The Architecture of Better Decisions

For decades, we've relied on the business case as the gateway to almost every important organisational decision.

Need investment? Write a business case.

Need more people? Write a business case.

Want to launch a new proposition? Write a business case.

The intention is entirely sensible. Before committing resources, organisations should understand the opportunity, the risks, the alternatives and the expected return.

But I wonder whether we've gradually started asking the business case to do something it was never really capable of doing.

Give us certainty about the future.

Markets change. Customers change. Technology changes. Competitors react. Costs move. Assumptions become obsolete.

The detailed financial model that looked convincing on approval day can look rather different six months later.

That doesn't make the business case pointless. Done well, it forces us to expose assumptions, understand risks and think properly about the economics of a decision.

The problem comes when greater detail creates an illusion of greater certainty.

I've seen organisations spend weeks, sometimes months, refining papers and financial models in pursuit of confidence that the analysis simply can't provide.

And the hidden cost isn't producing the document.

It's the delay.

While we're perfecting the spreadsheet, the opportunity may be disappearing.

The real issue is how decisions get made

I've worked in organisations where relatively straightforward decisions required multiple committees, repeated reviews and several layers of approval.

The intention was usually good: involve the right expertise, manage risk and make better decisions.

But something interesting can happen.

Everyone contributes.

Everyone has visibility.

Everyone effectively gets a vote.

And nobody feels genuinely accountable for the outcome.

This isn't really a business-case problem.

It's a decision architecture problem.

Decision by committee isn't necessarily good governance

Committees have an important role.

They bring together different perspectives. They challenge assumptions. They identify risks. They can materially improve the quality of thinking.

But contributing to a decision and being accountable for it are different things.

Finance should challenge the assumptions.

Risk should identify potential downsides.

Technology should assess feasibility.

Legal should advise on obligations.

Operations should explain the delivery implications.

Those perspectives make decisions better.

But there should still be clarity about who owns the decision.

And where a function genuinely has the authority to stop something, for regulatory, legal, financial or other reasons, that decision right should be explicit too.

Advice should improve accountability, not diffuse it.

Not every decision deserves the same governance

Jeff Bezos popularised the distinction between one-way door and two-way door decisions.

Some decisions really are difficult to reverse.

Buying a business. Entering a new market. Fundamentally changing an operating model. Making a significant regulatory commitment.

They deserve rigorous challenge, thoughtful analysis and appropriate governance.

But many organisational decisions are much more reversible.

• Run a pilot.
• Test a pricing change.
• Trial a new process.
• Launch with one customer.

If it doesn't work, stop. Learn. Adjust.

The appropriate question isn't simply:

“Have we done enough analysis?”

Perhaps it should also be:

“How reversible is this decision, and therefore how much governance does it actually require?”

Treating every decision as though it carries the same level of risk doesn't necessarily create better governance.

Sometimes it just creates bureaucracy.

Accountability needs authority

Another question I've found revealing is:

Does the person accountable for the outcome actually have sufficient authority to deliver it?

Surprisingly often, the answer is no.

An executive owns the result.

Technology owns the developers.

Finance owns the budget.

Operations own the people.

Marketing owns the communication.

Everyone has legitimate priorities and accountabilities of their own.

So the initiative gets approved.

And then very little happens.

Not necessarily because the strategy was wrong, or because anybody isn't doing their job.

The organisation has simply created accountability without sufficient authority or influence over the resources required to deliver.

That's not really a people problem.

It's an organisational design problem.

Governance should design the system, not make every decision

Perhaps this is where the role of governance is most important.

Good governance doesn't have to mean centralising decisions.

It can mean creating sufficiently clear guardrails that decisions can safely be decentralised.

Boards and executive teams can establish risk appetite, financial authorities, regulatory boundaries, customer expectations and ethical principles.

Within those guardrails, capable leaders can then be trusted to act.

That changes the role of governance.

Instead of asking:

“How do we make sure every important decision comes through us?”

we can ask:

“How do we design a system in which good decisions can be made safely without us?”

I think that's a much more interesting question.

There is a human consequence too

Poor decision systems don't just make organisations slower.

They change behaviour.

People quickly recognise when they're being held accountable for things they don't really have the authority to influence.

Over time, they become less inclined to take initiative.

Entrepreneurial thinking becomes harder.

Decisions move upwards.

Senior leaders become increasingly involved in issues that should never have required their attention.

And the organisation becomes slower still.

Which is why decision architecture isn't simply an operational issue.

It's a leadership issue.

A culture issue.

And ultimately, a business performance issue.

The best organisations probably won't eliminate business cases, committees or governance.

Nor should they.

But they may be better at recognising what each is actually for.

• Use analysis to understand uncertainty, rather than pretend it doesn't exist.
• Match governance to the consequences and reversibility of the decision.
• Seek broad expertise without diffusing accountability.
• Give decision owners sufficient authority to act.
• And, where uncertainty is unavoidable, favour learning over increasingly elaborate prediction.

Better decisions don't always require more analysis.

Sometimes they require a better system for making them.