The decision audit: trace your last ten significant decisions

Ask a leadership team who makes decisions and they will describe an org chart: this person owns product, that person owns hiring, the founders own strategy. Ask about the last ten decisions that actually mattered and the org chart usually stops being true.

A decision audit is simple to run and hard to argue with. List the last ten decisions that genuinely shaped the company: a pricing change, a hire, a pivot in the roadmap, a partnership signed or dropped. For each one, write down who actually decided, who was consulted along the way, how long it took from first raised to finally settled, and whether it bounced back up after someone thought it was already resolved.

What does a decision audit usually show?

Three patterns, almost every time.

The list of actual deciders is shorter than the org chart implies. Ten decisions, and the same two names show up eight or nine times, regardless of the domain on paper. Product decisions, hiring decisions, partnership decisions: same people, different letterhead.

Elapsed time has little to do with difficulty. A decision that should take a day takes three weeks, not because it is hard, but because it has to travel through people who were only ever meant to be informed, not asked. A genuinely hard decision sometimes takes a day, because one person had both the context and the authority to just make it.

Some decisions get made twice. Once in a meeting, and again a month later when someone senior reopens it, not maliciously, but because they were never told it had already been settled. Nobody designed that redundancy. It is what happens when decision rights live in people's heads instead of somewhere written down.

Why does this matter more than it sounds like it should?

Because decision speed is close to a proxy for organizational health. Companies rarely fail from making the wrong call. They stall from taking too long to make any call, or from making the same call more than once. A decision audit turns a vague feeling of slowness into ten concrete examples with names, dates, and durations attached, which is a different kind of conversation than "things feel slow lately."

It also tends to reveal the same bottleneck the work trace finds from a different angle. The work trace shows you where a task waited. The decision audit shows you why: someone downstream needed a call that only one or two people in the company were actually empowered, or willing, to make.

How to run one yourself

List ten decisions from the last quarter that changed something real. For each: who decided, who was consulted, how long it took, and whether it needed a second pass. Then look at the shape of the list, not the individual entries. If the same one or two names anchor most of it regardless of subject matter, that is not a resourcing problem. It is a decision-rights problem, and it is fixable by writing down, deliberately, who owns which category of call, rather than leaving it to default upward.