Retargeting

Retargeting explained

Retargeting can support a purchase decision already in progress: an ad recalls a viewed product or answers an unresolved question. A visit without a purchase is not always a lost customer. Some people return on their own; others have deliberately decided against an offer. These differences matter for the message, budget and measurement.

Common approaches include reaching previous website visitors, dynamic ads featuring viewed products and ads following an abandoned cart. Interactions within an advertising platform may also support audience creation, depending on its features. Customer Match matches eligible customer data to accounts on an advertising platform. An audience of new people with inferred interests is not, by contrast, a repeat contact with previous visitors.

A useful plan defines the relevant behaviour, time since the interaction, suitable content and exclusions. After a recorded purchase, the original purchase prompt should be reviewed or excluded. Contact frequency and campaign duration depend on the decision cycle, product and audience response. There is no universally appropriate number of ads per day. Tracking and cross-platform frequency controls also have limitations.

A high retargeting conversion rate does not establish an equally large advertising effect. The people reached already had an interest. The key question is which additional purchases or qualified enquiries the advertising causes. Suitable control groups or conversion lift tests can examine this contribution when enough data and the necessary platform features are available. Evaluation should include costs, margin, the comparison period and uncertainty in the result.

Before deployment, establish the legal basis, required consent, transparent information and applicable platform rules. An existing customer relationship or simply providing an email address does not automatically authorise its use for advertising matches. First-party data and server-side processing do not, by themselves, demonstrate compliance. Withdrawals of consent and deletion requests must be accounted for in the relevant process.

AI can assist with tasks such as ad variants and delivery decisions. Whether this produces better communication and an economic benefit must be measured against the specific objective. Relevant messages, suitable landing pages and additional outcomes are the criteria. The following scenarios illustrate possible applications, subject to the necessary data-use and consent requirements.