I've been running diagnostic engagements for a while now, and the part that still surprises me is how rarely the stated problem turns out to be the actual problem. Someone books a call about "alignment issues" and by the end of the first week it's clearly about one person who's been allowed to behave badly for years. Someone else says "our strategy isn't landing" and what they mean is "we have three conflicting strategies and nobody has been willing to say so out loud." The gap between the presenting complaint and the real thing is almost always there. The diagnostic interview is the main tool I use to cross it. I want to write this up honestly, including the parts that feel a bit uncomfortable, because I think there's a craft to this that doesn't get written about much — and I'm still refining it.
why meetings lie to you ¶
The meeting is a performance space. People know who's in the room, who reports to whom, and roughly what they're supposed to say. This isn't cynicism — it's just how groups work under social pressure. When I sit in on leadership meetings early in a diagnostic, I'm not there to gather information. I'm there to notice the gap between what's being said and what's being not-said: the quick glances, the topics that get technically addressed but never really discussed, the person who speaks last and least. That gap is my actual research question. The interviews I do afterward are designed to cross it. I try to be transparent about this with clients upfront — I tell them the interviews aren't a survey, they're a structured conversation, and I'm specifically interested in the things that are hard to say in a room full of colleagues.
the setup matters as much as the questions ¶
Before I ask anything substantive, I spend time on frame-setting — probably five to eight minutes that might look like small talk but aren't. I explain how I'll use what I hear (themes, not names, unless someone tells me something that suggests risk). I say explicitly that disagreement with the official narrative is not just allowed but useful to me. I usually say something like: 'The most helpful thing you can do is tell me what's actually frustrating, not what you think I want to hear or what would look good.' That sentence does a lot of work. It gives people permission they often don't know they needed. I also always conduct these one-on-one, never in pairs or small groups. The social dynamics of even two people in the room changes everything — someone becomes the spokesperson, the other person calibrates to them.
the questions that tend to open things up ¶
I have a loose sequence I come back to, though I adapt it constantly. These aren't verbatim scripts — they're orientations. The first one I ask, after basics, is some version of: 'If you could change one thing about how this organisation makes decisions, what would it be?' It's specific enough to be answerable and open enough to go anywhere. The second is: 'What's something that everyone here knows is true, but that doesn't get said in meetings?' That question lands differently every time. Sometimes there's a long pause — which is itself information. Sometimes someone laughs nervously and then tells me something that reframes the entire engagement. A third one I use toward the end: 'What would have to be true for this organisation to look significantly different in three years?' That one surfaces assumptions people are holding silently about what's possible and what isn't.
the follow-up is where the real work happens ¶
The questions above are entry points. The actual diagnostic work happens in the follow-up, and I think this is what separates a useful interview from a survey with a human attached. When someone says 'communication could be better,' I don't move on. I slow down. I ask: 'Can you give me a recent example?' or 'When you say communication — is that about information not flowing, or about decisions being made without input, or something else?' I'm interested in specifics: names of processes, names of meetings, recent events. Vague answers are usually an invitation to ask again more concretely. I also pay close attention to the moment when someone's answer shifts register — when they move from measured professional language to something that sounds more like how they'd talk to a friend. That shift usually means we've found something real.
what silence and hedging actually signal ¶
Two things I've learned to read carefully: silence and hedged language. A long pause before answering 'what's working well here?' is different from a long pause before 'what isn't working.' The first might just be genuine reflection. The second often means someone is deciding how much to trust me. I let silences run longer than feels comfortable. It's not a tactic — it's just that I've noticed how much information lives in the three seconds after someone finishes their first answer, before they decide they've said enough. Hedged language is similarly interesting. 'I think generally people feel like maybe the...' — that kind of stacking of qualifiers is usually someone gesturing toward something specific they don't quite want to name. My job is to create enough safety that they can name it, or at least point more directly.
the ethical side, which I think about more than I used to ¶
These conversations involve real trust, and I want to be honest that I've thought a lot about where the ethical edges are. I'm not a therapist and I'm not a journalist — I'm a consultant hired by the organisation, which means I have a client relationship that shapes everything. I'm always clear with interviewees about who hired me and what I'll do with what I learn. I don't name sources in reports. But I do synthesise — if six out of eight people say a variant of the same thing, that pattern appears in my findings, even if no individual is quoted. I've had moments where someone tells me something genuinely difficult — about a colleague, about leadership behaviour — and I've had to think carefully about what responsible handling looks like. I don't have a clean formula for this. What I do have is a practice of flagging when something crosses from 'useful organisational insight' into 'this person may need a different kind of support than I can offer.'
what I'm still working out ¶
I'm genuinely uncertain about a few things. One is how much my presence changes the answers — the observer effect is real, and I don't know how to fully account for it. Someone might tell me something because I'm an outsider and it feels safe, or they might perform a version of themselves they think will look good to a consultant. I try to design around this but I can't eliminate it. I'm also still experimenting with sequencing — whether it's better to interview front-line people before senior leaders or after (I've done both, and both have arguments). And I keep refining the question about 'what doesn't get said in meetings' — sometimes it's too direct and people deflect, and I'm looking for a softer approach that lands in the same place. If you've got thoughts on any of this, I'd genuinely like to hear them.
The diagnostic interview is not a neutral information-collection exercise. It's a designed conversation, and being honest about that — with yourself and with the people you're talking to — is, I think, what makes it useful rather than extractive. I'll keep writing up what I'm learning from these as I go.