A theme park, an F1 circuit and a cultural district do not report the same things. They still had to be run to one standard.

Qiddiya is not one business. Sports, entertainment, culture, theme parks and motorsport each run their own operation, with their own business units, their own teams and their own idea of what progress looks like.

• The same word meant different things. Ready, approved and complete each carried a different definition depending on which vertical you asked. One set of stages every business had to map onto, and freedom inside them.
• One platform would have fitted nobody. A circuit package and a cultural fit out do not share a workflow, and forcing them to would have made both worse. Separate platforms per business, built from shared parts.
• Nobody could see across. Leadership had reporting assembled by hand, arriving late and already out of date. A portfolio view that reads from the same data the teams work in.
• People left before they started. On the version this replaced, exit and bounce sat at 100 per cent. Users landed and went straight back out. A first screen that says what the session is for and the shortest way to finish it.

The instinct on a programme this size is to standardise the work. That fails, because the work genuinely is different. What can be standardised is the language it reports in, and what it feels like to be inside it. The version before mine functioned, and people still did not take to it, because nothing about it read as Qiddiya.

Not one platform. Seven, each shaped around how that business actually runs, sharing one set of stages, one set of components and one definition of every status word.

And one set of guidelines carrying Qiddiya's brand and its culture through all of them, so a person opening any of them recognises where they are before reading a word.

Most of what these platforms hold is data, so the charts, the scales and the way a figure sits on a page are the design rather than decoration laid over it. Getting someone to their number in fewer steps is an art direction problem as much as a flow problem.

Seven platforms rebuilt to one set of guidelines. Satisfaction rose from 2 to 8 out of 10, and conversion went up 300 per cent. Conversion here means a session that reaches its target: the person gets the data they came for, in the fewest steps, without dropping out on the way.

The version before it worked and people still would not use it. Exit and bounce sat at 100 per cent, so nobody was getting past the landing page to find out whether it was any good. It carried nothing of the brand and nothing of the culture, so it never felt like theirs. Very little of that movement is new functionality. It is recognition, and a route to the number that does not make people give up on the way there.