It was never about the taste.

I learned a very important life lesson from all the spinach I hated being fed as a child.

Green leafy vegetables made me gag through childhood, and for years I thought I didn't like the taste. After resisting my mum's and grandma's attempts through my teen years, when I became an adult, I started trying to solve for flavour. Different recipes, different seasonings. None of it worked.

What actually stood in my way was texture. One day, I realised I cannot tolerate edible leaves or flowers in my mouth. Made me gag, pretty much a binding constraint. Once I identified and named that correctly, the fix became simple. Blanch the spinach, blend with lemon and salt, drink it once a week.

It became one of the most consequential health decisions I have ever made.

I think about that often because now the same pattern shows up everywhere in business.

Most people misdiagnose what is actually in their way. The deepest professional skill of this era is constraint discrimination - the ability to accurately identify what is obstructing progress, distinguish between those constraints that can be reshaped and those that cannot, and respond differently to both types. Almost everything else in business is downstream of this one move.

This matters more now because the landscape is shifting fast. Senior leaders are questioning which parts of their expertise still hold. Mid-career professionals are testing whether their playbooks still work. Companies are watching competitive advantages erode faster than they can build them. The broader environment - inflation, geopolitics, labour-market anxiety - compounds it. Even highly capable people are struggling to orient.

Many of them are looking around and seeing mountains where there are actually rocks.

Rocks can be picked up and moved. Mountains cannot.

The problem is that, from where you are standing, mountains and rocks often look the same.

## Three failure modes

Before people arrive at a real solution, three failure modes usually appear first.

The first is misdiagnosing the constraint. You solve for the wrong problem entirely. Hiring more sales reps for coverage when your deal conversion is a problem. Discounting for what's actually a value-articulation problem.

The work feels productive but the outcome does not materially change because the lever being pulled is not connected to the thing that is actually stuck.

The second failure mode is brute-forcing. This is what happens when you correctly identify the problem but choose the highest-friction path toward solving it. Instead of redesigning the environment around reality, you attempt to overpower reality with consistency, discipline, or willpower.

This works temporarily, but it fails structurally. It cannot move the same rock every Monday for two years.

A simple example. A sales head sees the quarter slipping. The brute-force move is to run 9pm pipeline reviews every day until quarter-end, push every rep on every deal, jump on the biggest accounts personally, discount whatever needs discounting. You hit the numbers, the team is exhausted. The same fire drill happens the next quarter, and the one after.

A behavioural-design approach asks a different question. What revenue engine architecture would catch bad-fit deals at entry, at qualification, at forecast, so the quarter-end heroics become unnecessary? The goal is no longer heroic consistency. The goal is continuity without dependence on willpower.

The best systems are not the ones that work only when motivation is high. They are the ones that survive contact with real life.

The third failure mode is opting out. The environment feels too unstable, too politically difficult, too uncertain, or too overwhelming, so people disengage. They postpone decisions and defer experimentation, and wait for clarity that never fully arrives.

But opting out is not neutrality. It is surrendering your role in shaping what gets built. You stop influencing your own relevance, your own leverage, and your own trajectory inside the system.

## Diagnose. Then discriminate.

The way forward is to diagnose the constraint correctly and then discriminate.

Some constraints do not move. Your nervous system. Your team's actual cognitive capacity. The market's willingness to pay. Macroeconomic conditions. The structural realities of your operating environment.

These are design inputs. For these constraints, the move is accommodation. You design around them intelligently instead of pretending they do not exist. Trying to overpower them usually produces the brute-force failure mode.

Other constraints do move when you work on them honestly. How your team is structured. What you incentivise. How decisions get made. What work belongs to machines versus humans. What gets measured, rewarded, reviewed, and celebrated.

These constraints often appear fixed only because the surrounding architecture reinforces them. For these, the move is redesign.

Most organisations get this wrong in both directions at once. They treat fixed constraints as negotiable and burn themselves out trying to overpower reality. At the same time, they treat negotiable constraints as fixed and fail to redesign systems that could absolutely be changed.

That is the mountains-and-rocks problem at the level of the enterprise.

Constraint discrimination is the diagnosis. Behavioural design is the implementation.

## What this looks like in a revenue engine

Take a simple example. A firm has six strategic enterprise accounts, the most valuable customer relationships in the business. Leadership correctly identifies the opportunity - grow these accounts. The customers already trust the company, there is budget headroom, and there are adjacent problems to solve.

So the standard motions begin. Account managers are assigned. Cross-sell plans come into play. Quarterly review cadences are established. Growth targets are communicated.

Six months later, account expansion has barely moved. The architecture of the plan explains why.

- The original account executive still owns the real customer relationship, so the customer continues calling them first.

- The account manager is treated operationally as a service contact rather than a strategic advisor.

- The compensation structure still rewards new-logo acquisition more heavily than expansion revenue.

- Pipeline reviews focus on new business.

- Internal prestige still belongs to hunting rather than expansion.

The organisation says it wants account growth, but the environment rewards something else.

No amount of training can solve a structural contradiction.

Behavioural design rebuilds the architecture itself - who owns the relationship, what the compensation structure rewards, what shows up on dashboards, what gets celebrated internally, and what gets reviewed first in pipeline conversations.

I've been doing this work in some form my entire adult life. The frame has sharpened across analytics, strategy consulting, and the firm I've built around it, but it didn't start there. It started with the spinach that I knew was really good for me, but couldn't bring myself to swallow. Name what's actually in the way. Distinguish what moves from what doesn't. Build the architecture that makes the right thing easier.

AI is making this way of thinking unavoidable. Because the same both-directions failure is now playing out in real time. Organisations are forcing AI into places where human judgement still matters deeply while preserving human labour in places where machine execution is already clearly superior.

This is a constraint discrimination failure that shows up as a technology conversation. But AI is only the most visible example of a pervasive pattern.

I am writing this from inside the question, not above it. Constraint discrimination is not theoretical for me. It is the thing I am actively putting into practice almost everywhere, in real time.

Most people misdiagnose what's in their way.

The work begins with diagnosing challenges correctly. Everything else follows.

The method behind this piece is set out under How We Work.