Personalisation Strategy: Why Most Brands Are Doing It Backwards

Personalisation strategy is the discipline of delivering the right message, to the right person, at the right moment, based on what you actually know about them rather than what you assume. When it works, it reduces friction, increases relevance, and compounds over time. When it fails, which is most of the time, it produces the uncanny valley of marketing: a message that feels like it knows you but gets you completely wrong.

Most brands are not failing at personalisation because they lack data. They are failing because they have confused data collection with insight, and segmentation with relevance. The result is expensive infrastructure producing marginal returns.

Why Personalisation Gets Bought Before It Gets Thought About

I have sat in enough technology procurement conversations to know how personalisation projects usually start. A senior stakeholder sees a competitor doing something impressive, or attends a vendor presentation, and decides the business needs a personalisation platform. The platform gets bought. Then someone has to figure out what to do with it.

This is not a technology problem. It is a strategy problem that technology has been handed to solve. And no amount of machine learning fixes the absence of a clear answer to the question: who are we personalising for, and what are we trying to change about their behaviour?

The brands that do personalisation well start with audience architecture, not platform selection. They define the segments that matter commercially, the moments in the customer experience where relevance would change an outcome, and the signals that would tell them which segment a person belongs to. Then they buy the tools to execute that thinking. That order matters more than any specific technology choice.

If you are working through how personalisation fits into a broader go-to-market approach, the Go-To-Market and Growth Strategy hub covers the commercial framework that personalisation should sit inside, rather than operate independently from.

The Lower-Funnel Trap Most Personalisation Strategies Fall Into

Earlier in my career, I overvalued lower-funnel performance. I was impressed by the numbers, the attribution, the apparent precision of it all. It took time, and a lot of client P&Ls, to understand that a significant portion of what performance marketing gets credited for was going to happen anyway. The person who had already decided to buy and just needed to find the checkout button is not a personalisation win. They were already converted.

The same logic applies to personalisation. Most brands concentrate their personalisation effort on the bottom of the funnel: retargeting, cart abandonment emails, post-purchase recommendations. These are not unimportant, but they are the easy part. The person is already engaged. The signal is strong. The decision is already in motion.

The harder and more valuable work is earlier in the funnel, where personalisation can shift someone from unaware to curious, or from passively browsing to actively considering. Think of it like the clothes shop analogy: someone who tries something on is dramatically more likely to buy than someone who just walks past the window. Personalisation at the top of the funnel is the equivalent of getting the right item in front of the right person at the moment they are open to trying it. That is where commercial value is created, not just captured.

This connects directly to the tension between market penetration and existing customer optimisation. Brands that personalise only within their existing customer base are optimising a ceiling. The growth opportunity is in reaching people who do not yet know they need you, and making the first experience relevant enough that they stay.

What Personalisation Strategy Actually Requires

Strip away the vendor language and personalisation strategy requires four things to work: audience definition, signal identification, content infrastructure, and measurement that accounts for the counterfactual.

Audience definition means knowing which groups of people behave differently enough that they warrant different treatment. Not every difference in your customer base is commercially meaningful. A 45-year-old and a 28-year-old might buy the same product for the same reason and respond to the same message. Demographic segmentation alone is a lazy proxy for what you actually care about, which is intent, context, and stage of decision.

Signal identification means knowing what observable behaviour tells you which audience a person belongs to, and where they are in their decision process. Behavioural signals, what someone has read, clicked, searched for, or purchased, are almost always more reliable than declared preferences. People tell you what they think they want. They show you what they actually want.

Content infrastructure is where most personalisation programmes quietly collapse. You can have perfect segmentation and perfect signal capture, but if your content team cannot produce enough relevant variants to fill the matrix, the whole system degrades to a default experience with a personalisation label on it. I have seen this happen at scale. The technology works. The content does not exist. The result is a very expensive way to deliver the same message to everyone.

Measurement that accounts for the counterfactual is the hardest and most important. If you show a personalised experience to someone who was already going to convert, and they convert, that is not a personalisation win. Proper measurement requires holdout groups, incrementality testing, and a willingness to accept that some of what looks like personalisation lift is actually just demand that was already there. Forrester’s intelligent growth model has been making this argument for years, and it still gets ignored in favour of last-touch attribution.

The Segment-of-One Myth

The marketing industry has spent years selling the idea of the segment of one: hyper-individualised experiences tailored to every single person. It is a compelling concept. It is also, in most commercial contexts, a distraction.

True individual-level personalisation requires data density that most brands do not have, content volume that most teams cannot sustain, and a level of customer intimacy that most products do not warrant. You do not need to know that someone prefers blue to green if the decision they are making has nothing to do with colour preference.

What most brands actually need is meaningful segmentation: four to eight distinct audience groups with genuinely different needs, different decision drivers, and different moments of maximum receptivity. That is achievable, sustainable, and commercially significant. The segment of one is a vendor aspiration. Meaningful segmentation is a business outcome.

When I was growing the agency from around 20 people to over 100, one of the things that changed the quality of our client work was forcing ourselves to name the two or three audiences that actually mattered for each brief, rather than building elaborate persona documents that nobody used. Simplicity in audience definition creates clarity in execution. Complexity in audience definition creates paralysis.

Where Personalisation Creates Durable Commercial Value

There are three places where personalisation consistently earns its cost: onboarding, re-engagement, and category entry moments.

