Retargeting Campaign
A retargeting campaign serves paid ads to people who already interacted with your brand, such as visiting a page or starting a quiz but not converting.
Key takeaways
- An audience is a list with a membership rule and an expiry window.
- Behaviour-based lists outperform page-based ones because they describe intent.
- A shorter membership window converts better but drains the pool faster.
- Sequenced ads assume the previous message was seen and add something new.
- Only a holdout test separates retargeting's influence from returning intent.
In depth
A retargeting audience is a list with a membership rule and an expiry. Someone triggers an event on your site, the platform adds their identifier to the list, and membership lapses after the number of days you set. Lists can be built from any tracked event, which is why the useful ones describe behaviour rather than pages: reached question three, viewed pricing, opened the booking page. Ad delivery then bids only on people currently on the list.
Two forces pull against each other. A narrow, recent list converts best but empties quickly, so the same budget hits each person more often. A wide list lasts longer and costs more per conversion. The membership window is the dial that balances them. Bid levels should follow the same logic: pay more for someone who abandoned a form yesterday than for a visitor from three weeks ago. Rising costs in a retargeting campaign usually mean the pool shrank, not that the ads got worse.
The practical work is sequencing. Rather than one ad shown until it wears out, plan a short series where each ad assumes the previous one was seen: a reminder, then an objection answered, then a deadline or an incentive. Cap frequency per person per week and exclude anyone who completed the action. Where the funnel ends in a scored quiz, split the list by how far a person got, because someone who answered eight questions needs a different nudge than someone who bounced on the first.
Retargeting reports flattering numbers because it advertises to people who already intended to return. Some share of those conversions would have happened with no ads at all, and only a holdout test reveals how large that share is. The technique also depends on identifiers that browsers increasingly restrict, so lists built from third-party signals decay faster than they used to. And it cannot rescue a weak offer: showing the same unconvincing page again rarely changes the answer.
Example in practice
How to measure it
Read frequency and audience size together before anything else. Frequency divided into reach tells you how hard the campaign is working each person, and a climbing figure with flat conversions means fatigue. Then compare conversion rate by list: recent abandoners against older visitors, deep engagers against shallow ones. The gaps between those rates tell you where to move budget next week.
For the honest number, hold out a random share of each audience and show them nothing. Compare conversion rates between the held-out group and the treated group; the difference is the campaign's real contribution. Ten percent is usually enough to read within a month on a decent-sized list. Without that comparison, a retargeting report describes correlation and calls it performance.
Common mistakes
Forgetting the exclusion list is the most expensive oversight. Customers who already bought keep seeing ads to buy, which annoys them and quietly consumes the budget meant for people still deciding. Build a converters audience, exclude it from every retargeting ad set, and refresh it automatically rather than by hand. Check the exclusion is actually applied after every campaign duplication, since copied ad sets often lose it.
The other habit is retargeting everyone who touched the site, including people who arrived from an unrelated blog post and left in four seconds. That pool is large, cheap to build and worth almost nothing. Require a minimum depth of engagement before someone enters the list: time on page, scroll depth, or a specific step reached. A smaller list of genuinely interested people costs less to serve and converts far better.
Frequently asked questions
How is retargeting different from a regular ad campaign?
A regular campaign targets cold audiences based on interests or demographics, while retargeting only reaches people who already visited your site or engaged with your content. This warmer audience typically converts at a higher rate and lower cost per acquisition.
Do I need a tracking pixel to run retargeting?
Yes, you install a pixel or tag from the ad platform on your pages so it can build an audience of past visitors. Without it, the network has no way to know who interacted with your funnel.
What is a good audience to retarget for a quiz funnel?
Quiz starters who did not finish and result-page viewers who did not become leads are two of the highest-intent segments. They have shown clear interest, so reminding them of their unfinished score often recovers conversions cheaply.
What share of my budget should go to retargeting?
Enough to reach the warm pool a few times a week, and no more. Because these audiences are small, they saturate quickly, and extra budget buys repetition rather than reach. Work out the pool size, decide an acceptable weekly frequency, and calculate the spend that produces it. Any surplus belongs in prospecting, which is what refills the pool.
How long should someone stay in a retargeting audience?
Until the intent that put them there has plausibly expired. For a same-week decision, a few days; for a considered purchase, thirty days or more. Rather than guessing, look at how long converters actually took between first visit and conversion, and set the window slightly beyond that. Keep separate ad sets per window so the freshest people can be bid on harder.
How do I stop retargeting ads from annoying people?
Cap frequency, rotate creative, and enforce an end. Set a limit on impressions per person per week, change the ad every two or three weeks, and remove people from the list once the window closes rather than following them for months. Also exclude converters immediately. The complaint people voice is not that they saw an ad, but that they saw the same ad forever.
Does retargeting still work now that browsers block third-party cookies?
It works, but with smaller and shorter-lived lists. First-party approaches hold up better: server-side event collection, hashed email lists uploaded from your CRM, and platform-owned audiences such as video viewers or people who engaged inside the app. Expect a share of your visitors to be unreachable, and treat the reported audience size as a floor rather than the true number.
Should retargeting ads offer a discount?
Only after the cheaper reasons have failed, because a discount trains people to abandon and wait. Start with a reminder, then address the objection the page did not answer, and hold the incentive for the last ad in the sequence or for the oldest segment of the list. If discounted retargeting is the only thing that converts, the problem is upstream in the offer.
Can I retarget the same person across different ad platforms?
Only indirectly, because each platform maintains its own audiences and cannot read another's. What you can do is fire the same events to every platform's tag so each builds a comparable list, or upload the same hashed email list to each. Match rates differ, so the audiences will not be identical, and someone reachable on one network may be invisible on another.