Chapter 8: How We Know Organisations Aren’t Linear

Here’s draft chapter 8, and as always, I’d love to read your thoughts and feedback, as all my drafts are just that: drafts.

I’ve received lots of helpful pointers on earlier chapters 🙏

Chapter 8: How We Know Organisations Aren’t Linear

That was the theory, now the evidence. You just spent Q4 forecasting next year. Revenue targets, headcount plans, budget allocations, all built with impressive precision. Three-year strategies with specific milestones. Five-year plans projecting growth rates to two decimal places. How confident are you those numbers will be accurate? If you’ve worked in organisations for more than a few years, you know the answer. Last year’s forecast was wrong. The year before that was wrong. The five-year plan from 2019 became obsolete in 2020, and was quietly abandoned. Ditto last year’s. Yet here we are again, conjuring next year’s numbers with the same spreadsheets and the same pretence. This isn’t incompetence, it’s a category error.

We’ve a fetish for forecasts because we mistake organisations for linear systems that follow predictable rules. Input resources, execute plan, achieve predicted outcome. Lovely. It’s the same logic that builds bridges and manufactures cars, and it should work for organisations, provided we plan carefully enough and execute precisely enough. Except it doesn’t work. And it never has. But we keep doing it because we're making a fundamental assumption about what kind of system an organisation is. This shows up in everyday organisational life long before we get anywhere near transformation programmes. The stuff that happens when no one's trying to fix anything.

Business as Usual

You hire someone with the perfect CV, they nailed the interview and have outstanding references, then they don’t pass probation. Meanwhile, the risky bet you reluctantly took on becomes your best performer. We explain it away afterwards as cultural fit or some such, but before it happens, we’re confident. The CV and interview predict the outcome. Except they don’t, at least not reliably.

Or promotions. You promote someone everyone agreed had leadership potential. They’ve performed well for years and demonstrated all the right behaviours. Six months into the new role, they’re noticeably struggling. Decisions are slower, the team isn’t quite gelling. Meanwhile, on a different team, someone nobody expected to lead steps up when their manager leaves suddenly, and the team follows them. Not because anyone designed it that way. It just worked out; a nice surprise. If leadership ability were a predictable property of individuals, promotions would work consistently. They don’t.

Or strategy. You spend months on the five-year plan. Market analysis, competitive positioning, financial projections, scenario planning. Eighteen months in, the plan is obsolete. A competitor did something you didn’t anticipate. A technology shifted faster than predicted. Customer preferences changed in ways your crystal ball missed. Or, if you want a more recent example, a global pandemic happened. The plan wasn’t bad. The analysis was good. But the number of variables affecting outcomes and the way they interact make prediction very limited. The revenue target for Q3 was £10.2M; actual was £8.4M or £11.9M. Last year’s growth projection was 12%; actual was 7% or 18%. We’ve had decades to get better at forecasting. Better software, more sophisticated models, more data than ever before, AI. And still, forecasts are reliably wrong. They always will be. Nonlinear systems aren’t unpredictable in theory, but in practice forecasting one means tracing cause and effect through millions of local interactions, which would take something close to infinite precision and computing power. Not even AI will get there.

Or projects. They succeed, which is good. But when (if?) you debrief why, the reasons aren’t the ones in the plan. Someone happened to know someone in procurement who fast-tracked approvals. A team member had worked on something similar years ago and so there was no steep learning curve. The timing accidentally aligned with another initiative that created momentum. Luck, basically, disguised as execution. You can’t put “got lucky with timing” in the project retrospective, so you credit the plan. But you know. And of course, projects fail for reasons that also weren’t in the plan. A key person left. A dependency you didn’t know existed created a bottleneck. Something small cascaded into something large. The relationship between effort and outcome is unpredictable. Sometimes tiny problems derail everything; sometimes massive obstacles get overcome through routes nobody anticipated.

