Measure the decisions, not the activity
The question is not what they did this month. It is which decisions got made that were previously stuck, and whether they were made well enough that nobody has had to revisit them.
Keep a list of the stuck decisions from the brief. At ninety days, that list is the review. It is unambiguous in a way that a summary of activity never is, and it protects the operator as much as it protects you: a decision that could not be made because the company would not supply the information is a fact about the company.
The four questions at ninety days
- Which decisions on the original list are now made, and which are not, and why?
- What do we now know that we did not know in month one? A good engagement produces uncomfortable information early.
- What has changed that would survive them leaving tomorrow? Process, documentation, a hire, a system. If nothing would survive, you have bought advice rather than capability.
- What should the next ninety days be, and should it be fewer days than the last ninety?
Signals it is working
- Your team asks them things directly rather than routing through you.
- They have told you something you did not want to hear, in the first six weeks.
- The function produces something on a schedule that it did not produce before.
- They have started to reduce their own days, or have said the seat will not need them at this level for long.
Signals it is not
- Every output is a document. Recommendations that nobody is accountable for implementing are consulting, at senior rates.
- Nothing has been said that challenges the founder. A senior person who agrees with everything in month two is either not looking or not willing.
- The day is spent in status meetings. That is a sign the company is consuming the engagement rather than using it.
- You cannot describe what would break if they stopped. If nothing would, nothing was built.
What not to measure
Responsiveness between days. You are not buying availability, and an operator who replies within minutes on a day you are not paying for is either not busy or is stealing time from another client, which is a preview of what they will do to you.
Hours. A fractional engagement priced in days and measured in hours turns into timesheet arguments, and the person who wins those is never the client.
Questions
- What KPIs should you set for a fractional executive?
- Set outcomes rather than metrics: the specific decisions that need making and what the function should produce on a schedule by day ninety. Activity measures like hours or meetings attended reward the wrong behaviour, because you are buying judgement rather than time.
- How long before you know if a fractional executive is working?
- Six weeks is usually enough to know, even at one day a week. By then a good operator will have told you something you did not want to hear and your team will have started going to them directly. Ninety days is the right moment for a formal review.
- Should a fractional executive be available between their days?
- For genuine emergencies, yes. For everything else, no, and you should not want them to be. An operator who is instantly available on days you are not paying for is taking that time from another client, which tells you what will happen to yours.