Blog · AI
— AI··11 min read

The AI content audit: how to find and fix pages that are invisible to AI answers

Joona Heinonen· Choco Media · Rovaniemi

Retargeting ads strategy hasn’t fundamentally changed — you show ads to people who already know you, and those people convert at higher rates than cold audiences. But the mechanics have shifted enough since iOS 14 that many accounts we audit are running retargeting setups built for a world that no longer exists. At Choco Media, we’ve rebuilt how we think about retargeting from the ground up, and this post walks through the current approach: audience construction, creative fatigue management, and the creative patterns that actually move warm audiences toward a decision.

This is for anyone running paid media on Meta or Google who has a retargeting campaign live but suspects it’s either reaching too few people, burning out on the same creative, or working in ways that are hard to attribute. All three are common. All three are fixable.

What you’ll leave with: a working framework for sizing retargeting audiences in a post-cookie, post-ATT environment, a rotation schedule that kills creative fatigue before it kills performance, and a set of creative patterns we’ve tested repeatedly with warm audiences.

Why most retargeting setups underperform in 2026

The classic retargeting setup — pixel fires, user lands in a 30-day website visitor audience, sees three variations of the same ad for two weeks — worked reasonably well when pixel data was accurate and audiences were large enough to matter. iOS 14.5 and subsequent ATT changes removed roughly 40–60% of observable iOS events from Meta’s pixel, depending on your site traffic and audience composition. What this means practically: your “30-day website visitors” audience is a fraction of the people who actually visited.

The result is that many accounts have retargeting audiences that are either too small to exit the learning phase, or so narrow that the algorithm hammers the same people daily. Frequency rises, performance falls, and the temptation is to broaden the audience — which then dilutes the retargeting signal entirely.

The fix isn’t to abandon retargeting. It’s to rebuild the audience logic around first-party signals and to size the campaign budget to the actual observable audience, not the audience you had in 2021.

First-party data is the new retargeting foundation

If pixel-based audiences have shrunk, the replacement is data you actually own: email lists, CRM exports, phone numbers, and server-side event data. Meta’s Conversions API (CAPI) and Google’s enhanced conversions allow server-side event matching that partially compensates for pixel loss. But the more durable foundation is uploading your customer and lead lists directly.

What to upload and how often

The match rate on email-based custom audiences typically sits between 40–70% depending on how clean your list is and what email service your customers use. Even at 50% match, a 2,000-person email list gives you a 1,000-person audience — often larger than what the pixel is observing from website traffic.

Server-side tracking via CAPI doesn’t replace list uploads but improves pixel match rates significantly. In our implementation experience, adding CAPI alongside the browser pixel typically recovers 20–30% of previously unattributed events. That’s worth doing regardless of retargeting, but it directly affects retargeting audience size.

The retargeting question has shifted from “who visited my site?” to “who do I already know, and what do I know about them?” First-party data answers that question. Pixel data alone increasingly can’t.

Audience sizing: the minimum viable retargeting pool

There’s a practical floor below which retargeting becomes more harm than help. On Meta, audiences under 1,000 people struggle to exit the learning phase. Google Display has similar dynamics, though the threshold is different. The right move when your retargeting audience is below threshold isn’t to run it anyway — it’s to either expand the definition or fold retargeting budget into prospecting and use it there.

Expanding the audience without losing signal quality

The goal is to reach a combined retargeting pool of at least 5,000 people before you allocate meaningful budget. Below that, performance is too volatile to optimize against.

Creative fatigue: what it is, when it happens, what to do

Creative fatigue in retargeting happens faster than in prospecting because you’re showing ads to the same small group of people. A cold audience is enormous — millions of people who have never seen your ad. A retargeting audience might be 8,000 people. If you’re spending €500/week against 8,000 people, you’re hitting average frequencies of 3–5 per week within days. At that frequency, people start consciously or unconsciously tuning out your creative.

