Skip to content
Ninety days, eight deliverablesSee how
Nightshift

Delivery

Name the numbers before you build anything

The most valuable week of a data engagement is the one where nothing gets built.

July 22, 20265 min read

A baseline you cannot argue with later

A baseline recorded after go-live is not a baseline. Agreeing the measures, their definitions and their current values while the work is still hypothetical is the only way the final number survives scrutiny.

It also forces a conversation that otherwise gets put off forever: which three or four measures matter, out of the forty you could put on a screen.

Name a person, not a committee

Every definition needs someone with the authority to settle what it means. Without that, modeling stalls at the first disagreement and the schedule slips while everyone waits for consensus.

We make this a contractual condition, not a suggestion, because the engagements that skip it are the ones that overrun.

Check the systems are readable

We verify in week one that each source system can be read. Week six is too late to find out. If something cannot be connected inside the scope, drop it or price it there and then instead of absorbing the surprise later.

Want this against your own numbers?

A scoping call works out which measures are worth baselining and whether your systems can actually be read.

Let’s chat

More field notes

MeasurementAugust 8, 2026· 4 min read

Count from the catalog, not from the page

A screen reported seven products selling below cost. The catalog held twenty-six. The bug was not in the calculation, it was in what the count was computed over.

Read more

Start with the numbers you cannot see today

Half an hour works out which measures are worth baselining and whether your systems can be read.