Three years building a data science function inside a bank taught me things no consulting framework captures. Not because the frameworks are wrong, but because they describe the world from the outside — and the inside is where the actual blockers live.
The transition into consulting is often described as moving from practitioner to advisor. I’ve come to think that’s the wrong framing. It’s better understood as moving from one operational context to many. The job isn’t to leave the instincts behind — it’s to carry them across.
What are those instincts, exactly? They’re the ones you develop only by being accountable for something that runs in production.
You learn what “good data” actually means when you’ve spent three months trying to use data that turned out to be a reporting extract, not a source system record, and the difference invalidated your entire feature set two weeks before a model review. You don’t forget that. No amount of data quality framework reading produces the same calibration.
You learn that a risk committee can kill a technically sound project in thirty minutes if the framing is wrong. Not because the committee is obstructionist — usually the opposite, they want to approve things — but because the framing failed to answer the questions they were mandated to ask. Model risk appetite. Operational resilience. What happens when it fails, not just when it works. A consultant who hasn’t sat in that room tends to write recommendations that would never survive it.
You learn that failed pilots leave institutional memory. The team that tried something similar three years ago is still in the building. Their silence in the workshop isn’t apathy — it’s caution that hasn’t been given a reason to become engagement yet. The project history that doesn’t appear in any briefing document shapes more decisions than the one that does.
These aren’t soft skills. They’re operational instincts — earned in production, not transferable by reading about them.
The move into consulting doesn’t make these instincts less relevant. It multiplies their application. In a bank, you develop them in one context, applied to one institution’s specific constraints, politics, and risk appetite. In consulting, you apply the same instincts across a portfolio of contexts simultaneously. The pattern-matching accelerates. The second time you see a data governance programme stall for the same reason it stalled at a different institution, you recognise it immediately. The third time, you know what to look for before the client tells you.
The risk in the transition runs the other way, actually. The temptation is to leave the practitioner identity behind — to become a consultant who speaks in frameworks rather than one who speaks in production experience. That’s the version of the role that generates nice slides and frustrating implementations. The useful version keeps the practitioner knowledge active, even in contexts where the role is advisory rather than executional.
Banking in particular creates a useful kind of rigour. The approval processes, the regulatory obligations, the model risk governance — they force a specificity that more permissive environments don’t. You get used to being asked “what’s the failure mode?” before “what’s the use case?” That reflex doesn’t go away, and it’s valuable in consulting because clients who haven’t worked in banking often haven’t built it in.
The question I keep coming back to isn’t whether the transition from practitioner to consultant is a trade-off. It’s whether you carry the right things across. The people who carry the instincts — the production intuitions, the political pattern recognition, the calibrated scepticism about what actually works — tend to become consultants whose recommendations get implemented. The ones who leave the instincts behind tend to become consultants whose recommendations are intellectually sound but operationally stranded.
None of that shows up in a framework deck.
P.S. The most important thing a practitioner background gives you in consulting is a specific kind of credibility in the room — not the credibility of authority, but the credibility of someone who has been wrong about something similar before and learned from it. That’s harder to convey than a credential, and more valuable than most.