The early warning signals are specific and consistent across accounts:

The rotation schedule that prevents fatigue

Rather than waiting for signals of fatigue, we run a proactive rotation. The principle: no single creative should run against a retargeting audience for more than 3–4 weeks without a meaningful variant introduced. “Meaningful” means a different hook, a different format, or a different proof type — not just a different colour on the same layout.

If frequency is climbing fast because your audience is small, the rotation schedule accelerates. A 2,000-person audience might need new creative every two weeks. A 50,000-person audience might sustain the same creative for 8–10 weeks. Audience size and rotation cadence are directly linked.

What creative actually converts warm audiences

Warm audiences behave differently from cold ones. They’ve seen your brand, they know what you do at some level, and they made a decision not to convert — yet. Your job with retargeting creative is to address that “not yet.” That means different creative strategy than prospecting.

Creative patterns that work for retargeting

The creative mistake we see most often in retargeting: running the exact same creative and offer as the prospecting campaign. The prospecting creative was designed to introduce and hook. The retargeting audience doesn’t need an introduction. Running it anyway wastes frequency and budget on a message that’s already been seen.

Attribution in a retargeting context

Retargeting attribution is where reporting gets complicated. Someone sees your retargeting ad, doesn’t click, then converts via a direct visit two days later. Meta’s view-through attribution window claims that conversion. Google Analytics sees it as direct. Your actual retargeting ROAS might look very good in Meta’s dashboard and confusing everywhere else.

This is especially pronounced for retargeting because the audience is already warm — many of them were going to convert anyway, via email, organic search, or direct. The retargeting ad may have contributed or it may have simply been present.

For a deeper look at how we approach attribution across channels, our post on server-side tracking and GA4 attribution covers the implementation side in detail.

Budget allocation: how much to put into retargeting vs. prospecting

The general guidance we use with clients: retargeting should represent 15–25% of total paid media budget, not more. Above 25%, you’re typically over-investing in an audience that’s already warm and under-investing in filling the top of the funnel that feeds retargeting. The ratio isn’t fixed — it shifts based on funnel health and campaign objectives — but it’s a useful sanity check.

Signs you’re over-investing in retargeting

Signs you’re under-investing in retargeting

Our paid media service includes an initial audit that checks these ratios as part of account setup, because the split between prospecting and retargeting is one of the highest-leverage things to get right early.

Platform-specific notes: Meta vs. Google for retargeting in 2026

Meta and Google retargeting work differently in ways that matter for how you set them up.

On Meta, the retargeting audience system uses Advantage+ features that increasingly automate who within your defined audience sees your ads. This is mostly helpful — Meta’s algorithm is good at finding the highest-probability converters within a warm pool — but it reduces the precision you have about exactly who is seeing what. The practical response: define your audiences carefully (don’t accidentally include your prospecting audiences in retargeting), and let the algorithm work within those guardrails.

On Google, retargeting (called “remarketing”) works through both Search and Display. Search remarketing (RLSA — Remarketing Lists for Search Ads) is particularly valuable: you can bid higher on search queries from people who have already visited your site, or show ads only to that warm audience. This is often underused. Display retargeting on Google is lower intent but useful for keeping the brand visible during a consideration window.

The retargeting audit checklist we use on new accounts

When we take on a new client with existing paid media, retargeting is one of the first areas we check. The issues repeat so consistently that we’ve turned the diagnostic into a standard checklist.

This takes about 20 minutes per account and usually identifies at least two or three meaningful issues. If you’re running paid media for a client or your own business and want a second set of eyes on the retargeting setup, we cover this in our conversion rate optimization work and as part of paid media audits.

If you’d like to talk through your current retargeting setup or get a quick read on whether the audience logic and budget ratio make sense, reach out here — we keep the first conversation short and practical.

← All storiesNext story →
— Free tips, monthly

Get the playbook, for free.

One short letter a month — the prompts we use, the campaigns that worked, the AI tools worth the time. No sales pitch, just field notes.

— Want us to do it for you?

Hire the agency.

AI-accelerated content, paid media, brand and web — delivered by one small team that talks to itself. Currently taking on a handful of clients each quarter.

Book a call