The phrase “talk it through with someone smart” is almost right, and the part it gets wrong causes real problems.
Intelligence is not what makes a person useful to think with. Some of the sharpest people I know are actively unhelpful when I’m stuck — and a few people I’d never describe as brilliant are the ones I call.
Worth understanding why, because it changes who you ask.
What actually makes someone useful
They know your context, or one close enough to it. Someone who has run a service business with employees understands a set of constraints that a very clever person from a completely different world simply doesn’t have. They know what a bad month feels like, what collecting late payment involves, what it costs to lose a key person. That knowledge does more work than raw analytical ability.
They have nothing at stake in your answer. This disqualifies more people than you’d expect. Your business partner, your spouse, the employee whose job depends on the decision, the consultant who bills by the project — all of them may be entirely honest and none of them is neutral. You need at least one person whose life is unaffected by what you decide.
They ask more than they tell. The useful ones are noticeably light on pronouncements. They want to know what you’ve already tried, what you’re worried about, what the numbers actually are. People who arrive at a recommendation quickly have usually pattern-matched you to someone else.
They’ll say the uncomfortable thing. Plenty of clever, kind people will not tell you that your pricing is the problem, or that the person you’re defending should have gone a year ago. The willingness matters more than the insight, because the insight is frequently obvious and nobody has said it out loud.
Who to avoid, specifically
Not bad people. People who are structurally poor for this particular purpose.
- Anyone who needs to be right. You’ll spend the conversation managing their position rather than examining yours.
- Anyone who is impressed by you. Admiration is pleasant and it makes honest feedback nearly impossible. This is why founders surrounded by fans make such expensive mistakes.
- Anyone who will be hurt by the answer. Asking a team member whether you should restructure the team is not a fair question to put to them.
- Anyone selling a solution. They may be correct. They are not the person to help you decide whether you have the problem.
- And the person who only ever agrees. Comfortable, useless, and surprisingly hard to notice, because agreement feels like validation rather than absence.
The thing nobody says about finding these people
Most advice here amounts to “build your network,” which is true and no help at all.
Here’s what actually works: you find them by being one first.
The people worth thinking with are, almost without exception, people who get asked frequently — because they’re useful. They have limited patience for one-way relationships. If the first time they hear from you is when you need something, you’ll get a polite answer and not much more.
So the route in is to answer other people’s questions properly, with your real numbers and your actual mistakes, before you need anything. Not as a tactic — it doesn’t work as a tactic, people can tell — but because that’s what the relationship is made of.
In the network I run, the women who get the most useful answers are reliably the ones who answer other people’s questions the most specifically. Nobody enforces that. It’s just what happens.
How to actually have the conversation
- Bring the real version. Not the tidied-up problem. The one with the embarrassing detail in it, the number you don’t like saying, the thing you’ve been avoiding. Describe a sanitised problem and you’ll get advice about a problem you don’t have.
- Say what you think you should do. This invites disagreement, which is the entire point. Withholding your view to get an “unbiased” answer just means you get a generic one.
- Ask about their experience, not their opinion. “What happened when you did this?” is worth several times “what do you think I should do?” Evidence beats judgement, particularly judgement formed in four minutes.
- Write down what you heard, afterwards. Within the hour. You will otherwise remember the parts that agreed with you, which is the most common way a good conversation produces no change at all.
How many people you need
Fewer than you’d think. Three or four is plenty.
What matters is that they’re different from each other. Three people from the same industry at the same stage share your blind spots, and you’ll all reinforce each other cheerfully. One person running a bigger business, one running a different kind of business, and one who knew you before any of this started will cover most of what you need.
Regularity matters too. A relationship you use once a year, in a crisis, isn’t a relationship — it’s a favour, and it has to be re-established each time before you reach anything useful. Monthly, even briefly, keeps the context alive so the conversation starts somewhere other than the beginning.
The real reason this works
Almost none of it is about the other person being clever.
It’s that running a business inside your own head produces a particular kind of error, where an assumption you made three years ago has quietly become a fact, and everything since has been built on top of it. You cannot find those from the inside. They don’t feel like assumptions. They feel like how things are.
Someone outside your head trips over them immediately, usually by asking a question so basic it’s mildly irritating. Why do you charge by the hour? Why does that have to go through you? Why are they still a client?
That’s the value. Not brilliance — a functioning outside perspective, applied regularly, by someone willing to say the obvious thing you stopped being able to see.
More on why conversations beat advice, and why more courses won’t fix it.
