Facebook Ads: iOS 14 CPM Surge in 2026

That recent IAB report saying 78% of advertisers can’t measure their campaigns right after the iOS 14 changes? It’s just confirming what we all feel in the trenches. Attribution is a mess, which shoves all the focus onto metrics we can actually see, like engagement on our Facebook ads. This means marketers have to completely rethink their creative and targeting. So, what’s the actual plan for making money in this new reality?

The 20% CPM Surge: Efficiency Is Now The Whole Game

Apple’s App Tracking Transparency (ATT) framework in iOS 14 hit costs almost immediately. Industry data, including reports from eMarketer, showed that advertisers were suddenly paying about 20% more for their cost per thousand impressions (CPM) on Facebook and Instagram. For a lot of businesses, that’s a fundamental change to their unit economics. When your cost to reach people jumps that high, every single impression has to pull more weight. My take on this is simple: you can no longer afford to burn money on mediocre ads shown to vaguely defined audiences. The strategy of using so-so creative and hoping for the best is dead. Now, you have to carefully build every ad to stop the scroll and force a reaction. This means you have to be constantly running A/B tests on your headlines, your images, and your CTAs. These are the small optimizations that, when added up, can actually counteract that 20% CPM hike by making each impression do more work. If your creative is weak, you’re just paying a premium to get ignored.

Clawing Back Data with CAPI: The 15% Matching Boost

Meta’s big technical answer to iOS 14 was pushing everyone toward server-side event tracking with the Conversions API (CAPI). The old Meta Pixel sends data from the user’s browser (client-side), but CAPI lets you send event data directly from your server to Meta’s, getting around a lot of the new browser and device restrictions. Meta’s own data shows that setting up CAPI can improve your data matching by up to 15% compared to just using the pixel alone. This is how we start to get back some of the attribution clarity we lost. That 15% data matching improvement isn’t just a number. It means your Custom Audiences are more accurate, your segmentation is sharper, and Meta’s algorithm can actually optimize campaigns for purchases again. Without this data, Meta’s optimization algorithms are just guessing. I’ve seen it with my own clients, the ones who put in the work to get a CAPI implementation running properly recovered a big chunk of their conversion data. It won’t get you back to pre-iOS 14 visibility, but it’s the best tool we have. Not setting it up is just leaving conversion data, and money, on the table.

Short-Form Video’s Takeover: A 30% Engagement Bump

It’s no secret that people are glued to short-form video, and that’s completely changed the game for Facebook ad engagement. Platform data from Meta and others shows over and over that short-form video campaigns (under 15 seconds) get 30% higher engagement rates than old-school static image ads. This is about cramming information and an emotional hook into a tiny window of time. You can pack more personality, show off more product features, and build more brand feel in a 10-second video than a static image can ever hope to. All that extra engagement signals to the algorithm that your ad is relevant, which helps lower your costs and improve delivery. If you’re still just running static images, you’re basically invisible in the feed. You have to start thinking like a content creator, not a print advertiser. What can you show in ten seconds that makes someone stop scrolling, watch the whole thing, and actually want to click? The content has to be dynamic, feel real, and have a single, clear message.

Constant Testing Can Net a 10% CTR Lift

With targeting signals so much weaker now, your creative has to do all the heavy lifting. A disciplined approach to A/B testing, where you’re running at least three distinct creative variations, can easily produce a 10% lift in click-through rates (CTR). I’m not talking about changing button colors. I mean testing completely different angles: a problem-solution story vs. a benefit-driven one, or a user-generated content style video against a slick, polished brand spot. That 10% CTR jump isn’t some fantasy number. It’s a real benchmark I see clients hit all the time when they actually commit to testing. I still see so many advertisers running one ad set with a couple of variations for weeks on end. That’s just burning money. The algorithm needs a steady stream of new creative to figure out what works and find your audience faster, especially since it has less user data to work with. This process also teaches you exactly what your audience responds to, which is gold for your next campaign and your entire marketing strategy.

Go Broad: Let Strong Creative and the Algorithm Do the Work

This is where my advice often diverges from the old-school thinking, especially for marketers who grew up on hyper-specific targeting. A lot of people still want to build these tiny, perfect audiences with layered interests and demographics. But with the signal loss from iOS 14, I’ve found that going with broad audience targeting paired with really strong creative lets Meta’s algorithm find high-intent users much more effectively. This doesn’t mean you abandon targeting completely. It means you trust the machine learning to do its job. For instance, instead of layering on interests like “organic food,” “yoga,” and “sustainable living,” just target “women aged 25-55 in metropolitan areas” and let your ad creative, which should obviously scream those values, do the filtering. When you give the algorithm a bigger sandbox to play in, paired with ads that stop the right people, it finds patterns a human could never spot. You almost always get lower costs and find better scale this way because you’re not tying the algorithm’s hands. Bottom line: iOS 14 blew up the old playbook. The advertisers winning now are the ones who are obsessed with their creative, who have CAPI set up properly, and who are smart enough to give the algorithm broad targets and let it run to find those post-iOS 14 wins.