Remix something.
A 10-minute exercise for having an idea before you ask AI to have one for you.
- Time
- 10 minutes
- You will need
- Two bits of paper and a pen
Not every new idea needs to start from nothing.
Some of the most useful ideas are an existing model, behaviour or interaction moved somewhere it hasn't been used before.
I use a version of this exercise as a warm-up in ideation workshops.
It's deliberately quick, slightly stupid and very good at getting people away from the blank sheet of paper.
The concrete story
I once worked on an app that we jokingly described as:
Uber for concrete.
Concrete left in the back of a mixer has a problem: it's going to set.
Meanwhile, somewhere nearby, somebody has a small job that needs concrete but doesn't need an entire truckload.
Connect the two.
One person gets rid of something that's about to become waste. Another gets the amount of concrete they actually need.
Existing model. Different context. New proposition.
That's the basic idea behind this exercise.
Another example
You can see versions of the same provocation elsewhere.
Take the basic interaction idea behind Tinder:
- profiles
- preferences
- browsing
- matching
Now move that somewhere completely different.
Tinder for pets?
You start imagining people browsing profiles of dogs they'd like to spend time with. Which gets you somewhere surprisingly close to services such as BorrowMyDoggy.
The point isn't that one company literally copied another. It's that moving a familiar interaction pattern into an unfamiliar context forces you to notice what is actually useful about the pattern.
Try it.
Make two lists
On one side, write down 10 products, services or experiences with distinctive ways of working. For example:
- Uber
- Tinder
- Spotify
- Airbnb
- Duolingo
- Deliveroo
- Netflix
- Too Good To Go
- Vinted
- Strava
Don't think too hard about the list.
On the other side, write down 10 completely different contexts. For example:
- Pensions
- Pets
- Builders
- Schools
- Caterers
- Care homes
- Dentists
- Festivals
- Divorce
- Local government
Shuffle them
Cut them up, shuffle them, roll dice, close your eyes and point, or just choose randomly.
Take ONE from each list.
Do not pick the combination you think will work best. The awkward combinations are the point.
Make it exist
Give yourself five minutes.
Sketch what the resulting product, service or experience might be. Don't decide whether it's sensible first.
If you get Spotify for dentists, your job is to make Spotify for dentists exist for five minutes.
Ask:
- What gets recommended?
- What gets discovered?
- What's personalised?
- What's a playlist?
- Who shares what with whom?
- What part of Spotify are you actually borrowing?
Do not worry about whether anyone would fund it, buy it or build it. You are exercising your ability to make connections.
Do it again
Pick another pair.
The first combination gets your brain moving. The second and third are usually where it starts getting interesting.
The rule
Don't ask AI to solve the combination for you.
Not yet.
The point of the exercise is not to generate the maximum number of concepts as efficiently as possible. It's to make your own brain travel somewhere it wasn't already going.
Afterwards, absolutely use AI. Challenge the idea with it. Expand it. Research analogous services. Ask what you've missed. Prototype it.
But have the first thought yourself.
Why it works
The blank page asks: "Can you think of something new?" That's a surprisingly difficult instruction.
This exercise gives your brain constraints instead. You already have two things. Your job is simply to work out what happens when they collide.
It forces you to separate a product from its surface features and ask what underlying behaviour or model actually makes it interesting.
- Recommendation
- Matching
- Sharing
- Subscription
- Reputation
- On-demand access
- Habit formation
- Community
Then you transplant that mechanism somewhere else.
The objective isn't necessarily to produce a brilliant idea. It's to practise making unexpected connections.
When you know you're done
Almost every workshop eventually reaches the same point.
Someone looks up and says: "Wait. What about Tinder for [whatever industry we're working in]?"
At which point I generally know we've reached the end of the workshop.