Evaluation Is Not an Audit (So Stop Treating It Like One)
A partner told me last month that her team calls evaluation season "the inspection." Not affectionately.
I get why. So much of what organizations experience as "evaluation" actually behaves like an audit: someone shows up at the end, checks your work against a standard you didn't fully choose, and hands you a grade. Pass or fail. Compliant or not. Funded again or not.
No wonder people brace for it. No wonder they manage the story instead of telling it straight.
Here's the redirect I find myself making, gently but firmly, in almost every kickoff call: an audit and an evaluation are not the same tool. What separates them starts with the question underneath — but it doesn't stop there. It shows up in the whole approach.
The question underneath
An audit asks: did you do what you said you'd do? It looks backward. It's built to produce a verdict.
A good evaluation asks: what are we learning, and what should we do differently? It looks forward. It's built to produce a decision.
That distinction isn't semantics. It changes what people are willing to tell you. When someone believes they're being audited, they manage the story — they highlight the win, soften the miss, hold back the messy detail that might actually be the most useful thing in the whole dataset. When someone believes they're genuinely being asked to learn something together, they tell you what actually happened. That's the entire difference between data you can trust and data that's just optics.
Why this is corrosive in partnerships specifically
Funders and grantees. Boards and staff. Agencies and community partners. Most of these relationships already carry some power imbalance built in. Layer an audit mindset on top of that, and you've told the less-powerful party: anything you disclose here can be used to reduce your funding, your standing, or your job. Of course they hold back. You would too.
This is why "just ask better survey questions" is rarely the real fix. You can have a beautifully designed instrument and still get guarded, sanitized answers — because the problem was never the instrument. It was the implied threat sitting underneath it.
What this actually looks like in practice
We don't open a project by handing a client a template survey. We start by building an actual relationship with the team, taking the time to learn what genuinely moves the people they serve, and listening — not once, but over and over, well past the point where it starts to feel efficient.
Sometimes that leads us to a standard interview or survey. Sometimes it leads somewhere more creative — like photovoice, where the people closest to the work hand you the story in their own images and words, instead of just answering our questions about it. The method follows the relationship, not the other way around. You can't shortcut your way to that kind of trust with a better-worded question. It's built, slowly, by showing people — repeatedly — that what they tell you gets used to help them, not used against them.
Three questions that signal you've drifted into audit mode
Even well-intentioned teams slide into audit thinking without noticing. A few honest checks:
Are you asking questions to confirm what you already believe, or to learn something you genuinely don't know yet? Confirmation-seeking is an audit habit wearing evaluation's clothes.
When someone shares a disappointing result, is your first instinct to explain it away, or to get curious about it? Defensiveness is a tell — for you and for the people you're evaluating.
Would people tell you the truth if they knew there were zero consequences attached? If the honest answer is "probably not," that's not a data collection problem. That's a trust problem, and no instrument fixes it.
The redirect
If you're heading into an evaluation cycle and you want honest answers, the fix isn't a new tool or a longer survey. It's a sentence, said early and said often: we're not grading you, we're learning with you. Then you have to actually mean it — in how you ask questions, in what you do with a disappointing finding, in whether "we didn't hit the target" gets treated as a failure or as information.
One thing to try this week: before your next evaluation touchpoint — a check-in call, a survey launch, a data review meeting — say the words out loud to whoever's in the room. Watch what it does.
