The map was right. The traveller wasn't.

What I said

Nine stops, four questions, and a summit.

Three years ago I stood in front of a room and described what customer experience would look like in 2025.

The device was a map, drawn like the London Underground. Nine stops, each one an investment a business could make: user feedback, content, testing, web health, online quoting and contracting, personalisation, IP-based tooling, lead qualification, marketing automation. Four questions routed you between them. Do you regularly obtain user feedback? No: travel to Feedback. Are your key journeys good? No: Content, then Testing. Answer honestly and the map told you where to start.

Personalisation sat at the summit. It was the furthest stop, the one you earned your way to once the basics were solid.

It is now 2026. Three of those calls are worth checking, less because of whether I was right, more because of how each one aged.

Check two

The wall everybody braced against was quietly taken down.

Two of the nine stops, IP-based tooling and personalisation, rested on the same assumption: third-party cookies were about to disappear, and identity would have to be rebuilt on something else. Practically every roadmap of that period carried some version of that assumption.

On 22 April 2025, Google confirmed it would keep offering third-party cookie choice in Chrome and would not roll out the standalone prompt it had planned. On 17 October 2025, it retired ten Privacy Sandbox technologies, including the Topics API, Protected Audience and Attribution Reporting, citing low levels of adoption. Three components survived.

So the wall never arrived. A great many businesses spent several years preparing for a deadline that was eventually cancelled by the company that set it.

Meanwhile the change that did reshape the journey, assistants and agents standing between a customer and a website, wasn't on my slides, and I don't recall it being on anyone else's.

What the three have in common

The stops were mostly right. The traveller changed.

Put them together and the pattern isn't that the map was wrong. Most of those stops are still where I would put them. Feedback still matters. Content still matters. Web health still matters. It matters a lot - I have since evidenced a six-figure annual saving opportunity across these on a single client engagement.

What was wrong was the traveller.

The whole map assumed a person walking it. That assumption is buried in every stop, but it is loudest at the summit. Personalisation as the destination means the highest form of maturity is deciding what a human visitor sees. If a meaningful share of your journey is now mediated by something that does not see, is not delighted and does not scroll, the summit is in the wrong place, and not one of those nine stops was designed with that traveller in mind.

Nobody should feel foolish about this. The map was a fair reading of the evidence available in 2023. That is exactly the problem with maps.

One caveat, stated rather than buried

What you can check here, and what you can't.

The 2023 material was internal and has never been published, so you cannot verify what I said the way you could check a dated public post. The three external facts above are all verifiable and sourced below. My account of my own slides is not.

So this is a viewpoint piece, not a scored prediction. That is the standard I would want applied to anyone else making the same kind of claim, and it would be poor form to exempt myself from it.