From 40 ideas to 3 concepts in one afternoon
The wall from last week is still there: forty notes, a photo in the team chat. On Friday the sponsor will ask "so what are we doing?", and "we have forty ideas" is not an answer. This is the funnel between the ideation workshop and the decision workshop: about three hours, the core team of four to six, three concepts at the end.
Why a funnel and not a vote
Forty ideas cannot be decided; half are variations of each other and a few nobody can explain any more. A dot vote on that wall picks the catchiest wording, not the most important problem. The funnel narrows in six stages, each with a time box and a number going out. If a stage runs over, cut the discussion, not the stage.
The six stages
1. Clean the list (30 minutes, 40 to about 30)
Read every note aloud, one sentence each. Merge duplicates, split the notes that are two ideas, drop the ones nobody in the room can explain. Say it out loud: "nobody can explain this one, it goes."
2. Cluster by problem, not by solution (30 minutes, 30 to about 8)
Ask of every idea: which problem does this solve, for whom? Group by the answer. "Chatbot for returns" and "return label in every parcel" belong together because they attack the same pain. Each cluster gets a problem headline, never a solution name.
3. First cut with the matrix (30 minutes, 8 to 5)
Score the clusters, not the ideas. Impact: how much does solving this problem matter? Effort: what does it take to learn whether a solution works? Score silently, place the eight on the impact and effort matrix, keep the five the team would defend to the sponsor.
4. Build the concepts (45 minutes, 5 concepts)
For each cluster, combine its ideas into one concept statement: what it is, for whom, how it is different from what exists today. Three sentences. Then name the riskiest assumption, the one thing that kills the concept if it is false. That is what makes it a concept: a target, a difference and a known risk.
5. Score and discuss the gaps (45 minutes)
Agree three criteria before anyone scores, for example strategic fit, evidence of customer need, time to first learning. Everyone scores the five silently from 1 to 5. Discuss only where scores differ by two or more.
6. Pick three, park two (15 minutes)
Pick three for concept one-pagers. Park two with a written reason and a revisit date. "Parked until the pricing project is done, revisit in October" is a decision. "Maybe later" is a leak.
| Stage | In | Out | Time | Tool |
|---|---|---|---|---|
| 1 Clean | 40 ideas | about 30 ideas | 30 min | wall and pen |
| 2 Cluster | 30 ideas | about 8 clusters | 30 min | problem headlines |
| 3 First cut | 8 clusters | 5 clusters | 30 min | impact and effort matrix |
| 4 Build | 5 clusters | 5 concepts | 45 min | concept statement |
| 5 Score | 5 concepts | 5 scored | 45 min | three criteria, silent |
| 6 Pick | 5 concepts | 3 for one-pagers, 2 parked | 15 min | reason and revisit date |
Rules for the afternoon
- Nobody defends their own idea in stages 1 to 3. Ideas have no owners yet; they belong to the wall.
- Merging keeps the best wording and all the names. Whoever contributed to a merged idea stays on it.
- A cluster with one idea is allowed. Forcing it into the wrong cluster only hides it.
- Parked is written down with the reason. Tools like IdeaHone carry ideas into concepts with context and contributors; a photo of the wall does not.
Do stage 1 on the day of the workshop, while memory is fresh and everyone still knows what the cryptic note meant. It halves the afternoon a week later.
The afternoon produces concepts, not decisions. The decision belongs to the sponsor in the decision workshop, with the one-pagers on the table.
Where the funnel goes wrong
- Clustering by department. "Marketing ideas, IT ideas, ops ideas" is sorting, not clustering. Ask what problem each idea solves.
- The loud idea becomes its own cluster. The idea someone pitched three times gets a cluster of one with a solution name. Put it under the problem it solves and let it compete.
- Scoring 40 ideas instead of 8 clusters. Forty rows times six people is a spreadsheet, not a conversation.
- Three concepts that are the same concept in three sizes. Small, medium and large of one idea is one concept with an open scope question. The other two slots go to different problems.
- No revisit date for the parked ones. Without a date, parked means dead, and whoever had the idea knows it.
Score and prioritise ideas with your team
IdeaHone scores ideas against your criteria, shows them on a matrix and keeps the reasons, so the decision is clear and traceable.
How do you narrow down ideas after a brainstorm?
In time-boxed stages: clean the list, cluster by the problem solved, cut the clusters with the impact and effort matrix, build one concept per cluster, score against three agreed criteria, then pick three and park the rest with a reason and a date.
What is the difference between an idea and a concept?
An idea is a single suggestion, often one note on a wall. A concept combines related ideas into one statement (what, for whom, how it is different) and names the riskiest assumption behind it. Concepts can be decided on; ideas can only be collected.
How many concepts should go to a decision?
Three. Fewer leaves the sponsor no real choice, more turns the decision workshop into another scoring round. Each gets a one-pager so they are compared on the same facts.
