If your Meta retargeting feels more expensive than it used to, you are not imagining it. Cost per acquisition rises. Frequency climbs faster than conversions. Advantage+ keeps spending, but the incremental lift flattens. Your custom audiences shrink and your lookalikes fatigue faster than they once did.

For years, retargeting was the reliable safety net. Install the pixel, build a 30-day audience, and let conversion campaigns mop up demand. That worked when manual controls did more of the work and the platform leaned on advertisers to define who to chase.

That setup is now weaker, and broad pageview retargeting is the clearest example of why.

What changed underneath retargeting

In late 2024, Meta rolled out Andromeda, a rebuilt ad retrieval system, as part of a broader shift toward AI-driven delivery. Meta's own engineering blog documents this. The practical takeaway is that Meta's system is now very good at finding the people most likely to act, and it leans on broad targeting and creative far more than on the manual audience pools advertisers used to build by hand.

Pageview retargeting was designed for the older model. It assumes a website visit equals intent, so you scoop up everyone who landed on a page and ask Meta to optimize against that pool. When the platform was doing less of the finding, that was a reasonable thing to hand it. Now that the platform does more of the finding, the quality of what you hand it matters more, and a broad pageview pool is a noisy place to start.

Why a website visit is a weak starting signal

Website traffic is mixed by nature. A 30-day pageview audience blends competitors checking your positioning, students doing research, existing customers logging in, accidental clicks from broad prospecting, and casual browsers who arrived from social. You are bundling all of that into one audience and asking Meta to treat it as a buying signal.

Two problems follow from that.

First, browsing is not buying. A single product page view, a one-page exit, or a social-driven click looks very different from someone who compares multiple tiers, reads documentation, and returns within a few days. A pageview pool cannot tell those people apart, so the starting audience you hand Meta is diluted with people who were never close to buying.

Second, the window is usually too wide. Buying interest fades fast. Someone who researched yesterday is in a different place than someone who visited once three weeks ago, yet a 30-day or 60-day window treats them the same. The longer the window, the more stale visitors you fold into the starting point.

None of this means the platform is failing you. It means you are starting it from a wide, mixed pool when it would do better with a narrower, more relevant one.

The compounding cost

Pageview inefficiency rarely stays contained. When you build lookalikes from broad website visitors, you seed your prospecting with the same diluted starting point. As those audiences age, lower-quality traffic flows back into the retargeting pool, and the next lookalike is built from that. Each cycle starts the system from a slightly worse place.

The usual response is to raise the budget, refresh creative, or pile on exclusions. Budget scales whatever the campaign is already doing. Creative cannot fix a starting audience that points at the wrong people. Better inputs can.

What replaces broad pageview retargeting

The shift is from "who visited?" to "who is actually researching the category right now?"

In practice, that means a few things. Reserve your tightest retargeting for people who showed real evaluation behavior, such as multiple sessions within a short window, pricing engagement, or repeated product comparison, and let broad 30-day visitors sit in a lower-priority nurture tier rather than defining your performance structure. Keep those buying-mode segments tight and recent, because compressed interest is what you want the system starting from. And build your lookalike seeds from those tighter, more recent groups rather than from everyone who ever loaded a page.

The goal is not to retarget more people. It is to give Meta's AI a cleaner, more current place to begin, on both retargeting and the prospecting it feeds.

How to diagnose this in your account

A few questions surface whether pageview dilution is costing you. What share of your retargeting audience had more than one session within seven days? How far back do your pricing and product page audiences reach? Are your lookalikes built from broad traffic or from tighter buying-mode groups? Is frequency rising faster than conversion rate? Does retargeting performance fall off as the audiences age?

If broad 30-day website pools dominate your structure, you are handing Meta a wide, mixed starting point and paying for it in efficiency.

How Slopeside helps

Broad pageview retargeting underperforms because it starts Meta from a wide, mixed pool. Slopeside gives the system a narrower, more current place to begin.

Slopeside delivers audiences of people actively researching in your category, drawn from real-time search and browsing behavior across a verified base of more than 300 million people, not just on-site pageviews. Audiences refresh every 24 hours and sync straight into your Meta account as Advantage+ suggestions, so your retargeting and the lookalikes it feeds both start from recent buying behavior rather than aging site traffic. You are not replacing Meta's AI. You are pointing it at the right people first.

Start a $1 trial and test a Slopeside audience against your current retargeting setup.

Frequently asked questions

Is retargeting dead on Meta? No. Broad pageview retargeting is losing efficiency. Reinforcing recent, real buying behavior is what replaces it.

Should I delete my 30-day website audiences? Not necessarily. They can work as a lower-priority nurture tier. They just should not define your high-intent performance structure.

Does Advantage+ remove the need to think about audiences? No. Custom audiences act as starting suggestions rather than hard constraints, and a more relevant suggestion gives Advantage+ a better place to begin.

How often should high-intent audiences refresh? Daily refresh is ideal. Compressed buying windows mean a starting audience goes stale quickly.