Remarketing is advertising aimed at people who have already interacted with your business. Someone visits a product page, watches most of a video, opens your app or joins your email list. Later they see your ad in a social feed, on a news site or in a YouTube pre-roll. Google calls it remarketing; most other platforms and practitioners say retargeting, and the two words mean the same thing.
The audiences come from tracking. A tag such as the Meta Pixel or the Google tag records a visit, the platform matches the visitor to a logged in user, and that person joins a list you can target or exclude. Customer lists work the other way around: you upload hashed emails or phone numbers and the platform finds the matching accounts.
It works because a first visit rarely ends in a sale. People compare, get distracted, wait for payday or forget the name of the site. Put a reminder in front of the right people at a modest cost, and you win back a share of visits that would otherwise be gone for good.
Mechanics
How remarketing audiences are built and served
It all starts with a rule. Say you want everyone who reached a cart page in the last 14 days but never hit the order confirmation page. You define that audience once, and the platform fills it as the tag fires. Membership expires after the window you set, so the list refreshes itself and people who bought drop out through the exclusion.
Bidding then runs like any other campaign, except that the pool is small and every advertiser who touched the same visitor is bidding for the same attention. So a thousand impressions usually cost more on a warm audience than on a cold one, and frequency caps matter: without them, a handful of people see the same ad dozens of times.
Browser privacy changes have thinned these lists. Third party cookies are blocked in several browsers, tracking prevention on Apple devices limits how long identifiers persist, and consent banners remove visitors who decline. Server side tracking through a Conversions API and first party customer lists now carry much of the load that a simple pixel used to.
Example
A worked example for an online shoe store
Assume a running shoe store gets 30,000 visitors a month and 2 percent buy on the first visit, so 600 orders. That leaves 29,400 people who left, of whom 3,000 put a pair in the cart and abandoned it.
The store targets those 3,000 cart abandoners with a Meta campaign, keeping each one in the audience for seven days. It spends 1,500 in the month, and 150 of them come back and complete an order, a 5 percent recovery rate. At an average order of 120, that is revenue of 18,000 against spend of 1,500, a cost per recovered order of 10.
The honest question is how many of those 150 would have returned anyway. If a holdout test shows that 60 of them would have bought without seeing an ad, the campaign really added 90 orders, and the cost per incremental order is about 17. That's still a good result, but it isn't the 10 the platform reported.
Use
When remarketing earns its budget and when it does not
- It earns its budget when the site has enough traffic to fill a list, when the purchase decision takes days rather than seconds, and when the ad offers a reason to return.
- It works well for cart abandoners, pricing page visitors, demo watchers and past customers who are due to reorder, since each group has shown clear intent.
- It misleads when it takes credit for people who would have come back on their own. An occasional holdout group is worth the few sales it costs.
- It is a poor fit for a business with a few hundred visitors a month, because the lists are too small to serve and too small to learn from.
- Judge it by incremental orders and by the conversion rate of returning visitors, not by the platform's reported conversions alone.
Watch out
Common mistakes with remarketing
- Following buyers around for weeks with ads for the product they already bought, because nobody set up an exclusion for the confirmation page.
- Showing the same generic banner to every visitor instead of matching the ad to the page or product category the person actually looked at.
- Running with no frequency cap, so a small audience sees the ad so often that the brand becomes an irritation rather than a reminder.
- Reading platform reported conversions as incremental, when many of those people came back through a search or an email without noticing the ad.
- Building lists around sensitive categories, such as health conditions or financial hardship, which breaks platform rules and leads to disapproved campaigns.
FAQ
Questions about Remarketing
Is remarketing worth it for a small business?
It can be, provided your site gets enough visitors to build audiences of useful size and your offer has a reason for people to return. A business with little traffic may find the lists too small to serve. Compare remarketing results against a holdout or past sales to check that the ads bring back buyers who would not have returned anyway.
How often should people see my remarketing ads?
There is no fixed number; it depends on how long your customers usually take to decide and how varied your creative is. A short buying cycle calls for a brief window and a modest cap, while a considered purchase may justify a longer window. Watch frequency reports, and rotate ads when click rates fall.
What is the difference between remarketing and retargeting?
In everyday use the two words mean the same thing: showing ads to people who have already interacted with your business. Google has historically called it remarketing, while many other platforms and marketers say retargeting. Some people use remarketing more narrowly for email follow ups, so clarify the meaning when briefing a team.
