handbooksInside an existing company
Running this where there is already a business to protect. What an innovation team needs, and how to keep new products coming rather than treating each one as a one-off project.
- An established company can do this, but the team needs three things: a small budget it can count on, its own authority to build, and a personal stake in the result.
- Give experiments a sandbox with clear limits, so they can run without threatening the main business.
- Treat innovation as a standing function rather than a project that ends.
- The real obstacle is rarely the method. It is that a new product looks unimpressive next to a mature one.
What an innovation team needs
Ries names three conditions. Teams that lack any one of them tend to stall.
| Condition | Why it matters |
|---|---|
| Scarce but secure resources | A small budget the team can rely on. Less money than the main business, but not money that disappears mid-quarter |
| Independent authority to build | The team can design, build and ship without approval from every function. Cross-functional, and it decides for itself |
| A personal stake in the outcome | People need to care how it turns out, through ownership, recognition, or simply being publicly attached to it |
Large companies find the second hardest to give.
A sandbox for experiments
Rather than asking permission per experiment, agree the boundaries once:
- Which customers experiments may touch, or what share of traffic.
- How long an experiment may run.
- Which metrics it will be judged on, agreed in advance.
- That the same team sees each experiment through and reports the result.
Inside those limits the team moves without asking. Outside them it comes back.
The fear behind all this is real: an experiment can damage a brand or upset customers the main business depends on. A cap on affected customers turns that open-ended worry into a number, which is why limits agreed once beat reviewing every experiment.
Make it a standing function
A company that treats each new product as a one-off project gets an idea, staffs it, ships it, and disbands the team. The learning goes with them.
Keeping a permanent group whose job is starting new things means the skill accumulates. The second product is cheaper than the first, because the people running it have done this before.
A new product measured against a mature one always looks bad. Small revenue, few customers, uncertain future. This is why innovation accounting matters here more than anywhere: without learning milestones, the only available comparison is the one the new product is guaranteed to lose.
Ask whether your innovation team can ship to real customers without approval from outside the team. If not, that is the constraint, whatever else is in the process.