
It's early but growing fast. Adobe Analytics reported traffic from generative AI sources to US retail sites grew more than 1,200% year over year heading into the 2024 holiday season. It won't replace paid social in 2026, but you can start building for it now.
How do Shopify stores show up in ChatGPT and Perplexity?
AI assistants pull from crawlable product data, structured schema, and third-party sources like reviews and roundups. Clean product feeds, Product schema markup, and real editorial mentions make your store easier for a model to cite in a recommendation.
Should I stop running Meta ads and go all in on AI discovery?
No. Meta still moves volume. Keep paid social running while you build an AI discovery channel in parallel, so you own an acquisition path that doesn't get more expensive every quarter.
