Behavioural targeting is a marketing technique in which advertising or marketing decisions are based on information about a person's or audience's observed behaviour, such as pages visited, searches performed, products viewed, purchases made, content consumed or advertisements interacted with. The basic idea is that previous behaviour can provide evidence about likely interests and can therefore help determine which advertisement or marketing message is most relevant.
Historically, behavioural targeting became strongly associated with online advertising because websites, advertising networks and tracking technologies could record digital interactions at scale. A retailer, for example, might identify a visitor who repeatedly viewed cameras and subsequently serve that visitor advertising for camera equipment. Behavioural targeting can operate using first-party information within a company's own environment or through broader cross-site systems involving external platforms. The term describes the targeting method rather than one specific technology.
Behavioural targeting usually involves several stages. First, behavioural events are collected, such as page views, searches, clicks, purchases or app interactions. The events are then associated with an identifier, account or audience record where technically and legally permitted. The system can classify the resulting activity into a segment, such as "recently viewed running shoes", "frequent purchasers" or "customers who have not purchased for six months". Advertising or marketing platforms can then use the segment to select messages.
Some systems use simple rules, while others use statistical models to infer interests or predict conversion probability. The quality of the targeting depends on the quality and representativeness of the behavioural signals. A person may view a product because they are researching a gift, comparing prices, writing an article or studying a subject, so an observed action does not necessarily prove personal interest. Behavioural targeting therefore involves inference and uncertainty even when the underlying event is accurately recorded.
Behavioural targeting is different from contextual advertising because the former uses information about behaviour while the latter primarily uses the context in which the advertisement is being displayed. If someone reads several articles about hiking and subsequently receives hiking advertisements across different environments, that can represent behavioural targeting. If an advertisement for hiking boots appears simply because the current article is about hiking, that is contextual advertising even if the reader's previous behaviour is unknown.
The two approaches can also be combined. A platform could select hiking advertising because the page is about hiking and then optimise the choice based on the user's consented previous interactions. The distinction matters because contextual advertising can often operate without constructing an individual behavioural profile, whereas behavioural targeting generally depends on some form of behavioural information being linked to an audience, account or identifier.
Historically, third-party cookies were a significant technical mechanism for behavioural targeting because they allowed advertising services to recognise a browser across participating websites. This made it possible to aggregate browsing events and classify browsers into interest audiences. Modern browser restrictions and privacy controls have reduced the reliability and availability of unrestricted third-party-cookie tracking. As a result, behavioural targeting can increasingly rely on first-party customer data, authenticated accounts, platform-owned data environments, consented identifiers and other technologies.
The decline of third-party cookies does not mean behavioural targeting has disappeared; rather, the sources and technical mechanisms used to create behavioural signals have changed. An organisation can still perform behavioural targeting within its own website or application using first-party interactions, subject to applicable privacy and consent requirements. The technical implementation therefore needs to be distinguished from the broader marketing concept.
Behavioural targeting can be useful for remarketing, retention and audience segmentation. A customer who abandoned a shopping basket may be eligible for a reminder, a customer who repeatedly uses a particular feature may receive relevant educational material, and a customer who has not purchased recently may be included in a re-engagement campaign. These applications can reduce irrelevant advertising and make limited marketing resources more focused. However, behavioural targeting can also produce unwanted repetition, inaccurate assumptions and excessive frequency.
If a customer has already purchased an advertised product, continuing to show the same advertisement can waste advertising expenditure and create a poor experience. Frequency capping, purchase suppression and recency rules can reduce such problems. Behavioural signals also decay over time: a search or page view from two years ago may have little relevance to a person's current interests. Effective systems therefore consider recency, frequency and context rather than treating every historical action as equally informative.
Behavioural targeting involving personal data can engage privacy and electronic-communications requirements. In the UK, the ICO states that organisations are likely to need consent under e-privacy rules for many website cookies and online tracking methods, and that where consent is required it must satisfy the applicable UK GDPR standard. Organisations also need to consider data-protection obligations where behavioural information identifies or can reasonably be linked to an individual.
This includes understanding the purpose of processing, providing appropriate information, managing retention and ensuring that data is not used beyond the relevant permissions or legal basis. Behavioural targeting should therefore not be treated as merely an advertising configuration. It can involve a chain of data collection, identity linkage, profiling, audience creation and advertising activation, each of which may require governance.
Behavioural targeting is also connected with profiling, but the terms are not identical. Profiling generally refers to automated processing of personal data used to evaluate or predict aspects of a person, such as interests, behaviour or preferences. Behavioural targeting can be a practical marketing application of behavioural information without necessarily involving every element of formal profiling.
A marketer could create a simple segment based on "visited Product A twice in the last seven days", while another system might use machine learning to predict purchase probability from hundreds of behavioural signals. Both involve behaviour-based targeting, but the analytical complexity differs. This distinction is useful because regulatory requirements can depend on the actual processing rather than the marketing label assigned to it. Organisations should document what data is collected, how it is combined, whether individuals are identifiable, what inferences are made and how the resulting audience is used.
The principal limitation of behavioural targeting is that behaviour is evidence, not certainty. Browsing activity can be shared between household members, performed for research or triggered by a temporary interest. Automated models can also amplify historical biases or misclassify people. Targeting can become particularly sensitive where inferred characteristics relate to health, finances, political views or other sensitive subjects. Strong governance therefore requires careful selection of data sources, appropriate exclusion rules, transparency, security and controls over sensitive inferences. From a marketing perspective, behavioural targeting should be evaluated against incremental outcomes rather than assuming that more precise targeting automatically produces better results. Controlled tests can determine whether behavioural targeting actually improves conversion or profitability relative to a broader or contextual strategy.
