I keep getting asked for the audit as something you can print, so here it is.

One page, blank, free, no form to fill in first.

It scores one thing: whether the context a live workflow depends on is actually usable at the moment the workflow needs it. Not what your architecture diagram says you bought.

What a running decision can reach.

What it scores

Four properties, scored zero to two.

1. Queryable asks whether the workflow can get the right question answered fast enough to act on.
2. Addressable is about identifiers, and whether the answer comes back against the one the rest of your system uses.
3. Current means the freshness matches the decision, which is a different bar for a churn score than for a cart trigger.
4. Accountable is the awkward one. Can you reconstruct afterwards what the workflow actually saw when it acted?

I set those out in what the context layer actually has to do, and turned them into this scoring method in the context audit, which has a worked example if you want to see one filled in before you start.

Eight points possible.

Six is functional.

Four is a sign there is real work to do.

And score three workflows rather than one. As I said in June, one score tells you about one workflow, and three start to show you a pattern.

The third block, and why it isn't scored

The bottom section asks something the four properties do not. Not whether the context arrives, but which version wins when two systems both hold the same fact and disagree about it.

That one refuses to sit in the same table.

It belongs to a fact rather than a workflow, it doesn't change between workflows, and there is no version of it that scores cleanly zero to two. So it sits underneath as its own short grid, one row per fact, with room for the system it comes from, whoever decided that, and the date you checked. A provisional answer with a date on it counts. Waiting for someone with the authority to declare a winner mostly means writing nothing at all.

I worked through why in "Which context wins when two systems disagree about your customer?".

Download it!

Look, it is a piece of paper with boxes on it. Print it, or work through it on screen if you prefer. If you want to formalize what comes out of it, freshness commitments and named owners are exactly what a data contract is for, and the Open Data Contract Standard is a reasonable place to start.

Martech Therapy also runs this with teams as a half-day working session, three workflows and the people who own them in one room. If it is useful, get in touch.