Innovation framework
Definition
Innovation framework
An innovation framework is the named path one idea walks from concept to launch, with set stages and a decision gate between each. It is not the innovation strategy that picks the arena, and not the innovation portfolio that holds every bet at once.
Stage and gate is the common shape. Robert Cooper published the original version in the 1980s, and most corporate frameworks are a variant of it: a run of work stages, each ending in a go, kill, hold or recycle decision.
The gate is the part firms skip. A stage list with no kill rule is a conveyor belt — and a conveyor belt is how a firm ends up funding nine projects that should have stopped at three.
Key takeaways
- The repeatable path one idea walks, with named stages and a named decision at each gate.
- Every gate carries a go, kill, hold or recycle call and a named person who makes it.
- Frameworks differ by stage count, not by principle; the gate criteria do the real work.
- Without a kill rule the framework becomes a queue and starves the better ideas.
How it works
Work moves one stage at a time and nothing passes a gate without evidence. Stages get longer and more expensive as risk falls, so the cheap tests always happen first. Gate criteria tighten in the same order.
| Stage | Question at the gate | Evidence required | Typical decision |
|---|---|---|---|
| Discovery | is the problem real | user interviews, market sizing | go or kill |
| Scoping | can we serve it | rough concept, rival scan | go or hold |
| Business case | is it worth building | costed plan, revenue model | go, hold or recycle |
| Build | does it work | working product, pilot results | go or kill |
| Launch | will it sell | pricing test, first customers | scale or stop |
| Review | did it pay | results against the business case | keep, fix or retire |
Gate criteria have to be written before the first idea enters. A gate invented on the day turns into a negotiation, and a negotiated gate almost always ends in hold rather than in kill.
Recycle is the underused option — sending an idea back with two specific questions costs a fortnight and avoids the sunk cost argument that follows a reluctant go.
Evidence beats enthusiasm at every gate. The sponsor’s job is to ask what would have to be true for this idea to work, then check whether anyone went and looked.
Regulated sectors run the same shape under law. Drug development is not an innovation framework in the business sense, but it is the clearest example of stages with hard gates between them.
Clinical research runs in three phases and most candidates stop at one of them. On the U.S. Food and Drug Administration’s figures, about 33% of drugs move from phase 2 to phase 3.
That attrition is the point of a gate. Roughly 70% of candidates pass out of phase 1 and about a third of those pass phase 2, so each stage buys information before the expensive one begins.
A framework is not a rebuild of existing work. Business process reengineering changes how today’s operation runs, while an innovation framework governs what new work is allowed to start.
Nor is it a delivery method. Agile transformation changes how a team builds once it has been told to build, and the framework decides whether it gets told at all.
Business process transformation is the downstream consequence — once a gated idea ships, the operating model has to change to carry it.
Firms buy the stages they cannot staff. Innovation outsourcing hands discovery or prototyping to a partner, and product outsourcing takes over the build once a gate has been cleared.
Smaller manufacturers rent the whole method. The Manufacturing Extension Partnership, run by the National Institute of Standards and Technology (NIST), works through nearly 1,400 advisors at more than 450 service locations.
Examples
Consumer goods firms run the classic version. A shampoo line extension passes five gates inside eighteen months, and most of the killing happens at gate two, because concept testing costs almost nothing.
Banks in Singapore and Australia run a shorter variant for digital products — three gates covering problem validated, pilot live with real customers, and commercial case proven.
Government technology programmes gate on maturity rather than margin. A part has to reach a stated readiness level before it may enter an integration stage, which is the whole reason the ladder exists.
An outsourcing provider in Cebu gates automation ideas the same way: two weeks of discovery, a four week pilot on one queue, then a single rollout decision with a named owner.
A medical device maker in Ireland adds a regulatory gate between build and launch, because approval timing rather than engineering sets the launch date in that market.
A software firm in Poland runs a lighter version for internal tools. One gate before build and one before rollout is enough governance for work that never leaves the building.
A retailer in the United Arab Emirates runs the same gates for new store formats as for software. One register and one shared language across very different work is worth more than two perfect models.
Related terms
- Innovation outsourcing: buying a stage of the framework instead of staffing it.
- Agile transformation: how teams build after a gate says yes, not whether they should.
- Business process transformation: the operating change a shipped idea forces downstream.
- Business case template: the document a build gate usually asks to see.
- Knowledge transfer outsourcing: moving what a partner learned back into the firm.
FAQ
What is the difference between an innovation framework and a process?
A process describes how work gets done. A framework adds the decision points, so each stage ends with someone choosing to continue, stop or rework. The decisions are the difference, not the steps.
How many stages should a framework have?
Three to six. Fewer than three leaves no chance to kill a weak idea early, and more than six usually means the firm is reviewing rather than building. Count the gates before you count the stages.
Who owns the gate decision?
One named person with budget authority, supported by a small panel. Shared ownership of a gate is the most reliable way to stop anything ever being killed. Put the name in the document.
Does agile replace the stage gate model?
No. Most firms run both, with agile delivery inside a gated frame: the gates release the funding and the sprints do the building. The two answer different questions.
Can a small company use one?
Yes, and the lighter the better. A two gate version with a written kill rule beats an elaborate model that nobody actually follows.
Find a partner for the stages you would rather not staff in the Outsource Accelerator network of outsourcing providers.
Independent
