Personalisation for CRO: When It Works and When It Is Marketing Theatre

Personalisation gets marketed as the future of CRO. Every visitor sees the perfect experience for them, conversions double. The reality most UK businesses find is different. The tools are expensive, the implementation is slow and the measured impact is usually smaller than the vendor case studies suggested. Personalisation does have valid CRO uses. Priority Pixels adds it later in engagements inside conversion rate optimisation for UK B2B and public sector organisations, once the fundamentals are working. The wider framework sits in the CRO framework guide. The measurement discipline behind good personalisation matches the discipline behind good A/B testing.

Priority Pixels adds personalisation as a later workstream on CRO engagements. Fundamentals first, personalisation only after those fundamentals are stable and producing measurable output.

The right sequence: fix the fundamentals first. Get analytics right, get form CRO right, run structured tests on the busiest pages. Once those are in place, personalisation becomes the next layer worth exploring. Skipping ahead produces expensive tooling that never earns its keep.

What personalisation actually is

The vendor definition is broad: any experience that adapts to the individual user. In practice, personalisation splits into a small number of use cases with very different value profiles. Some earn their keep, some do not. The distinctions matter since the effort involved varies wildly.

Campaign-specific landing pages that mirror ad copy are the highest-value use case and often not called personalisation at all. It is just disciplined landing page hygiene. Returning-visitor CTAs and sector-specific messaging via firmographic data sit close behind. Individual-level personalisation from CRM data on the specific visitor is expensive to build, hard to measure and rarely produces the lift the vendors promise.

The use cases that actually work

The pattern seen across many engagements is that a small number of well-defined personalisation moves produce the majority of the value. Product recommendations on B2B service sites usually underperform static featured content the marketing team chose. Real-time behavioural personalisation based on scroll or mouse movement is complex to build and rarely produces measurable lift. Where the effort pays back is set out below.

Full detail on the tools that support these use cases sits in the CRO tools guide.

The GDPR consideration

Personalisation in the UK operates under data protection rules. Behavioural personalisation based on cookies requires consent. Firmographic personalisation based on IP-derived company identification sits in a legal grey area most legal teams treat conservatively. The ICO’s consent guidance is the authoritative source. UK cookie consent rates on most B2B sites materially reduce the addressable audience for any tool that depends on cookies. First-party personalisation (URL parameters, referrer-based routing, IP segmentation) works around this and is usually the personalisation that produces the highest lift.

The vendor demo economy pushes personalisation harder than the evidence supports. Personalisation cases in vendor pitches almost always show the “with” line rising and the “without” line flat, but the comparison is rarely apples-to-apples. Real-world tests with proper hold-out groups produce more modest results and often no measurable lift at all on B2B service sites where the content library is not deep enough for meaningful recommendations. HubSpot on personalisation statistics gives useful context on where the vendor claims tend to overshoot actual impact. McKinsey research on growth marketing is another reference for the more measured view. Where personalisation fits into the wider tooling market sits in the CRO tools guide.

The measurement problem

Personalisation is harder to A/B test than static changes since the personalised variant means different things to different visitors. Comparing “personalised” against “not personalised” is comparing apples to oranges. Proper testing needs a hold-out group and measurement of lift against the same audience segment the personalisation was targeting. Testing method sits in the A/B testing guide.

The right sequence: fix the fundamentals first. Get analytics right, get form CRO right, run structured A/B tests on the highest-traffic pages. Then personalisation becomes the next layer. Where personalisation fits alongside search-driven work is covered in the CRO vs SEO overview. For service businesses the pattern differs slightly, covered in CRO for service businesses.

Most personalisation implementations skip the hold-out. The team switches it on, watches conversion figures, sees a rise (which may or may not be attributable) and declares success. Without proper controls the attribution is guesswork. CXL research on personalisation testing.

The pattern above is the practical starting point rather than the only shape a working programme can take.