← Essays Essay · Decision Work

When Your Gut Says No and the Numbers Say Yes

By Stan Tscherenkow · Published April 2026 · 11 min read

Quick Answers

Why does the gut say no when the numbers say yes? Because the gut is reading signals the numbers have not yet captured. Pattern recognition on the counterparty, on the commercial structure, on the operating assumptions inside the model. When the two disagree, it is not the gut being irrational. It is the gut working on inputs the analysis has not made legible.

When is the gut actually right? The gut is most reliable on counterparty reads, on pattern-matching against prior situations the founder has lived through, and on sensing the direction of drift in a situation that has not stabilized yet. It is least reliable on pure probability work, on scale effects the founder has not operated at before, and in domains where the founder is systematically biased.

What is the structural process when the two disagree? Four steps. First, make the gut signal explicit: what specifically is it saying no to. Second, check whether the model's inputs include the thing the gut is reading. Third, ask whether the situation matches one the gut has a reliable track record on or one where bias is likely. Fourth, if the gut is reading a real signal the model has missed, the model is incomplete, not the gut wrong.

When should the numbers actually win? When the decision is a probability call at scale the founder has not operated at. When the situation is outside the founder's pattern library. When the gut is running against a specific source of bias. The numbers win when the analysis is on solid ground and the gut is running a known bias. The reverse is also true: the gut wins when the analysis is incomplete and the gut is reading real pattern.

When analysis and instinct disagree, most founders default to the numbers. The numbers are not always what they look like. And the gut is not always what it feels like.

What the gut is actually doing

The gut is not magic. It is not an intuition that arrives from outside the person. It is the compressed pattern-recognition output of everything the founder has experienced, every counterparty they have read, every deal that closed, every deal that fell apart, every hire that worked, every hire that did not. The gut has read thousands of situations. What it outputs, fast and without explanation, is a weighted similarity score against all of them.

The numbers are a subset of what the gut is processing. A financial model takes a few dozen inputs and runs them through an explicit logic. The gut takes hundreds of inputs, many of which were never written down anywhere, and runs them through a logic that does not generate its own explanation. When the two agree, the decision is easy. When they disagree, it is almost always because the gut is processing an input the model does not have.

This is not a claim that the gut is always right. It is not. The gut has biases, blind spots, and systematic failures that are as predictable as the failures of any formal method. But treating the disagreement as "the numbers are rigorous and the gut is feeling" is backwards. Both are outputs of pattern recognition. One of them operates on a narrower input set and produces a cleaner explanation. The other operates on a wider input set and cannot produce the explanation at all.

The question when they disagree is not which output is more trustworthy in general. It is which one has processed the signal that actually matters in the specific situation in front of you. Most of the time, the work is to find out what the gut is reading that the numbers have missed. Sometimes the work is the opposite.

Where the gut is reliable, and where it is not

The gut is not uniformly reliable. It is very reliable in some domains and systematically unreliable in others. Knowing which domain you are in is what tells you how much to weight the signal.

Where the gut tends to be right

Counterparty reads. You have met thousands of people. You have been burned by a specific type of person before. When your gut says the person across the table is the same type, it is almost always reading real pattern. The analysis will not see what the gut is seeing, because the relevant signals do not fit in a memo.
Situations that match your pattern library. If you have lived through a similar commercial situation before, the gut has a real case file to reference. The signals that mattered last time are the signals the gut is processing this time. Track record of the gut in this kind of situation is what you are relying on.
Direction of drift. The gut is good at sensing whether a situation is stabilizing or deteriorating, before the data catches up. The numbers describe the snapshot. The gut reads the momentum. In fast-moving situations, the momentum is what matters.

Where the gut tends to be wrong

Pure probability at scale. The gut is bad at expected value over large numbers. It systematically over-weights recent events, dramatic outcomes, and situations that pattern-match to personal experience even when the match is superficial. In a decision that is genuinely a probability call over many repetitions, the numbers are usually more reliable than the gut.
Scale effects outside your pattern library. If the decision is at a scale you have not operated at, the gut has no case file. It will still produce an output. The output is extrapolation from smaller-scale cases that may or may not apply. Confidence in the gut scales with proximity to your actual experience. At the edges, it should be low.
Domains of systematic bias. Every founder has them. The business they built is one. The co-founder they recruited is another. The market they fell in love with is a third. In these domains the gut is not reading pattern, it is running a bias dressed as pattern. The bias is usually invisible to the founder. That is what makes it a bias.

Before weighting a gut signal against a number, ask which of these categories the decision is in. A counterparty read in a domain you have experience in, with no bias attached, is a gut call you should weight heavily. A scale-effect probability decision outside your pattern library, with a known bias at play, is one you should weight the other way.

The four-step process when they disagree

When the disagreement is live, the work is to structure it, not to pick a side. Four steps.

Gut versus numbers, structured

Make the gut signal explicit. Not "I have a bad feeling." That is not a signal, that is a summary. What specifically is your gut saying no to. The counterparty. A term. The timing. The structure. The scale. The person who will run it after close. Force yourself to write the specific object of the gut's resistance in a single sentence. If you cannot, the gut has not finished processing, and the right move may be to give it more time rather than overriding either side.
Check whether the model's inputs include the thing the gut is reading. Look at the assumptions. Look at the sources of data. If the gut is saying "the retention assumption is wrong because the product does not do what the deck claims," is that assumption actually in the model, or is the model taking the deck at face value? Nine times out of ten, the gut is reading something the model is not explicitly testing. When you find it, you can choose to test it or not, but you no longer have a gut-versus-numbers conflict. You have an incomplete model.
Ask which domain the decision is in. Counterparty read where you have experience? Weight the gut heavily. Probability call at a scale outside your pattern library? Weight the numbers heavily. Domain where you are known to be biased? Weight the numbers heavily even when the gut feels certain. Domain where the gut is reading direction of drift that the snapshot does not yet show? Weight the gut. The weighting is not arbitrary. It is a function of where the gut has a track record and where it does not.
Document the override. Whichever way you go, write down what you decided and why. Specifically name which signal you weighted and what you chose to discount. Six months and eighteen months later, come back and check. Over time, this builds a real track record of when your gut was right and when it was not, which is the only way to get more reliable at weighting it in future disagreements.

The gut is almost never saying the numbers are wrong. It is usually saying the numbers are measuring the wrong thing.

Most of the time, what you find in steps one and two is that the gut and the numbers are answering different questions. The numbers are answering "does the commercial case work." The gut is answering "is this the right thing to do." Different questions produce different answers, and treating the conflict as if they were the same question is the error. The architecture for getting to one answer, when the decision is what is stuck rather than the data, lives in the stuck decision.

When the numbers should actually win

This essay is not a defense of trusting the gut over the numbers. It is an argument for reading both accurately. The numbers should win in specific cases.

Four cases where the numbers are the better signal

Probability calls at scale outside your experience. If the decision is a high-repetition probability question at a scale you have not operated at, the statistics are more trustworthy than your pattern library. Insurance decisions, capital allocation across many small bets, hiring at pace for a role you have not personally done. The numbers are on solid ground. The gut is guessing.
Decisions inside your known bias domains. If you know you are biased toward this product, this person, this market, the gut cannot be trusted as a pattern reader. It is running the bias. In those cases the numbers, imperfect as they are, are still less biased than the gut signal. Proceeding with the gut here is self-deception dressed as decisiveness.
Situations where the gut signal is a preference, not a pattern. "I do not like this" is not "I am reading danger." The two feel similar. They are not the same. A preference dressed as a gut read will fail you repeatedly. A pattern-based gut read will be right more often than chance. The difference is whether you can name the specific pattern the gut is matching against. If you cannot, it is preference, not pattern.
When the gut is running on insufficient data. Early in a situation, before you have met the people, seen the numbers close up, or spent real time in the domain, the gut is working on surface-level inputs. Its output will be confident anyway, because the gut is always confident. But confidence and reliability are not the same. At the beginning of a situation, weight the analysis, keep gathering, and let the gut develop a real read before asking it to make the call.

When the gut and numbers agree, the decision is easy. When they disagree and the gut is reading a real signal the numbers missed, the model is incomplete and the gut is the signal. When they disagree and the gut is running bias, preference, or under-developed pattern, the numbers are the signal. The work is knowing which case you are in, not defaulting to one side.

The founders who make the fewest expensive mistakes in this space are the ones who have stopped pretending that analysis is rigorous and gut is feeling. Both are pattern recognition. One of them operates on a narrower, cleaner input set. The other operates on a wider, messier one. A serious decision process uses both, weights them against the specific domain of the call, and writes down the reasoning so the track record can be read later. The rest is confidence dressed as method. The generalized essay on where unmade decisions go when this process fails is the hidden cost of delayed decisions.

If you are in front of a decision where the analysis says one thing and the gut says another, and you cannot tell which is reading the real signal, the conversation that names it takes one hour. Bring it before you commit either way.

Apply

Stan Tscherenkow Private Business Advisor Two decades operating across Europe, Russia, Asia, and the United States.

About Stan →