Retargeting strategy has changed more in the past three years than in the decade before. At Choco Media, we’ve watched clients pour budget into retargeting setups built for 2020 — broad audiences, long windows, the same creative running for months — and we’ve seen those setups quietly drain cash with almost nothing to show for it. The good news is that a leaner, more deliberate retargeting strategy still works in 2026. The bad news is that it requires a different mental model than what most paid media guides describe.
This post is for anyone managing or overseeing paid media who has noticed retargeting performance dropping off and isn’t sure whether to cut it, fix it, or rebuild it from scratch. We’ll walk through the audience segments that actually convert, the creative rotation logic that prevents fatigue, and the exclusion rules that stop you from wasting money on people who will never buy.
We’ll also be direct about what we can’t know without your specific data — retargeting is one area where generalisation breaks down quickly, so we’ll tell you what to test rather than what to assume.
Why most retargeting setups have stopped working
The shift happened in two waves. First, iOS 14.5 in 2021 broke the tracking infrastructure that retargeting depended on — pixel-based audiences shrank overnight, attribution windows became unreliable, and audience sizes that once felt manageable became too small to run meaningful campaigns against. Second, the platforms responded by pushing advertisers toward broader, more algorithmic approaches, which quietly deprioritised retargeting in favour of prospecting at scale.
The result: most retargeting pools are smaller than advertisers realise, the audiences are less accurate than they were three years ago, and the creative assets — often repurposed from prospecting campaigns — aren’t doing the specific job that retargeting creative needs to do.
- Pixel audiences have shrunk by 30–60% on Meta for most advertisers since iOS 14.5 (figures vary significantly by industry and audience age)
- 7-day click, 1-day view is now the default attribution window on Meta — longer windows inflate retargeting credit
- Broad retargeting audiences on Google often include users who visited a single page for under 10 seconds and have no real intent signal
- Creative fatigue hits retargeting audiences faster than prospecting because the pool is smaller and frequency builds quickly
None of this means retargeting is dead. It means the approach that worked in 2020 doesn’t translate directly to 2026.
The audience segments that still convert
The segments worth spending on in 2026 share one characteristic: they represent meaningful intent, not just a visit. A bounce after two seconds is not intent. A cart abandonment, a product page viewed three times, a pricing page visit — these are different.
High-intent behavioural segments
- Cart and checkout abandoners (0–7 days): Still the highest-converting retargeting segment for e-commerce. The window matters — beyond 7 days, conversion rates typically drop sharply. Keep the audience tight.
- Product page repeaters (0–14 days): Users who visited the same product or service page more than once. This requires setting up a custom event or URL-based audience, but the intent signal is strong.
- Pricing or “how it works” page visitors (0–7 days): For B2B and SaaS, these pages indicate late-stage research. Short window, tight targeting.
- Video viewers — 75%+ completion (0–14 days): If you run video in prospecting, the people who watched most of it are a warm retargeting segment worth isolating.
- Email list custom audiences: First-party data. Upload your email list and target subscribers who haven’t purchased. The audience is usually small but the signal is clean.
Segments to deprioritise or cut
- All-site visitors with no time-on-page or page-depth filter — too broad, too much noise
- 30-, 60-, or 90-day windows as primary retargeting audiences — intent has degraded significantly
- Lookalikes of retargeting audiences — this is prospecting, not retargeting; treat it as such
We typically see retargeting budgets perform best when the audience is small and specific rather than large and approximate. A 2,000-person cart abandoner list will outperform a 200,000-person all-visitor list almost every time — if the creative is right.
Creative rotation logic for retargeting
Frequency is the quiet killer in retargeting. When the same person sees the same ad five times in a week, the sixth impression doesn’t just stop working — it starts generating negative associations with your brand. Small audiences and high budgets make this happen faster than most advertisers expect.
The rotation framework we use
We structure retargeting creative in three phases, each addressing a different psychological moment:
- Phase 1 — Re-engagement (days 0–3): Direct, product-specific. “You looked at this. Here it is again.” Simple, clear, no clever copy. The goal is to get back in front of them while the memory is fresh.
- Phase 2 — Objection handling (days 3–7): Address the most common reasons people don’t convert. For e-commerce this might be returns policy, shipping cost, or social proof. For B2B it might be implementation complexity or pricing structure.
- Phase 3 — Last-chance or urgency (days 7–14): Scarcity, limited-time offer, or simply a different angle. Don’t manufacture false urgency — if there’s no genuine hook, skip this phase and let the audience exit the funnel.
Rotate between at least 3 creative variants per phase and set frequency caps at the ad set level. On Meta, we typically cap at 2–3 impressions per week for retargeting audiences under 10,000 people. On Google Display, the equivalent is roughly 3–5 impressions per week per user.
For more on building the creative infrastructure that makes this sustainable, see our post on creative-led paid media and UGC ad engines.
Exclusion rules: where retargeting budgets quietly disappear
Exclusions are the part of retargeting setup that gets skipped most often. They’re not glamorous, but they’re where a significant portion of wasted budget hides.
