Regression to form

I’ve observed something over the years that I’ve started calling “regression to form”. It’s the depressing inevitability that if you give people something that looks even vaguely like a form, they will treat it as a form.

As a result, the objective shifts from thinking to form-filling. From conversation to completion. From understanding to box-ticking.

Take something like the Business Model Canvas. It’s a brilliant thinking tool – nine interconnected boxes designed to help you explore and visualise how a business works. The whole point is the thinking, the questioning, the iterating. But roll it out in an organisation and watch what happens. Teams start treating it like a form to be filled in. The quality of thinking becomes secondary to whether all the boxes have something written in them. The submission of the completed document becomes a critical milestone

Performance appraisals are probably the worst example of this phenomenon. What were supposed to be meaningful conversations between managers and team members about development, growth, and performance become exercises in prose generation. Both parties know they’re going through the motions, but the form demands to be filled.

This reminds me of Goodhart’s Law: “When a measure becomes a target, it ceases to be a good measure.” We could adapt this for our purposes: “When a thinking tool becomes a form, it ceases to be a good thinking tool.”

The arrival of large language models makes this considerably worse. We now have machines that are exceptionally good at filling in forms. And they will be used to automatically populate these thinking tools that were never meant to be forms in the first place.

Imagine an LLM-generated Business Model Canvas, complete with plausible-sounding value propositions and customer segments, generated from a one-line brief. No thinking required. No human conversation. No iterating or challenging assumptions. Just industrialised form-filling of things that weren’t supposed to be forms.

This terrifies me.

So what can we do about it?

One approach I’ve had success with is to take anything that looks remotely like a form and make sure it looks nothing like a form at all.

Instead of giving teams a canvas template, I’ve created card games. Physical cards or digital versions using tools like Miro. Each card represents a different element and they’re tangible objects that can be moved around, combined, discarded.

The act of physically manipulating these objects changes the conversation. People play rather than fill. They experiment rather than complete. The cards become props for thinking, not boxes to be ticked.

I’ve done similar things with workshops using photographs, Lego, dice and even magnetic rubber ducks to help folk think.

The key is to externalise the thinking in ways that encourage manipulation and conversation rather than completion and submission.

And these days, if you want a record of the conversation, you can record it, transcribe it, and analyse it later. The thinking happens in the interaction, not in the documentation. Post activity documentation is a trivial task with LLMs.

The challenge is fighting the organisational urge to systematise everything. There’s comfort in forms. They’re measurable, comparable, and they scale. But when we turn our thinking tools into forms, we stop thinking.