A navigable dashboard in week two: what has to happen for it to be true

A navigable dashboard with real data in week two of a BI project is a promise that is a little scary written down. And it is good that it is scary, because it forces a specific way of working from day one. KubiqX has it on its website, in the description of Core, and it is not a slogan: it is a consequence of how the work is ordered.

It is not about being fast. It is about the fact that correcting course in week two costs a conversation, and correcting it in week eight costs the project. These are the conditions for it to hold, and what happens when they are not met.

The five conditions of the first week

• Access to the sources before Friday. Not a "we will send it over": credentials, an export or a working connector. Without real data there is no real dashboard, there is a mock-up, and mock-ups deceive.
• One business person who answers within 24 hours. Not a committee. Someone who knows what "sales" means in that company, whether it includes VAT and from which date the history is valid.
• Five indicators agreed, not fifty. The five the manager looks at on Monday morning. The rest will come, but the first version is built on those.
• Excel as the starting point, no shame. The history almost always lives in spreadsheets. It is taken as the initial source and built upon. Migrating later is easier than starting from nothing.
• A first version that is ugly and incomplete, on purpose. It is there so someone can point at it and say "that is not right". Every correction at that moment is worth gold.

The four places where it breaks

• The ERP does not export. Or it does, but only the vendor can do it, and the vendor takes three weeks. When that happens it has to be said in the first meeting, and the timeline adjusted before starting, not in week four.
• Nobody owns the numbers. If "margin" has three definitions and nobody can pick one, the dashboard cannot pick for them. It stops until someone decides.
• Wanting the final version in week two. It does not exist. The final one is the week-eight version, and it is good precisely because there was a week-two version that was wrong.
• Changing the five indicators every week. Iterating is changing how something looks. Changing what is measured every week is starting from scratch every week.

Why it is worth insisting on

Because a BI project delivered whole at the end of the quarter is a blind bet: for eight weeks nobody sees anything, and at delivery it turns out half the figures make no sense. It is better to show something imperfect early and fix it in front of the people who will use it.

What delays teach

Of the BI projects that drag on, almost none drag on because of technology. They drag on because of access that never arrives and definitions nobody wants to close. Week two is not a speed target: it is the date on which those two problems come to light, while they are still cheap.