Founder as Architect

7 days ago

5 min read

Updated: 18 hours ago

Founders tend to describe themselves by what their businesses do. We build houses. Sell software. Run restaurants. Advise companies. Manufacture things. But increasingly, I think that description misses the most important thing a founder actually builds.

A structure. A system. We build an environment.

We decide how decisions get made, what gets rewarded, what gets ignored, who has authority and something that came up over dinner this evening (thank you to Fraz Khan and Matt Jones for the interesting conversation), what happens when somebody makes a mistake.

And we decide whether bad news travels upwards, whether customers are listened to, whether meetings produce action or merely more meetings.

Then… something interesting happens.

The environment starts producing behaviour.

Behaviour, repeated for long enough, becomes culture. And the culture itself starts producing results.

Founders are rather more like architects than many of them realise.

The Story

Let’s look at Toyota, as it offers one of the clearest examples to my mind. Toyota’s famous production system isn't simply a collection of clever manufacturing techniques. It changes the environment in which people work.

One of its best-known principles is that problems should become visible immediately. Toyota developed the famous andon system (アンドン), where a worker who spotted a problem could pull a cord to summon help and, if necessary, stop the production line. Quality becomes something built into the process, rather than inspected at the end.

What is the behaviour that architecture encourages?

If an employee discovers a problem and stopping production gets them shouted at, problems will be hidden. If stopping the line is expected when quality is threatened, problems surface. So the same people, the same team, in a different environment, performs a different behaviour and a different result is produced.

Paul O'Neill did something similar after becoming CEO of aluminium giant Alcoa in 1987.

Rather than arriving with the usual grand pronouncements about shareholder value, growth and margins, he made worker safety the obsession. When an employee was injured, the relevant unit president had to report what had happened and how a recurrence would be prevented within 24 hours. That requirement forced information to travel quickly through the company and made managers investigate the systems behind accidents. During O'Neill's 13-year tenure, Alcoa says lost work days per 100 employees fell from 1.86 to 0.2. A 90% reduction. The slogan wasn’t the important bit of O’Neil’s mantra. It was the environment created.

O'Neill wasn't simply asking people to behave more safely. He redesigned what happened when they didn't. That's business architecture.

The Company You've Designed

This idea connects much of what I've been exploring through this series.

I looked at businesses making perfectly rational decisions inside flawed systems.

I looked at symptoms being mistaken for problems.

The challenge with feedback loops that reinforce themselves.

The idea that lag between decisions has consequences.

The metrics that tell us what happened, while disguising the why.

And first-order decisions whose second-order consequences appear months later somewhere else entirely.

All point towards the same conclusion, and it’ll be a little uncomfortable to many.

The business we have and the results we are getting is, to a large extent, the environment we designed and installed along the way.

So what’s the uncomfortable bit?

We didn’t necessarily design and install it deliberately. It just happened. Like evolution.

Nobody sits down and designs a company where nobody takes responsibility. But when every problem requires the founder's approval before anyone can solve it. People soon learn to wait for the founder. About more and more things. Then, the founder is normally to be found complaining that nobody takes responsibility.

Imagine the situation where somebody delivers bad news and receives twenty minutes of irritation and interrogation. Nobody introduces a policy saying: Please conceal emerging problems until they become expensive. They don't have to. The environment teaches people to do that by default.

Culture is often imagined as something soft and mysterious. I don't really think it is. It is hundreds of small lessons, delivered by leadership, about what happens around here.

Which salesperson gets praised?
What happens when a deadline slips?
Does anyone follow up the actions agreed in meetings?
Can somebody admit they made a mistake?
Does the company recruit brilliant people and then require them to ask permission for everything?
Does the fastest worker get rewarded with everybody else's unfinished work?
Or worse, can employee A repetitively not turn up and force employees B, C and D to do their work?

Those are design decisions, whether we recognise them as that or not. And once you see a business this way, the job of the founder, the boss, changes.

You stop asking only:

How do I get people to behave differently?

And start asking:

What is it about this environment that makes this behaviour sensible?

Isn’t that a much more useful question?

The Four Seasons

I started this series, The Shape of Business, several weeks back with the article Smart Businesses. Stupid Decisions.

I’ve mentioned the season each relates to along the way. The founder support system we developed works in seasons, so it has seemed relevant.

This article is perhaps where GamePlan makes most sense and needs a small coherent explanation. I don’t mean so much as a productivity system. This is a rhythm for designing the business.

Winter is where we think.

Autumn gave us evidence. Winter is where we decide what it means, what needs changing and where we're going next. This is drafting. Architecture, before construction begins.

Spring is where we test.

Our ideas meet reality while there is still room to change them. We try out processes. We test offers. We clarify responsibilities. The mistakes in our winter thinking are found and corrected. This is where andon really shines.

Summer is where we execute.

Then we can enjoy the season everyone wants. Sell like buggery. Deliver. Let the systems operate under real pressure, and let them flow.

And while summer runs, we observe and record. Where does work jam up? What breaks when volume increases? Which customers return? Where does the founder get dragged back into operations?

Ideally the production line doesn’t need stopping during this phase. But we must collect the evidence.

Autumn is where we sort the data.

What worked? What didn't? Which problems were isolated and which reveal something structural?

Then we take that knowledge back into winter.

Think. Test. Execute. Learn. Repeat.

Every year. Every project. Every hire. Every product.

Do founders build businesses? Or do we build the conditions in which businesses emerge?

If we don't like what the company keeps producing, eventually we have to stop blaming the occupants and take a look at ourselves and the architecture we implemented.