What the post-event survey actually tells you
The standard five-question survey sent the next morning measures three things: whether people enjoyed themselves, whether the logistics worked, and how they feel about the person who organised it. All three are worth knowing and none of them is impact.
Enjoyment is also a poor proxy. The activities that change how a team works — a pre-mortem, a failure résumé, a session that removes a pointless meeting — frequently score lower on 'did you enjoy it' than a day of go-karting that changed nothing.
If you are going to send a survey anyway, ask two better questions: what will you do differently, and who will you now talk to that you would not have before? Both are specific, both are answerable, and both can be checked later.
Three things you can honestly measure
Network change. Before and after, ask everyone to list the people they have asked for help in the last month. Count the cross-team links. This is a real, countable change and team activities move it more reliably than they move anything else.
The specific behaviour you targeted. If the session was about handovers, count handover defects for the following quarter. If it was about raising problems earlier, count how early issues were raised. Pick the measure before the session, not after — choosing it afterwards guarantees you find a flattering one.
Follow-through on the commitments. Every good debrief ends with named actions. Three months later, how many happened? This measures your programme honestly, and a low number is more useful than a high survey score.
What you cannot claim
You cannot attribute a change in retention, engagement scores or performance to a team building session. Too many variables move at once, the effects are small relative to the noise, and the people who get to decide the budget have usually seen that argument made badly before.
Claiming it anyway is the single fastest way to lose credibility for the whole programme. The honest position — 'this is a small input into something with many causes, and here are the specific things it did change' — is both more defensible and, in practice, more persuasive.
Be equally careful in the other direction. The absence of measurable impact is not evidence of no impact; it is usually evidence that the effect is smaller than your instrument.
A reporting format that survives scrutiny
One page per quarter. What we ran and who attended. What the stated purpose was. What the debrief produced, as a list of commitments. Which commitments were completed. Any measure you committed to in advance, reported whether or not it moved.
Including the things that did not work is what makes the rest of the page believable. A programme report with no failures in it is read as marketing.