Personalisation is the practice of changing what a website, email or advert shows to a person based on what is known about them, such as where they are, how they arrived, what they have looked at or what they bought before. The aim is to make the content more relevant to that person than a single version for everyone would be.
How personalisation works
Every personalised experience needs two things: a signal and a rule. The signal is what you know; the rule is what you change because of it.
- Context signals need no history: the visitor’s approximate location, the device, the time of day, or the ad or email they clicked. A heating engineer can show “Serving Bristol and Bath” to visitors from that area, and a landing page can repeat the exact offer from the ad.
- Behaviour signals come from what someone has done on your site: pages viewed, items left in a basket, a returning visit.
- Customer data comes from accounts and purchases: past orders, loyalty status, preferences someone has told you. This is first-party data, collected directly.
Rules range from simple (“show the trade pricing banner to logged-in trade accounts”) to machine-learned recommendations (“customers who bought this also bought”). Email has done this for years through dynamic content blocks; websites increasingly do the same.
Most useful personalisation is really segmentation: a handful of meaningful groups, each shown something more relevant, rather than a unique page for every individual.
Why it matters
Relevant content removes work for the visitor. Someone who arrived from an ad for commercial cleaning should not have to find their way past domestic cleaning offers. A returning customer should not be shown the first-time buyer discount they already used. Done well, personalisation shortens the path to what people came for.
In the UK it also carries legal obligations. Personalisation based on personal data needs a lawful basis under UK GDPR, and your privacy notice should explain it in plain terms. If it relies on cookies or similar storage on the visitor’s device, PECR applies, and non-essential storage needs consent before it is set. A visitor who rejects analytics and marketing cookies should still get a working site, just a less personalised one.
There is a trust question too. Personalisation that feels helpful when it is obvious (“based on your last order”) can feel unsettling when it reveals data people did not expect you to have.
Common mistakes
- Starting with complex software before the basic pages and offers are right.
- Creating more versions than anyone can write, check and keep up to date, so some show stale offers for months.
- Never measuring whether the personalised version beats the standard one.
- Loading personalisation tools before consent, or treating “personalisation” as an essential cookie purpose.
- Using sensitive inferences, such as health or financial difficulty, to change what people see.
How to act on it
Start with context, which needs little data and little legal complexity. Match landing pages to the ads and emails that send traffic to them, show location-relevant details where your service area differs, and treat returning customers differently from new visitors where your platform supports it.
For each personalised element, keep a small holdout group who see the standard version, and compare results over a meaningful period. Drop variations that do not earn their maintenance.
Before adding tools, check your consent banner and privacy notice cover what you plan to do. Matching ads, pages and follow-up to the same audience is a large part of how I run performance marketing across Google and Meta.