Insights
How to Kill an Idea Properly
The most valuable thing a development process does is stop things – early, cheaply, and on evidence
Most organisations describe their new product process in terms of what it builds. The better question is what it stops, and how cheaply.
A process that has never killed a project at a gate is not a process. It is a queue with paperwork.
The eight-stage decision process
The sequence below is Kotler’s. Each stage has a question, and each question has a real no.
| # | Stage | The question that must be answered yes |
|---|---|---|
| 1 | Idea generation | Is the idea worth considering at all? |
| 2 | Idea screening | Is it compatible with our objectives, strategies and resources? |
| 3 | Concept development and testing | Can we find a concept consumers say they would try? |
| 4 | Marketing strategy development | Can we find a cost-effective, affordable route to market? |
| 5 | Business analysis | Will this product meet our profit goal? |
| 6 | Product development | Do we have a technically and commercially sound product? |
| 7 | Market testing | Have product sales met expectations? |
| 8 | Commercialisation | Launch. |
What matters is not the list. It is that a no at any stage has three legitimate destinations, and only one of them is failure.
- Drop. The idea does not work and will not. Stop spending.
- Send back. Return to product development, or modify the marketing programme, and re-enter the gate.
- Lay future plans. Keep it for future reference. The idea is sound and the timing, capability or capital is not. This is the option most processes lack, and its absence is why good ideas get destroyed rather than parked.
What a real gate looks like
A working stage-gate has three things most diagrams omit.
- A named document at every stage. An idea brief. A project plan. A concept brief. Minutes of the review meeting. An evaluation report. If a stage produces no artefact, it did not happen – it was a conversation.
- A defined re-engineering loop. Not everything that fails a gate should die. The loop back into development, with the reason recorded, is what turns a filter into a pipeline.
- A file of evidence. A single location holding the proof of compliance with the process for every project. Unglamorous, and the only thing that makes the process auditable rather than aspirational.
What the gates actually test
Grouped by the three questions underneath the eight stages.
Gate one: should this exist?
- Where did the idea come from – a competitive product, a concept from another market, a modification of something existing, market feedback, a research partner, or a genuinely new category?
- Product category, target market, differentiation point, initial market size, competitive structure – and critically, manufacturing capability.
- The output is a concept brief that either continues, re-engineers, or gets kept for future reference.
Gate two: can we actually make it?
- Recipe or specification, and the process technology required.
- Make or outsource – and the capital implication of each.
- Packaging category and design, shelf life, quality control.
- Operations capability across supply chain and manufacturing.
- Cost evaluation, run properly, before the concept is finalised rather than after.
Gate three: will the market take it?
- Product testing with real users, and consumer evaluation of the finished thing.
- Packaging concept tested, not assumed.
- Alignment with the target market as defined at gate one – has it drifted?
- Then registration, supplier contracts, final samples for quality control, sensory evaluation and a trial batch.
The filter that actually filters
One discipline separates processes that work from processes that theatre.
A filter stage that has never killed anything is not a filter.
Filter research is cheap relative to development. Its entire purpose is to reduce the pipeline. If every concept that enters your screening stage emerges from it, the stage is adding cost and delay while removing no risk.
The related discipline: run the pricing and profitability study as soon as a concept survives the filter, including production set-up and any additional people required. Skipping it is how a product reaches launch before anyone establishes that it cannot make money at a price the market will pay.
After launch, the step everybody drops
Two evaluations, and they answer different questions.
- Project evaluation. How did the process itself perform? Planning, milestones, what we learned about running a launch. This improves the next project regardless of what was launched.
- Performance evaluation. Manufacturing issues, supply chain issues, quality issues in trade, sales performance, consumer feedback. This tells you whether the product is working.
Without both, an organisation does not learn. Each launch teaches only the person who ran it, and that knowledge leaves when they do.
The question worth asking your own business
Not “what is our process”. Everyone has a diagram.
When did we last formally stop a project at a gate – and what did it cost us to have got that far?
If nobody can remember, the gates are not gates. They are milestones, and milestones only ever point one way.