Why Digital Marketing Needs Fewer Tools and More Theory
Digital marketing theory is becoming more important as the number of platforms, tools, and automated systems available to marketers continues to expand.
Personalization under EU digital regulation is entering a period in which the ability to predict consumer behaviour is becoming less important than the legitimacy, transparency, and proportionality of the methods used to influence it.
For much of the digital economy’s development, personalization was presented as an almost unquestionable improvement. More data would produce better predictions, more relevant experiences, and stronger commercial outcomes. The underlying assumption was that increasing the precision of digital interactions would benefit both organisations and consumers.
That assumption is becoming more difficult to sustain.
European regulation is not bringing personalization to an end. Instead, it is challenging the conditions under which personalized experiences are designed, delivered, and evaluated. The distinction matters because the future of personalization will depend not only on technological capability, but also on how organisations interpret the relationship between commercial relevance and consumer autonomy.
Personalization has traditionally developed around an expanding capacity to collect and connect information. Browsing behaviour, purchase histories, location signals, engagement patterns, and inferred interests have been combined to construct increasingly detailed representations of individuals.
The commercial logic is understandable. Better knowledge of customer preferences can reduce irrelevant communication, improve product discovery, and make digital services easier to use.
Yet collecting more information does not necessarily produce a better understanding of consumer needs. Behavioural signals are incomplete, context-dependent, and open to interpretation. A product viewed once may indicate genuine interest, temporary curiosity, or an entirely unrelated purpose.
The General Data Protection Regulation (GDPR) places important limits on how personal data can be processed. Its principles include purpose limitation, data minimisation, transparency, and the requirement for an appropriate legal basis. Consent is one possible legal basis, but not the only one; the appropriate basis depends on the processing activity and applicable rules.
These principles change the strategic question facing organisations.
Instead of asking how much customer information can be collected, businesses must increasingly consider which information is genuinely necessary to provide a particular experience and whether its use can be justified.
This does not make personalization inherently less effective. It creates an opportunity to distinguish meaningful relevance from unnecessary surveillance.
The Digital Markets Act (DMA) addresses a different dimension of the personalization economy: the structural advantages associated with large digital platforms.
Certain platforms have historically benefited from their ability to connect information across multiple services. Search activity, social interactions, advertising engagement, and other behavioural signals can create extensive profiles that are difficult for smaller competitors to replicate.
For designated gatekeepers, the DMA restricts specified forms of combining and cross-using personal data without valid user consent. It does not prohibit personalized advertising generally, nor does it apply the same gatekeeper obligations to every business.
Its significance extends beyond the immediate compliance requirements.
When access to combined personal data becomes more dependent on meaningful consumer choice, the commercial value of unrestricted data accumulation becomes less certain. Competitive advantage cannot be assumed to follow automatically from operating the largest interconnected data environment.
This may encourage businesses to reconsider the role of their own customer relationships.
A retailer, for example, may have a legitimate reason to use a customer’s explicitly stated preferences to improve product recommendations. That relationship differs from constructing extensive profiles through unrelated interactions across multiple platforms.
The distinction is not simply between first-party and third-party data. Information collected directly by a business must still be processed lawfully. The more important distinction concerns the purpose, necessity, and reasonable expectations associated with its use.
The Digital Services Act (DSA) introduces another important consideration: personalization influences not only what consumers see, but also the choices they are likely to make.
Recommendation systems determine which products, information, and opportunities receive attention. Advertising systems influence visibility, while interface design can make some decisions appear more natural or convenient than others.
These mechanisms are not inherently harmful. However, they become problematic when individuals cannot understand the basis of recommendations or when digital environments systematically discourage meaningful alternatives.
The DSA requires online platforms to provide information about the main parameters of their recommender systems. Very large online platforms and very large online search engines must also provide at least one recommendation option that is not based on profiling.
The regulation additionally restricts certain profiling-based advertising practices, including advertising based on sensitive personal data and profiling-based advertising directed at minors under the conditions specified by the law.
These requirements reflect a broader shift in how digital choice is understood.
Personalization can help consumers navigate complex environments, but it can also narrow the range of alternatives they encounter. As discussed in How Platforms Reshape Consumer Choice Architecture, the organisation of digital options can influence decisions without explicitly removing any available choice.
The future of personalization therefore involves more than making recommendations accurate. It also requires attention to whether individuals retain a meaningful capacity to understand, question, and influence the systems shaping their experiences.
Artificial intelligence introduces additional complexity because personalization systems can increasingly infer preferences, generate individualized messages, and adapt interactions without continuous human intervention.
The EU AI Act establishes a risk-based regulatory framework for artificial intelligence. Among other provisions, it prohibits certain harmful manipulative or exploitative AI practices and introduces transparency obligations for specified AI applications.
However, it would be incorrect to interpret the AI Act as a general prohibition on AI-driven marketing or personalized recommendations.
Ordinary commercial personalization does not automatically become a prohibited or high-risk activity simply because artificial intelligence is involved. The applicable obligations depend on the system’s purpose, characteristics, deployment context, and effects.
This distinction is particularly important for organisations seeking to understand the practical consequences of regulation.
