Stick, Twist, Fold: Clearer Innovation Experiment Decisions
Make a clear call on every experiment at every checkpoint: stick, twist or fold. This simple rule for innovation experiment decisions keeps effort on what works and makes stopping a normal, healthy outcome.
Why experiments drift without a decision rule
Experiments drift when nobody has to decide. Teams keep going because stopping feels like failure. Budget and attention spread across too many half-alive ideas.
Two problems make this worse:
- No structure. Teams experiment independently. Results can't be compared, and there is no shared way to pick which to fund.
- No focus. Teams default to the easiest experiments. The most valuable ones get less effort than they need.
Innovation programmes rarely fail because of bad ideas. They fail because of organisational gravity. A clear decision rule is one way to push back against it.
Stick, twist or fold: three choices
Every experiment gets one of three decisions at each checkpoint. The names come from a card-game decision. Everyone understands them straight away.
- Stick: it's working. Scale it.
- Twist: it's promising. Adjust it and iterate.
- Fold: it isn't working. Stop and redirect the effort.
The language matters. Three plain words give the whole organisation a shared vocabulary. People can discuss a decision without defending a reputation.
Each decision has its own signals:
- Stick when customers keep returning to the same value, in their own words, across a round.
- Twist when customers like the need it meets but stumble over how it works.
- Fold when reactions stay polite and flat, and the why never gets stronger.
Every decision should lead somewhere. A stick moves the idea towards build. A twist sets the change for the next round. A fold sends the effort to a better bet.
Why normalising stopping matters
Stopping is a result. A fold frees money, people and senior attention for something with more promise. That is good use of an innovation budget.
Stopping needs to feel safe. If folding an idea looks like personal failure, teams will avoid it. Weak experiments then linger and drain the strong ones.
Twisting protects good ideas from early exit. First attempts are often rough. Without structured iteration, teams can conclude an idea isn't ready and abandon something that could have worked.
How to make innovation experiment decisions at each checkpoint
Base each decision on customer evidence. Customers settle debates that internal opinion can't. Their reactions give every checkpoint a shared basis.
Use these questions at every checkpoint:
- What did customers react to? Look at what they valued and the reasons they gave.
- Is the signal consistent? Check whether the same themes repeat across one audience.
- Have we proved the value first? Ask whether customers want it before asking if we can build it or make money from it.
- Did the team hear it directly? People back what they've heard themselves.
Bring the team into the room. A weekly workshop means the people making the call have seen the evidence. Decisions land faster because nobody has to be persuaded second-hand.
Linking decisions to your strategy
Stick, twist or fold works best inside a clear innovation strategy. Random experiments that don't connect to what the business needs next are hard to judge. A strategy gives each checkpoint a reference point.
Keep experiments close to the business. A separate lab can drift from commercial reality. Its work becomes impressive but irrelevant.
Build the habit internally. Programmes that depend on permanent outside support get cut when budgets tighten. When your own teams own the decision rule, the capability stays after we leave.
Make the call, every time
Good innovation programmes make clear decisions often. Stick, twist or fold gives your teams a simple way to do that and makes stopping part of the process.
If your experiments need sharper decisions, we'd be glad to help. Get in touch to talk about how we run checkpoints inside a sprint.
Read next: concept testing four rounds · ai adoption framework four forces · sprints