Retargeting and remarketing

Building lists worth retargeting

Retargeting quality starts with the list, and most accounts only have one: all website visitors. That pool mixes someone who bounced off the blog in four seconds with someone who abandoned a full basket. Showing both the same ad wastes money on the first and under-serves the second.

The lists worth building are the ones that mean something. Viewed a specific product or service page. Started checkout and stopped. Spent more than a threshold time on site. Watched a meaningful share of a video. Opened a lead form and abandoned it. Each of these describes a different person with a different objection.

Sources go beyond the website. Customer lists uploaded from your CRM, email subscribers who never bought, social engagers, video viewers and app users all make usable audiences. The mistake is building these only when you need them. Lists take time to fill, so set them up now and they are ready when you want them.

• Segment by page and action, not just by visit
• Build cart and form abandonment lists separately, they are the highest value
• Upload customer and subscriber lists as both audiences and exclusions
• Create lists before you need them, they fill over weeks

Segmenting by intent depth

The useful way to organise retargeting is by how close someone got to buying, then match both the message and the budget to that. Someone who abandoned a basket needs a nudge and possibly a reason to act now. Someone who read one blog post needs a reason to care at all.

A workable three-tier split looks like this. High intent covers cart abandoners, pricing page visitors and form starters, and it gets the highest bids and the most direct ads. Mid intent covers product and service page visitors, and it gets the case for choosing you. Low intent covers blog readers and brief visitors, and it gets content or a soft offer rather than a purchase request.

Exclusion between tiers is what makes this work. Without it, your cart abandoner sits in all three audiences and receives three different messages from three campaigns competing against each other. Each tier should exclude the tiers above it, so everyone sits in exactly one.

• High intent: cart, pricing and form abandoners, direct offer, top bids
• Mid intent: product and service page visitors, the case for choosing you
• Low intent: brief visitors and readers, useful content or a soft offer
• Exclude upward so nobody sits in two tiers at once

Recency windows and why they matter

Intent decays. Someone who visited yesterday is far more likely to buy than someone who visited five weeks ago, and treating both the same means overpaying for the second. Recency windows let you split a single audience by how long ago someone visited and bid accordingly.

The right window follows the buying cycle. An impulse purchase might justify a one to three day window with everything after that treated as cold. A considered B2B purchase can reasonably run a ninety day window because that is how long the decision genuinely takes. Look at your own time to purchase data rather than guessing.

A practical approach is tiered windows within the same audience. Days one to three get the highest bid and the most direct message. Days four to fourteen get moderate bids. Days fifteen to thirty get low bids and lighter frequency. Beyond that, most audiences are better left alone.

Frequency caps and burn pixels

Frequency caps are what separate retargeting from harassment. Without a cap, a small audience and a normal budget will show the same person your ad dozens of times a week, which produces irritation, complaints and eventually ad blindness. A cap of a few impressions per person per day is a reasonable starting point, tightened if the audience is small.

Burn pixels are the other half. A burn pixel fires on the confirmation or thank-you page and removes that person from the retargeting audience. Without one, you keep paying to advertise a product to somebody who already bought it. This is the single most common retargeting fault we find, and it is both the most wasteful and the most annoying to the customer.

Exclusions should go wider than the confirmation page. Exclude existing customers from acquisition retargeting. Exclude current subscribers from subscription offers. Exclude people who have seen the campaign a set number of times without acting, because at some point the answer is no and continuing to ask costs money.

• Cap impressions per person per day, tighter on small audiences
• Fire a burn pixel on confirmation and remove buyers immediately
• Exclude existing customers from acquisition campaigns
• Set an overall exposure limit, after which you stop asking

Sequencing messages across a longer decision

The best retargeting tells a story in order rather than repeating one ad. The first exposure reminds them who you are. The second handles the most common objection. The third brings proof, whether a review, a demonstration or a comparison. The fourth makes a direct offer.

This works because the reason someone did not buy is rarely that they forgot. It is usually a specific doubt: price, trust, timing, or uncertainty about whether it fits their situation. Sequenced retargeting answers those doubts one at a time instead of shouting the same thing louder.

Building it needs the platform features that support ordering, which most now have in some form, and it needs enough audience size for the later steps to get a real run. On small audiences a simpler two-step sequence works better than an elaborate five-step one that nobody reaches the end of.

Running retargeting across platforms without double paying

The same person can be retargeted on Google Display, Meta, YouTube and the open web at once, and each platform will happily claim the conversion. Run these without coordination and you both overspend and misread the results, because your reported returns add up to more conversions than you actually had.

Coordination means deciding what each platform is for. Meta and Instagram tend to suit visual reminders and social proof. Google Display and the open web give reach and cheap frequency. YouTube suits demonstration. Search retargeting, where you adjust bids for previous visitors when they search again, is often the quietest performer and the most profitable.

For measurement, pick one source of truth outside the ad platforms, usually your analytics or your order system, and reconcile against it. Then judge retargeting as a whole rather than platform by platform. The honest question is what total retargeting spend produced in total incremental revenue, and answering it usually leads to consolidating spend rather than spreading it.

• Give each platform a defined job rather than running the same ad everywhere
• Set a shared exposure budget per person across platforms
• Reconcile against one source of truth outside the ad dashboards
• Test by holding out a group, the only clean way to measure incremental value

Consent, privacy and the shrinking cookie

Retargeting depends on recognising people across visits, and the mechanisms for that have narrowed. Third-party cookies are restricted or blocked in most browsers, tracking prevention limits how long first-party identifiers survive, and consent requirements mean a share of visitors are never added to a list at all.

The practical response is to lean on first-party data. Customer lists, logged-in identifiers, email subscribers and server-side event tracking hold up better than browser cookies. Audiences built from people who gave you their details directly are both more durable and more accurate.

It also means setting expectations honestly. Retargeting audiences are smaller than they were, and reported numbers understate what happened. That is not a reason to skip it, since retargeting remains one of the cheapest conversion sources available. It is a reason to build lists on data you own and to judge results on total business outcomes rather than platform reporting alone.