An AI system recommending products based on declared preferences presents a different set of concerns from one designed to exploit a person’s vulnerabilities through manipulative techniques that meet the legal threshold for prohibition.
Similarly, an automated customer service interaction may raise transparency questions without necessarily falling within the AI Act’s high-risk classification.
The strategic implication is that organisations need to understand what their systems actually do, rather than treating artificial intelligence as a single regulatory category.
One of the more significant consequences of the European regulatory environment may be a renewed interest in forms of personalization that do not depend on extensive individual profiling.
Contextual relevance provides one example.
A digital retailer can prioritise seasonal products, regional availability, or information relevant to the page being viewed without necessarily constructing a persistent behavioural profile of each visitor.
Similarly, a customer may explicitly select preferred product categories, communication frequency, or service interests. These preferences can support useful personalization while reducing the need to infer intentions from unrelated behaviour.
Neither approach is automatically exempt from data protection obligations. Contextual information can still involve personal data, and directly provided preferences must be handled appropriately.
Nevertheless, they demonstrate that personalization does not have to be synonymous with comprehensive behavioural tracking.
This is a meaningful strategic distinction.
The commercial objective should not be to know everything possible about an individual. It should be to understand enough to provide a useful experience without creating unnecessary informational or psychological costs.
Regulatory restrictions are sometimes described as obstacles to digital marketing performance. There are legitimate operational concerns behind this argument.
Reduced access to behavioural data can complicate attribution, audience modelling, campaign optimisation, and the measurement of incremental commercial effects. Businesses that depend heavily on cross-platform profiling may need to reconsider established assumptions about targeting and performance.
However, the relationship between personalization and commercial value is not linear.
A technically accurate recommendation may still be perceived as intrusive. A highly targeted advertisement may generate immediate engagement while undermining confidence in how personal information is being used.
Conversely, a less precise recommendation may be acceptable when the consumer understands why it appears and can easily modify the experience.
These differences suggest that performance should be evaluated beyond immediate conversion metrics.
Consumer trust, willingness to share preferences, repeated engagement, and the quality of the customer relationship are also relevant. They are harder to attribute to a single personalization intervention, but they influence the sustainability of digital business models.
This connects personalization with the broader question of digital reputation as a long-term asset. The way an organisation uses personal information can become part of the evidence through which customers assess its credibility.
Regulatory compliance and consumer trust are not identical. A lawful practice may still feel excessive, while a reassuring interface does not establish that the underlying processing is compliant. Responsible strategy must consider both dimensions.
The most consequential change may occur before any recommendation is generated or advertisement is displayed.
Organisations are increasingly required to reconsider the assumptions embedded in the design of their digital systems.
Should every interaction be recorded? Is a persistent profile necessary to improve the service? Can a consumer decline personalization without losing access to an unnecessarily degraded experience? Does the organisation understand how its automated systems arrive at consequential decisions?
These are not exclusively legal questions. They concern product design, organisational priorities, and the definition of customer value.
A business that regards personalization primarily as a mechanism for extracting behavioural information will approach these questions differently from one that regards it as a service intended to reduce friction and improve relevance.
European regulation does not resolve every tension between those approaches. Nor does it eliminate the commercial incentives that encourage organisations to collect information and influence behaviour.
It does, however, make certain assumptions more difficult to defend.
The ability to personalize an experience is no longer sufficient justification for doing so. Increasingly, organisations must also explain why a particular form of personalization is appropriate, what information it requires, and how it affects the people involved.
The next phase of personalization is unlikely to be defined by the complete disappearance of behavioural data or by the abandonment of artificial intelligence.
More plausibly, it will involve greater differentiation between forms of personalization.
Some experiences will continue to depend on individual preferences and behavioural information. Others will rely more heavily on contextual signals, explicit customer choices, or recommendations that do not require persistent profiling.
The appropriate model will depend on the nature of the service, the expectations of its users, and the legal and ethical implications of the processing involved.
This introduces a more demanding understanding of relevance.
Personalization has often been measured by how accurately a system anticipates what someone might do next. Under a more mature regulatory and strategic framework, relevance must also account for whether the system’s intervention is justified, understandable, and proportionate.
The future of personalization under EU digital regulation is therefore not a choice between technological sophistication and consumer protection. It is a test of whether digital organisations can develop sophisticated systems without treating unrestricted access to personal information as a prerequisite for commercial value.
Article by Dario Sipos.
Dario Sipos, Ph.D., is a Digital Marketing Strategist, Branding Expert, Keynote Public Speaker, Business Columnist, Author of the highly acclaimed books Digital Personal Branding and Digital Retail Marketing.
Readers who wish to explore the underlying research, citations, and peer-reviewed publications can find them via his Google Scholar Profile.
His verified academic identifier is available through ORCID.
Digital marketing theory is becoming more important as the number of platforms, tools, and automated systems available to marketers continues to expand.
Strategy is becoming more important than execution again because digital execution is increasingly automated, standardised, and accessible to a much wider range …
Consumer choice architecture is increasingly designed by digital platforms that determine which options become visible, comparable, and easy to select..