Essential exclusions for every retargeting campaign
- Recent converters: Anyone who purchased, signed up, or completed your goal in the last 30–60 days. Without this exclusion, you’re spending money to tell customers to buy something they already bought.
- Current trial or onboarding users (SaaS/B2B): These users are in a sensitive phase. Retargeting them with acquisition messaging can feel pushy and undermine the onboarding experience.
- Unsubscribes and suppression lists: Upload your email suppression list as a custom audience and exclude it. These are people who have actively opted out of your marketing.
- High-frequency, zero-conversion users: If a user has seen your retargeting ads 15+ times without converting, they’re not going to. Exclude anyone who has hit a frequency threshold without a conversion event.
- Existing customers for acquisition-focused creative: If your retargeting creative is acquisition-focused (discounts for new customers, first-purchase offers), exclude your existing customer list entirely.
We cover how to set up clean conversion tracking that makes these exclusions reliable in our piece on attribution in 2026 and server-side tracking.
Budget allocation within retargeting
The question we get most often is how much budget to put into retargeting versus prospecting. There’s no universal answer, but the rough principle we apply is: retargeting should never exceed 20–25% of total paid media spend for most businesses, and for businesses with small audiences or long sales cycles, it should be lower.
Why over-indexing on retargeting is a trap
- Retargeting can only convert the people prospecting already found. If prospecting is underfunded, you have a small and shrinking retargeting pool.
- Attribution models tend to credit retargeting heavily because it touches users close to conversion — but those users may have converted anyway.
- Small retargeting audiences + large budgets = high frequency + creative fatigue + wasted spend. The math doesn’t work above a certain threshold.
Within retargeting budget, we prioritise in this order: cart/checkout abandoners first, high-intent page visitors second, longer-window or broader audiences third (and often not at all, if budget is limited).
Our paid media service includes a full audit of how retargeting budget is currently allocated before we make any changes — because what the numbers say and what’s actually happening are often different.
Platform-specific considerations in 2026
Retargeting behaves differently across platforms, and the gaps between them have grown wider.
Meta (Facebook/Instagram)
Advantage+ audiences have gradually absorbed manual retargeting setups for advertisers who don’t actively resist it. If you’re running Advantage+ Shopping Campaigns, retargeting audiences are handled algorithmically — you can’t fully control them. For retargeting-specific setups, you need to use manual campaign structures with custom audiences, and you need to monitor them carefully for audience overlap with prospecting campaigns.
- Use the 7-day click, 1-day view attribution window as your baseline
- Set up a dedicated retargeting campaign, not ad set — this gives you cleaner budget control
- Monitor audience overlap between retargeting and prospecting ad sets using Meta’s Audience Overlap tool
Google Ads
RLSA (Remarketing Lists for Search Ads) remains underused and often outperforms display retargeting in terms of intent quality. If someone searched for your target keyword, visited your site, and is now searching again — that’s a high-value signal. Bid adjustments on RLSA for high-intent visitors can be effective without requiring large audience sizes.
- Use Customer Match with first-party data where possible — it’s more reliable than pixel-based audiences post-iOS
- Display retargeting on Google requires frequency cap management at the campaign level
- YouTube retargeting to high-intent visitors can work for brands with strong video creative
Measuring retargeting honestly
Retargeting is one of the most attribution-inflated channels in paid media. Because it touches users who are already warm and close to converting, last-click models systematically over-credit it. A user who was going to convert anyway — because they already decided to buy — shows up as a retargeting win.
The measurement approaches we use to get a more honest picture:
- Holdout testing: Exclude a random 10–20% of the retargeting audience from seeing ads. Compare conversion rates between the exposed and holdout groups. The difference is the true incremental lift from retargeting.
- Incrementality experiments: Meta offers lift studies for larger accounts. Google has similar tools. These are more resource-intensive but give a cleaner read on whether retargeting is actually driving conversions or just claiming credit.
- View-through conversion scrutiny: View-through conversions in retargeting are almost always overstated. We recommend treating them as directional signals only, not hard performance metrics.
If you run a holdout test and find the conversion rate difference is small, that’s valuable information. It may mean retargeting is claiming credit for organic conversions. It doesn’t necessarily mean you should cut retargeting entirely — brand presence matters — but it should change how you allocate budget.
When to rebuild rather than optimise
Sometimes the right answer isn’t adjusting the existing setup — it’s starting over. The signals that tell us a retargeting account needs a rebuild rather than a tune-up:
- Frequency above 8–10 across the retargeting audience with no conversion rate improvement from creative changes
- Audience sizes below 1,000 people on Meta — too small to run effectively
- No exclusion lists in place, ever — the historical data from this account is not reliable
- Attribution window set to 28-day click — all the numbers are inflated and incomparable to current benchmarks
- The same creative has been running for more than 90 days with no rotation
If three or more of these are true, we typically recommend a clean rebuild: fresh campaign structure, new creative, properly configured exclusions, and a holdout set up from day one so you have a baseline.
If you’d like a second opinion on your current retargeting setup, get in touch — we’re happy to take a look before you make any changes.