Retargeting, also called remarketing, is the practice of showing advertisements to people who have previously visited a website or used an application. Audiences are built from site activity, such as viewing a product, abandoning a cart, or reaching a particular page, and those individuals are then reached across other sites, social platforms, and video services.
It consistently reports the strongest performance of any paid channel, and the reason deserves scrutiny. Retargeting reaches people who have already demonstrated interest and who are, by definition, further along in their decision than a cold audience. Many of them would have returned and purchased without seeing any advertisement. Because attribution credits the last advertisement touched before conversion, retargeting collects credit for a large volume of purchases it did not cause, which is why its reported return on ad spend is typically excellent and its incremental return frequently modest.
This does not make the channel worthless, and the honest position is more nuanced than either enthusiasm or dismissal. Retargeting demonstrably does recover some abandoned sessions, particularly for considered purchases where the visitor genuinely forgot or was interrupted, and it maintains presence during longer decision cycles. What is required is measuring that contribution with a holdout rather than accepting attributed figures, which almost always reveals a smaller but still positive effect.
Segmentation determines whether the spend is efficient. Treating all past visitors as one audience wastes budget on people who bounced within seconds, on existing customers who have already purchased, and on job applicants and competitors. Meaningful segmentation distinguishes depth of engagement, recency, and stage: someone who abandoned a cart yesterday warrants different treatment from someone who read one article three months ago, and existing customers should usually be suppressed entirely from acquisition campaigns.
Frequency control is the difference between a channel that works and one that damages the brand. Without caps, retargeting will follow the same individual across every site they visit for weeks, which produces irritation, ad blocker installation, and negative brand association that no click-through metric captures. Setting frequency limits and membership durations that reflect the actual purchase cycle, rather than leaving platform defaults in place, is basic practice that many campaigns still omit.
The technical foundation has weakened considerably. Third-party cookie restrictions, tracking prevention in browsers, and mobile platform permission requirements have all reduced the size and accuracy of retargeting audiences, and the erosion is uneven across browsers and devices. Audiences built from first-party data, such as customer lists and logged-in behaviour, have become correspondingly more valuable, though they carry their own consent obligations.
Creative treatment is frequently neglected in this channel, with many programmes serving the same product image repeatedly for the duration of the membership window. Since the audience has already seen the site, repeating what they have seen adds nothing, whereas addressing the reason they did not convert, whether that is delivery cost, uncertainty about fit, or a need for comparison, gives the impression a purpose.
Because the channel sits at the boundary between advertising and on-site experience, its performance depends heavily on what happens after the click. In practice a marketing services engagement manages the audience and frequency strategy, the incremental measurement is designed with data analytics, and the underlying reasons visitors left in the first place are addressed through CRO service work, since recovering abandonment is cheaper than paying repeatedly to bring the same people back.