Personalisation is only as good as the data behind it. Many of the organisations we work with are on the same path: moving through Optimizely CMS 12 to CMS 13, and expecting the new platform to deliver more relevant, more measurable digital experiences.
The CMS gives you the tools to show different content to different people. What it can't do on its own is tell you who those people really are, what they've just done, or what they're likely to do next. That's the job of a data platform, and it raises a fair question for any digital leader: when does a data platform like the Optimizely Data Platform (ODP) actually earn its place?
This article sets out where ODP adds value, the data and timing conditions that make it pay back, examples from sectors we know well, and an example of the checklist you should use for knowing when you're ready.
ODP delivers its strongest return when a customer journey can be materially improved by combining behavioural and customer data from more than one source, and acting on it in near real time, across channels you already control.
That breaks down into two tests.
ODP's core job is joining things up. It stitches an anonymous website visitor to a known customer, and merges what they do online with what your CRM, billing, policy or membership systems already know about them. The result is one profile per person, available to your CMS, experimentation and messaging tools.
If a single source can already answer the question - say, a CRM field that tells you someone's membership grade - you may not need a data platform at all. CMS audiences can often handle that on their own. The value appears when the useful signal only exists once sources are combined. Example: 'on a support register and browsing tariff pages; renewal due in 60 days and no CPD logged this year.'
ODP's real-time audiences update in under 90 seconds, using the last 28 days of behaviour. That matters when the moment of intent is short: someone is researching a decision right now, has just hit a problem, or is about to leave. If a nightly export from your CRM would do the job just as well, the case for real-time data is weaker.
If one data source can answer the question, you probably don't need ODP. If the answer only appears when you join them up - and it matters now - you probably do.
ODP grew up in ecommerce, and its out-of-the-box predictions lean towards purchase behaviour. In the sectors below, the value comes from designing the right events and bringing in CRM data. Here's where both tests are passed.
A customer on your support register visits the “struggling to pay” page after a missed direct debit. On their own, each signal is weak. Combined, they're a clear case for surfacing support content and suppressing sales messaging - while they're still on the site, not in next month's campaign.
The same pattern supports smart meter journeys (show booking prompts only to eligible customers who haven't booked) and fixed-term retention (combine contract end dates with visits to comparison pages).
A member within ten years of retirement age starts reading guidance articles and using the retirement calculator. Joining their scheme data with that behaviour lets you offer timely next steps and a prompt to book guidance.
It also gives you something regulators increasingly expect: evidence. Consumer Duty asks firms to show good outcomes and understanding across different groups of customers, and one profile per member, with events tracked across the journey, makes those outcomes measurable.
Many manufacturers know what a customer bought, through product registration or warranty data, but their website treats every visitor the same. Connecting that data with browsing behaviour means spares, service contracts and upgrades can be targeted to the equipment someone actually owns, and how old it is.
For B2B journeys, ODP can also link anonymous research by engineers and specifiers - spec sheets, CAD downloads, configurator use - to known accounts in CRM, if using Microsoft Dynamics. The connector released in June 2026.
Falling engagement is the clearest warning sign that a member won't renew: fewer logins, no CPD activity, missed events. Those signals usually sit in different systems. Brought together, they give you an early-warning audience and the chance to act months before the renewal date, not after it.
The same data powers the first 90 days of membership: prompting new members towards the benefits they haven't used yet.
For most Optimizely customers, the best moment to invest in ODP is during the move to CMS 13, not after it.
The upgrade is the cheapest moment to lay the foundations. Adding the ODP tracking tag, agreeing an event model and connecting your CRM are far easier while templates, components and integrations are being rebuilt anyway. Adding them later means reopening work you've just signed off.
Optimizely's documented real-time audience integration supports CMS 12 on .NET 8 and CMS 13 on .NET 10, ODP audiences appear alongside CMS audiences, so editors can personalise content blocks without leaving the tools they know. CMS 13's Content Variations also make it easier to hold several versions of the same content for different audiences or tests.
Think in terms of a pairing, not a product. ODP delivers the most when it works alongside the rest of the platform:
ODP decides who - the joined-up profile and real-time audience.
The CMS decides what - the content each audience sees.
Experimentation proves it worked - ODP audiences can now be built directly for Feature Experimentation.
CMP keeps the variants coming - personalisation creates demand for content, and producing variants is where programmes can stall.
Other signs the time is right: you're already personalising, but by hand (manual rules, CRM exports, many separate email lists); the same person exists in several systems with no link between them; or your board or regulator is asking for evidence of customer outcomes.
Before you invest, check you have these in place - or a plan to get them:
A selection of priority use cases, maybe 3-5 each tied to a business outcome (retention, conversion, cost to serve).
Named data sources for each use case, and an owner for each one.
A consent model that your legal and privacy teams have agreed, reflected in your privacy policy.
Identity keys - how you'll recognise the same person across systems (email, member or customer number).
CRM data you trust: deduplicated, reasonably complete, and kept up to date.
An agreed event model - the handful of actions that matter, named consistently across web, app and back-office systems.
A supported CMS: CMS 12 on .NET 8 or CMS 13, or have an upgrade already planned.
Content capacity to produce and maintain variants for each audience.
An experimentation plan, so every personalisation can be measured against a control.
A business function that owns audiences day to day, not just at launch.
A data platform won't fix a data problem. If you can't yet name your use cases, don't have a consent model, or don't trust your CRM data, start there. The platform will be far more valuable once those foundations are in place.
The same applies if nobody will own audiences after launch. ODP's value comes from acting on the data, week after week. Organisations that bring the data together but never act on it see little return.
Imagine this
A pension scheme member, 58, logs in for the first time in two years. Before the page loads, you already know their retirement date is approaching, that they read a guidance article last week, and that they haven't nominated a beneficiary. So instead of the generic homepage, they see their own next three steps - and a prompt to book guidance.
No export. No overnight batch. No guesswork. Just the right conversation, at the right moment.
That's what joined-up data makes possible. And for organisations already planning their move to CMS 13, the foundations can be laid now, as part of work that's already happening.
If you'd like to test whether ODP passes the two tests for your organisation, talk to us about a readiness review.