Twenty-four months to rebuild a precinct's digital estate.

What the assessment said

Before any of it could be sequenced we read the estate through our Digital Asset Value Model, which scores a digital asset across four dimensions and then does something unusual with the result: instead of averaging them, it takes the lowest. The argument is that a digital asset performs at the level of its weakest dimension, because the other three can only spend what that one lets through. At the V&A the weakest dimension was data, comfortably, and everything the estate was capable of was capped by it.

That single finding is why the programme opened where it did. Experience was decent, the technology was serviceable, the strategy was clear enough. None of it mattered while the estate could not identify a person across two properties or report one number that meant the same thing in two places. Put data at the heart of everything became the commitment for the next two years, and it was a conclusion the model reached rather than a slogan anyone chose.

The roll-out plan

A transformation that runs in sequence takes a decade and loses its sponsor somewhere in year three. This one ran as four parallel streams with deliberate dependencies between them: data first because it set the ceiling, campaigns immediately because they earned attention while the foundations were still being poured, product on the longest arc, and analytics threaded through so that everything shipped could be judged.

The campaign stream

Running a campaign stream alongside a platform rebuild is usually where transformation programmes come apart, because marketing cannot wait two years for a foundation and ends up commissioning yet another orphan site. The multi-site framework is what stopped that happening. Each campaign got a real digital home with its own art direction and its own logic, built off shared foundations, which is why four of them shipped inside the same year the platform work was still underway.

Measuring it

The analytics stream existed so the rest of the programme could be judged rather than admired. An omni-channel reporting layer was set up across every web and application property, and behind it a standing procedure: every digital product we deploy has its success KPIs documented before launch, monitored after it, and reported on a cadence. It is a boring discipline and it is the reason the next section has numbers in it at all.

Search was rebuilt on the same logic. With every web property now stemming from one strategy, a single integrated SEO programme could scale across all of them instead of fighting itself, sequenced over six months and then handed to an ongoing cadence tied to the content team's calendar.