TBM 50/52: Fragile Models

Scenario…

I do a lot of research. The research is messy! It looks like this:

I notice three broad groups:

I take those three broad groups, and create a model/tool. For example, I generate three personas to adapt messaging for different buyers.

I leave the company. My research rots in a folder somewhere. All that remains are the three personas and the messaging framework.

From this point forward, there are three personas in the official Persona Deck.

Say the messaging framework is wildly successful. At that point, you almost don’t even need three personas. One will suffice. If a salesperson (or writer, or designer) can master The Message, they can do a fine job. They might be able to ignore differences altogether.

How convenient is that? One message (or design) for everyone!

“All models are wrong, but some are useful,” said George Box. My model is useful. But without a way back to my research, and to my thought process, things are fragile. And risky!

• No one can see how I got from A to B (or A-Z to A,B,C)

• People start believing the model without the necessary caveats.

• People lack context. They forget why I designed the model.

• They can’t add to my research and/or see how the model changes

• They can’t challenge my work, especially if the end-result “seems reasonable”

• There might not even BE any research. Who would know?

If the nature of the game/activity changes, and people need more in-depth awareness...they will not have it. They could do bad things. Designing for the average...not the edges. Overconfidence. It could lead to marginalization and malpractice. Silencing people.

It could lead to dangerous mono-process (see this new video of mine).

So...

When making models, watch out.

Useful models can be especially dangerous because we equate usefulness to truth.

Make the “Job” of the model explicit. Target customers? Messaging that resonates? Guide design decisions? Make that clear.

Leave a pathway back to the research.

Model with care.