This is everyday organisational life. Hiring, promoting, strategising, executing. The core operations. And none of it behaves the way linear systems behave. Inputs don’t reliably produce predicted outputs. Small causes sometimes have large effects. Large investments sometimes produce no results. Expertise doesn’t guarantee good predictions, and careful planning doesn’t prevent surprise. Then we try to deliberately change things, and we're met with unpredictability and unintended consequences. Annoying, isn’t it.

Comfortable Fiction

A new CEO arrives with a vision. Let’s say it’s “Agile.” Consultants are hired, Scrum Masters appointed, stand-ups mandated. The vocabulary spreads, squads, tribes, sprints, ceremonies. Training runs for weeks and three months in there’s energy. Six months in, it’s waning. Stand-ups happen, but real decisions still get made in the same small group afterwards and retrospectives identify improvements that never get actioned. Twelve months in, the ‘transformation’ is a pile of dust. Teams may have their ceremonies and use new language, but the actual work happens the way it always did. It’s the same people with influence, the same bottlenecks and patterns, just with new vocabulary or a new org chart.

Or restructuring. My very first office job was at a 6,000-person organisation operating in over 110 countries. Within little over two years, we’d gone through two major organisational transformations. One global, one UK-only. The first was announced with the usual fanfare. Senior leadership had been off-site for weeks, working with consultants on a grand plan. Town halls explained the rationale of better alignment, clearer accountability etc etc. New org charts were drawn and distributed, reporting lines changed, some teams were merged, others split, and of course we played musical chairs. For months, the organisation was consumed with the mechanics of transformation.

Eventually, things settled. We learned the new structure, leapfrogged the new processes, and got back to work. Then, eighteen months later, another transformation. Different consultants and rationale, but the same pattern. More town halls, more org charts tweaks, more moves, more months of disruption. After both transformations, the organisation looked more or less identical to before. A few reshuffles here and there. Some title changes. But the actual work, how it got done, who had real influence, where decisions were made, this was all pretty much the same. It turned out, the organisation’s history had more say in its shape than any transformation plan did.

The Transformation Nobody Ran

The same pandemic that shredded five-year plans the world over also delivered the biggest transformation of working life in decades. And nobody designed it. In March 2020, organisations that had spent years explaining why remote work was impossible went fully remote in a fortnight. No consultants, no change curve, no town halls building the case. The thing that a decade of flexible-working advocacy had failed to achieve arrived in a matter of days, delivered by a virus with no methodology whatsoever.

Complexity has a name for events like this, gateway events. Shocks that tear through a system and shunt it into a new stable pattern. The Industrial Revolution was one. The internet was another. Probably AI, too. They arrive unplanned, open unexpected opportunities for some and disaster for others, and afterwards the new conditions simply become the norm. What sorts the winners from the casualties isn't planning. A system needs enough variety and flexibility to survive the shock in the first place, and only then can it profit from whatever the new conditions offer. This is how complex systems change, long stretches of apparent stability followed by a sudden shift, then a new stability. The pandemic changed the conditions, and the system reorganised around them. Grand plans and persuasion never came into it.

Executives have spent the years since trying to command it back. Return-to-office mandates, attendance tracked like a school register. The results have been about as good as the transformation programmes, because it’s the same mistake. A linear plan issued to a system that has already moved to a different stable state. So the one global transformation of our era wasn’t achieved by management, and nor will it be reversed by management.

The pattern is the same whether we’re talking about hiring or transformation programmes, strategy or restructuring, everyday operations or deliberate change efforts. Outcomes are unpredictable. Small things sometimes cascade while large investments sometimes disappear. What works in one context fails in another. Nothing sustains without continuous energy, and patterns emerge from below regardless of what's designed from above.

