How to run an idea challenge that ends in real decisions

If you want to know how to run an idea challenge that produces decisions instead of a spreadsheet nobody opens again, the work is mostly in the framing and the follow-through, not in the collection window. An idea challenge asks a group of employees one focused question for a fixed period, gathers their proposals, evaluates them against what the company is trying to achieve, and closes with a visible decision. This guide walks through each step, from writing the question to telling every author what happened to their idea.

The format works for innovation teams, HR leaders running an employee suggestion program, operations managers chasing cost or quality problems, and leadership teams that want input from people closer to customers. The steps below are the same whether you invite fifteen people or five hundred. What changes with size is how much time you need for merging and scoring.

Why an open suggestion box is not enough

An always-open suggestion box asks people to guess what the company wants. Most will not guess, and those who do send ideas that are hard to compare because they answer different questions. A challenge fixes both problems. It tells people exactly where to point their thinking, it creates a deadline that gets ideas written down, and it gives reviewers a set of proposals that address the same goal and can be ranked against each other.

Write a challenge question people can actually answer

The question is the single biggest lever. A vague prompt such as "How can we innovate?" produces vague ideas. A narrow prompt with a measurable outcome produces ideas you can test. A good challenge question names the problem, the people affected and the result you want, and it leaves the solution open.

• Too broad. "How can we improve customer experience?"
• Too narrow. "Should we add a chat widget to the pricing page?"
• Useful. "How could we cut customer onboarding time in half without adding headcount?"

The useful version tells people what success looks like, sets a constraint that forces creative trade-offs, and does not smuggle in the answer. If your team already uses the "how might we" format, the same rules apply, and our guide on how to write a how might we question covers the common traps. Add two or three sentences of context under the question: why it matters now, what has already been tried, and any hard limits such as budget, regulation or systems that cannot change.

Set the scope and the deadline

Scope answers three questions before anyone submits. Who can take part, what kinds of ideas are in bounds, and what the company is prepared to do with the winners. Be honest about the last one. If the best idea will get a small pilot and a named owner, say so. If there is a budget, state it. People calibrate their effort to the stakes you describe.

For the deadline, two to four weeks of collection is a sensible default. Longer windows rarely add quality, because most submissions arrive in the first and last days. Publish the whole calendar up front, including when reviews will finish and when decisions will be announced. A decision date is a promise, and keeping it is what earns participation in the next challenge.

Invite the team and seed the challenge

The invitation should come from someone with authority over the outcome, not from a generic program mailbox. Keep it short: the question, why it matters, the deadline, what happens to the best ideas, and where to submit. Ask managers to mention it in their team meetings.

An empty challenge page is intimidating. Before you open submissions, add a handful of starter ideas. They show the expected level of detail, they give people something to build on or argue with, and they break the silence that makes the first submission feel risky. Starter ideas can come from the challenge sponsor, from a short internal workshop, or from AI suggestions that you then edit. In Brainstormer, admins can generate starter ideas for a challenge and keep the ones worth publishing, and they can also run an AI strategy workshop on the same question to surface angles from several perspectives before the team joins.

Collect ideas in a consistent shape

Comparing ideas is much easier when each one answers the same few questions. A short template works well:

1. What problem does this solve, and for whom?
2. What is the proposed solution?
3. What would a first, small step look like?
4. How would we know it worked?

Do not make the template a barrier. Many good ideas arrive as one sentence from someone who has lived the problem for years. Accept the short version and fill in the rest later, either by asking the author or by having AI draft the missing parts for the author to confirm. What matters is that every idea reaches review with a problem, a solution, a first step and a success measure written down, so reviewers are comparing like with like.

Keep the collection open and visible. Letting people see and vote on each other's submissions increases the quality of later ideas, because authors build on what is already there rather than repeating it. Voting is a useful signal of what resonates with colleagues, but it is not the decision. Treat it as one input among several.

Merge duplicates and group ideas into themes

Any challenge with more than a few dozen participants will produce duplicates. Five people will propose the same self-service onboarding checklist in five different words. If you score those separately, you waste reviewer time and split the votes of the idea that clearly has the most support.

Merge duplicates into a single idea that keeps the strongest wording, credits every author, and combines their votes. Then group the remaining ideas into themes, such as "self-service setup", "sales handoff", "documentation" and "account configuration". Themes help reviewers see where energy is concentrated and often reveal that the real answer is a combination of three smaller ideas. Doing this by hand is tedious at volume, which is one of the jobs AI does well. Brainstormer flags likely duplicates for an admin to merge and proposes themes that the admin can rename or rearrange.

Evaluate ideas against your goals

Evaluation is where most challenges drift into opinion. The fix is to decide the criteria before you read a single idea, and to tie them to the company's strategic goals rather than to general enthusiasm. A practical set is impact on the stated goal, cost to implement, risk, and strategic fit, with weights that reflect what leadership cares about this year.

Ask for a written reason next to every score. A number without a reason cannot be challenged or learned from, and it is the reason that authors will want to read later. If you want to go deeper on building the scorecard, see idea evaluation criteria and the weighted scoring model. AI scoring is a useful first pass because it applies the same criteria to every idea and writes out its rationale, which reviewers can then accept or overrule. It should never replace the people who own the decision.

Build a shortlist

The shortlist is the bridge between a long list and a decision. Aim for a number that decision makers can actually discuss in one meeting, usually three to eight ideas. Rank by weighted score, then look at the list as a portfolio. You may want one quick win that can ship within a month, one larger bet, and one idea that is cheap to test even if its score is lower. Note anything you are deliberately leaving off and why, because those notes answer the question "what happened to mine?" before it is asked.

Decide and close the loop with every author

A challenge ends with a decision, not a ranking. For each shortlisted idea, record one of a small set of outcomes: approved with an owner and a first step, parked with a reason and a date to revisit, or rejected with a reason. Move approved ideas into the pipeline where implementation is tracked, so the challenge connects to real work rather than ending at the announcement.

Then tell every author what happened, including those whose ideas did not make the shortlist. A short, specific note is enough: the decision, the reason, and whether the idea might return in a future challenge. This feedback is the single strongest predictor of whether people will bother next time. Publish a summary for the whole company with the winning ideas, the themes that emerged, and the next steps. A board report per challenge, with the ranking, themes, stage counts and a recommendation, makes the same summary easy to share upward. Brainstormer produces one that can be printed to PDF or exported as Markdown or CSV.

Common mistakes to avoid

• A question with no measurable outcome. People cannot aim at "innovation". Name the result you want.
• No decision date. Without it, review slips, authors hear nothing, and the next challenge gets a fraction of the entries.
• Treating votes as the decision. Popular ideas are not always the ones that fit the strategy. Use votes as a signal only.
• Scoring before merging. Duplicates split support and waste reviewer time. Merge first, then score.
• Criteria invented after reading the ideas. This lets reviewers justify favorites. Fix criteria and weights in advance.
• Silence for rejected ideas. Everyone who submitted deserves a reason. It is also how people learn what a strong idea looks like.
• No owner for approved ideas. An approved idea without a named owner and a first step is a rejected idea with better manners.

Much of this is covered in more depth in why employee suggestion programs fail, which is worth reading before your first challenge.

Where to start

Pick one real problem that leadership cares about this quarter, write the question with a measurable outcome, set a four week calendar with a public decision date, and seed the page with a few starter ideas. Run it end to end once, including feedback to every author, before you scale to several challenges at a time. If you want the collection, duplicate merging, scoring against goals, pipeline and board report in one place, Brainstormer's innovation management software is built around exactly this cycle.