Skip to content
Launch Promenade

Sharing results with the business

An analysis that cannot be checked will be argued with rather than acted on. This page is about making it hold.

Lead with the sentence, not the diagram

Section titled “Lead with the sentence, not the diagram”

Open with the finding stated plainly: population, measure, gap, suspected cause. The process map goes on the second slide, as evidence.

Leading with the diagram invites a debate about the diagram — why is that box there, why is that arrow so thick — and the room never reaches the point. Leading with the claim makes the diagram do its actual job, which is supporting it.

Pick views for the audience

Section titled “Pick views for the audience”
  • A process map is for people who work in the process. They will spot what is wrong with it faster than you will.
  • The metro map view works for people who have never seen a Petri net: object types drawn as transit lines over a shared spine, readable without training.
  • A dotted chart is unbeatable for batching, because vertical stripes need no explanation.
  • A single number with its population — “38% in this region, 6% elsewhere” — is what survives into a steering meeting.

Be explicit about what you filtered

Section titled “Be explicit about what you filtered”

Every analysis makes choices: a time window, a case ID, a noise threshold, cases excluded as incomplete. State them on one slide. It sounds like weakening the argument; it does the opposite, because the first question from anyone senior is “does this include X?”, and having the answer ready is the difference between a finding and a discussion.

Hand over something reopenable

Section titled “Hand over something reopenable”

Export the workspace as a bundle. It carries the log, the artifacts derived from it, and the provenance — which action produced what, with which parameters, in which order.

That means the analysis can be reopened rather than described. Someone who doubts a number can trace it back to the step that produced it. Someone who wants the same analysis next quarter can re-run the chain on a fresh export instead of reconstructing your clicks from a document.

Note that the bundle contains your event data — share it under the same rules as the export it came from.

Plan the second analysis before finishing the first

Section titled “Plan the second analysis before finishing the first”

The first analysis is usually a proof that the data supports the method. Its most valuable output is often a better question:

  • The same process, but object-centric, because order/item/delivery interaction turned out to be the story.
  • The same question on a longer window, to test whether a finding is seasonal.
  • A different process, now that the extraction pattern is established and IT knows what you need.

Where to go next

Section titled “Where to go next”