Baffling If Linear

This is what Chapter 7’s theory predicts. If organisations are complex systems, not linear machines, we should see exactly these patterns. And we do, constantly. Start with transferability. Best practices should work across contexts if organisations are linear. There can be minor adjustments, sure, but the core best practice should translate. Instead, the same intervention produces wildly different outcomes. I taught the Advice Process, the practice behind self-management pin-up Buurtzorg's decision-making, to teams across dozens of organisations. The same practice, the same training, and wildly varying results. This pattern appears everywhere. Agile transformations succeed in some organisations and flop in others. Psychological safety enables innovation in some teams and becomes an excuse for avoiding accountability in others. The reason is in the history. Organisations develop along their own evolutionary paths, each accumulating its own habits, relationships and quirks along the way. A practice distilled from a handful of success stories can’t account for what any particular system has become. There is no one right way, and copying Spotify’s squads won’t make you Spotify.

Then there’s the relationship between investment and results. Linear systems should show correlation. Thirty years of engagement programmes and billions spent should really move Gallup’s numbers. Instead they barely budge. Two-thirds of employees worldwide aren’t engaged. Ninety percent in the UK. And this despite an entire industry working the problem. Chapter 3 showed the data I tried to ignore. In a linear system, sustained competent effort produces results. In a complex system, interventions get absorbed and the system returns to its stable state. And the relationship runs both ways, sometimes tiny interventions have an enormous impact, and sometimes massive efforts change nothing.

Then there’s sustainability. Properly implemented practices should sustain themselves in linear systems. They’re structural. Instead, practices require continuous energy to maintain against the system’s natural state. Stop the energy and they revert. That’s exactly what Chapter 3’s eighty percent fade rate demonstrated. The experiments worked during the course because the course provided structure. Once that scaffolding disappeared, most couldn’t sustain themselves. The two self-managing organisations I worked for? Both reverted within eighteen months when their champion left. Not because the practices were poorly embedded. Because they required constant energy to hold in place, and when that energy stopped, entropy took over. A building left alone becomes a pile of rubble. A pile of rubble left alone never becomes a building.

And finally, there’s design versus emergence. If organisations were linear, you could design them from above and the design would stick. Instead, patterns emerge from below regardless of what’s designed from above. That’s what the informal structure is. Every organisation has networks of relationships, decision-making habits and power dynamics that don’t match the org chart, grown from local interactions rather than anyone’s design. Whether those patterns settle into anything useful is a different question. In Sami’s terms, self-organisation is the stable, productive layer between deathly order and wasteful chaos, and a system only reaches it where people are genuinely free to interact, capable of interacting, and guided by simple rules most of them actually support. Those conditions exist in degrees, and in most organisations they’re thin. So the system finds the layer only partially, in pockets, running well below what it could do, and where the conditions give out it swings between the two states instead, rigidity one year, firefighting the next, which is most organisations’ lived experience. Where the conditions do exist, you can get a Buurtzorg or a Morning Star. Not because they implemented a methodology, because they created conditions where self-organisation could occur.

The evidence is overwhelming. Best practices don’t transfer reliably. Investment doesn’t correlate with results. Small changes can have large effects while massive interventions get absorbed. Change doesn’t sustain without continuous energy. Patterns emerge from below, not from design above, and only the right conditions turn them into something stable. None of this is surprising if organisations are complex. All of it is baffling if they’re linear.

The consulting industry has spent decades treating organisations as linear. Diagnosing problems, designing solutions, implementing changes. All producing disappointing results. The standard response is “we need better implementation.” But after several decades and billions of dollars, perhaps the problem isn’t implementation quality. Perhaps it’s treating complex systems as if they were linear ones. That’s the category error.

The practices aren’t wrong. The toolkit isn’t useless. The problem is the implicit promise that applying the right practices and mindset will predictably produce desired effects. That promise only makes sense if organisations are linear. If they’re complex, that promise is impossible to keep. Which brings us to the uncomfortable question: if organisations are complex systems that can’t be reliably transformed through planned intervention, what does that mean for the rare examples where transformation did happen? And why are those examples so rare? That’s the rare bird problem.

A not-so-rare bird in Lyme Regis, Dorset, a few years ago (and my best Martin Parr impression 📸).

---