The problem with optimisations are unknown-unknowns. These are things that you don’t know you don’t know.
Say, for example, you want to go abroad on a road trip. You have the known-known, of how many guests vs car seats there are, the known-unknown of how much fuel you’ll need and the unknown-unknown that you’ll need a visa for the country you are visiting.
The known-known is fine, as long as there aren’t more guests than seats the road trip is good to go.
The known-unknown can be mitigated. You can either work out how much fuel you’ll need before you leave and ensure you have it or you can check that there are places to refuel along the way. Doing either will work.
The unknown-unknown will sink the road trip. You’ll have no way to mitigate it until you first learn that it exists. At the point it becomes a known-unknown, you’ll be at the border, car fully loaded, unprepared and on a timeline you can’t change.
That’s why optimisations are hard. It’s (relatively) straightforward to improve upon known-unknowns to optimise something, but to truly optimise something, you need to improve upon the unknown-unknowns.
By definition, you won’t know what you don’t know, but someone else might. By collaborating with others from different fields or those with very different perspectives, you’ll be more likely to uncover unknown-unknowns and create better innovations.
I recently posted about this in the N8 AgriFood Food Systems Policy Hub, regarding the way in which farmers aren’t brought into the conversation with agricultural innovations. By not involving them, unknown-unknowns are often only uncovered when the innovation is in its final stages, making applying that fix needlessly more difficult and expensive to do.
Tools to uncover unknown-unknowns
Involving others from different fields as early as possible in a project is a simple way of uncovering unknown-unknowns. It’s not high-tech or new, but it works. It works particularly well when the meeting is facilitated or follows a stakeholder development framework.
Researching failure, using the premortem technique or simply researching the mistakes of others are both ways to get lists of potential unknown-unknowns to then compare with your project. At a minimum, you’ll have alternative perspectives to asess your project through.
Framing is one of the most useful tools, but the hardest to use well. To me this always sounds dangerously vague, but when wielded correctly, actually creates the most clarity.
Framing is where you ask a series of open-ended, but directed questions. These ensure that you are trying to solve the most fundamental problem and so can create the best innovation.
For example, say you are trying to win the Le Mans 24hr race.
A reasonable question might be how can we get the car to go faster?
Create a more powerful engine, ensure that all of the power can be put down to accelerate the car, make sure the driver has the best training, etc.
A better-framed question would be: How can we get the car around the track faster?
This is a much more interesting question. The bendy bits of race tracks are the most problematic and it’s a problem that going faster won’t solve, going faster will actually make it worse.
In fact, improving the car’s breaking so that it can break later and still get around the corners will decreasing its lap time. Improving the car’s ability to slow down would actually increase its average track speed.
This is exactly what Jaguar did in 1953 when they won the Le Mans 24hr race by 4 laps, half an hour ahead of the next car to finish. The next year, everyone had adopted Jaguar’s innovation of disk breaks.
The hard thing with unknown-unknows is that you can’t confirm how many remain uncovered and trying to uncover every last one will likely bring any project to a standstill. The important thing is to acknowledge that they exist, and knowing this, trying to uncover as many as possible. They can then be fed back into whatever system you are using to ensure you are addressing the most fundamental problem possible, and so, creating the best innovation.