Onboarding is the highest-leverage moment in most customer relationships. Someone has just made a decision to try your product or service. They are paying attention. They are open to guidance. A personalised onboarding experience, one that reflects what you know about why they signed up and what they are trying to achieve, dramatically increases the probability that they reach the moment of value before they disengage. Generic onboarding is one of the most expensive mistakes in retention marketing, and it is hiding in plain sight on most brands’ dashboards.

Re-engagement is where personalisation can recover value that would otherwise be lost. A lapsed customer is not the same as a new prospect. They have history with you. They have context. A re-engagement message that acknowledges that history, that reflects what they bought, what they used, what they might have outgrown, is categorically different from a generic win-back campaign. The signal is there. Most brands just do not use it.

Category entry moments are the most underused opportunity in personalisation strategy. These are the moments when someone enters the market for a product or service, often for the first time. They are forming preferences, building mental models, and deciding which brands deserve further attention. BCG’s research on brand and go-to-market strategy has consistently shown that the brands which win at category entry are those that show up with relevance before the competition does. Personalisation at this moment is not about knowing the individual. It is about knowing the context well enough to be useful.

The Content Problem Nobody Wants to Talk About

I have judged the Effie Awards. I have seen what genuinely effective marketing looks like at its best. And one of the consistent patterns in effective personalisation is not sophisticated technology. It is disciplined content strategy.

The brands that personalise effectively have made deliberate choices about which messages vary and which stay consistent. They have not tried to personalise everything. They have identified the two or three points in the customer experience where a different message would produce a meaningfully different outcome, and they have invested in building the content variants to support those moments. Everything else stays consistent.

This is the opposite of how most personalisation programmes are designed. Most programmes start with the technology capability, which can personalise almost anything, and work backwards to fill it. The result is personalisation theatre: a lot of variation in message that produces no meaningful difference in outcome, because the variation was not grounded in a genuine insight about what different audiences need to hear.

Content infrastructure for personalisation also needs to account for the long tail of edge cases. What happens when your personalisation engine encounters a signal it does not recognise? What is the default experience? In most systems, the default is mediocre. Building a genuinely good default experience, one that works for people the system cannot classify, is not glamorous work, but it is where a lot of personalisation value leaks out.

Tools like Hotjar’s feedback and behaviour analytics can help identify where the default experience is breaking down, which sessions are showing frustration signals, which content is being ignored. That kind of qualitative signal is often more useful than aggregate conversion data when you are trying to understand why personalisation is not working.

Personalisation and Privacy: The Tension That Is Not Going Away

Any honest treatment of personalisation strategy has to address the privacy dimension, not as a compliance exercise, but as a commercial reality. The data infrastructure that personalisation depends on is under sustained pressure from regulation, browser changes, and shifting consumer expectations. Third-party cookies are largely gone. Signal loss is real and growing.

The brands that are handling this well are not the ones with the most sophisticated workarounds. They are the ones that have invested in first-party data relationships: email lists built on genuine value exchange, loyalty programmes that give customers a reason to identify themselves, content experiences that earn attention rather than track it.

First-party data is not just a privacy-compliant alternative to third-party tracking. It is a qualitatively better signal. Someone who has actively given you their email address and told you what they are interested in is a more reliable personalisation input than a probabilistic inference from browsing behaviour. The signal loss from cookie deprecation is real, but it is also forcing brands toward data relationships that were always going to be more durable.

BCG’s work on understanding evolving customer needs makes the point that the brands which build genuine understanding of their customers, rather than just tracking their behaviour, are the ones that sustain commercial advantage over time. Personalisation built on first-party insight is harder to build, but harder to replicate.

How to Build a Personalisation Strategy That Actually Works

The practical sequence for building a personalisation strategy that delivers commercial results is not complicated. It is just different from how most brands approach it.

Start with the commercial question, not the technology question. What behaviour are you trying to change, in which audience, and what is the commercial value of changing it? If you cannot answer that question specifically, you are not ready to build a personalisation programme. You are ready to do more audience research.

Then identify the moments in the customer experience where a different message would change that behaviour. Not every touchpoint warrants personalisation. Most of them do not. The discipline is in identifying the two or three moments where relevance genuinely changes an outcome, and concentrating your effort there.

Then map the signals that tell you which audience a person belongs to and which moment they are in. These signals need to be observable, reliable, and available at the moment of delivery. Signals that require days of data processing are not useful for real-time personalisation. Signals that are available immediately, search query, referral source, content consumed, are almost always more actionable.

Then build the content. Not the platform. The content. Write the variants, test them against a default, and establish whether the variation produces a meaningfully different outcome before you invest in automating it at scale. This is the step that most brands skip, and it is the step that determines whether the whole programme works.

Finally, build a measurement framework that accounts for incrementality. Use holdout groups. Test the counterfactual. Be honest about what personalisation is actually driving versus what was going to happen anyway. The brands that do this well tend to find that personalisation works in fewer places than they expected, but works much better in those places than their original estimates suggested.

Personalisation strategy does not sit in isolation. It is one component of a broader growth framework, and how it connects to acquisition, retention, and commercial planning matters as much as the personalisation mechanics themselves. The Go-To-Market and Growth Strategy hub covers the wider planning context that makes personalisation decisions more coherent and commercially grounded.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions