The usual framing of this question is wrong. The choice is not between data and instinct - experienced judgement is genuinely valuable, and organisations that ignore it make sterile decisions. The real question is which decisions your gut is reliable for, and which it is systematically bad at.
Where intuition is genuinely reliable
Experienced intuition is pattern recognition built from thousands of repetitions with fast feedback. It works well when:
- You have seen many similar situations and learned the outcome each time
- The feedback loop is short (a salesperson reading a room, a chef judging a service)
- The situation is qualitative and the variables are not measurable
- Speed matters more than accuracy
In these conditions, trust the gut. It is compressed experience and it outperforms slow analysis.
Where intuition fails predictably
Human judgement has well-documented, replicable weaknesses that no amount of experience removes:
- We overweight vivid recent events. One angry customer outweighs a hundred quiet satisfied ones in our sense of "what customers think".
- We see patterns in randomness. Three good weeks feels like a trend. It usually is not.
- We look for confirmation. Having formed a view, we notice supporting evidence and discount the rest, without any awareness of doing it.
- We are terrible at proportions and base rates. "Half our complaints are about delivery" means nothing until you know what share of orders involve delivery.
- We remember our successes. Which makes our track record feel better than it was, and our instincts feel more reliable than they are.
Every one of these is a place where a number beats a feeling, and none of them announce themselves.
The decisions that specifically need data
- Anything involving small differences (pricing, conversion, process improvements)
- Anything where the outcome is delayed, so your intuition never got feedback
- Anything about a large population - customers in aggregate, staff in aggregate
- Anything expensive and irreversible
- Anything where you have a stake in the answer being a particular way
That last one is the important one. The stronger your prior conviction, the more you need the number.
How to combine them properly
The sequence that works in practice:
- Use intuition to generate hypotheses. Experience is excellent at suggesting what to look at.
- Use data to test them. Including the ones you are confident about.
- Decide in advance what result would change your mind. This is the whole discipline in one sentence.
- Use judgement to interpret and to weigh what the data cannot capture - brand, relationships, strategy, risk appetite.
What "data-driven" should not mean
It should not mean paralysis, ignoring qualitative evidence, or hiding behind numbers to avoid accountability for a decision. Data narrows uncertainty; it does not remove the need for someone to choose. Organisations that pretend otherwise produce slow decisions and unowned failures.
Building the capability
Making this work requires enough statistical literacy to know when a number is trustworthy. The Statistics for Business course at London School of Business UK builds exactly that. Enquire today.