Most idea campaigns fail before they start. Not because people don't have good ideas, but because the question they were asked was either too vague to be useful or too narrow to spark anything new.
This guide gives you a simple framework for writing an Idea Challenge that pulls in high-quality, relevant submissions and keeps people engaged from launch to close. It covers the five components every strong challenge needs, the one sentence that doubles submission quality, ready-to-use templates, sector patterns, warning signs, what to do during the live window, and an FAQ.
Why the question matters
The framing of your Idea Challenge determines the quality of what you receive. A vague prompt like "share your ideas for improving our processes" tells people nothing. A prompt like "ideas to reduce changeover time on Line 3 by 15% using existing equipment" is so narrow it excludes half the people who might contribute.
The sweet spot is a focused challenge with room to breathe. Specific questions consistently outperform generic ones on submission rate, idea quality, and implementation rate, because they signal what kind of expertise is being asked for and the right people self-select in. The opposite mistake also happens: a challenge that already specifies the solution ("we need ideas for installing sensor X on equipment Y") is a procurement request, not a challenge. Leave room for the answer while constraining the question.
The anatomy of a good idea challenge
Every strong Idea Challenge has five components. Each one earns its place; the challenge weakens when any of them is missing.
| Component | What it does | Length | Failure mode if missing |
|---|---|---|---|
| 1. Context | Why this matters now | 2-3 sentences | Submitters can't tell whether their idea is wanted; engagement drops |
| 2. Core question | Single, specific, answerable question | One sentence | Bundled questions produce split submissions |
| 3. What you're not looking for | Rules out off-scope ideas upfront | 2-4 bullets | Evaluators waste time triaging things that should never have been submitted |
| 4. What good looks like | Calibrates the size and shape of an answer | 1-2 sentences | Submissions cluster at the wrong altitude |
| 5. What happens next | Names the review timeline and feedback path | 1 sentence | Submitters who have been burned by silence don't bother |
Context. Two or three sentences on the situation you're trying to improve and why it matters now.
Core question. One specific sentence. "How might we reduce the number of customer complaints reaching the support team?" beats "How can we improve customer satisfaction?"
What you're not looking for. Most people skip this and pay in evaluation time. If budget is approved for a new system, say so. If solutions must work without IT involvement, say that.
What good looks like. Quick wins testable in a week, or bigger bets across a year? Cost reduction or speed? Calibrate without locking people in.
What happens next. "We'll review all submissions by [date] and share the outcome with everyone who participated." Without this line, most people will skip the challenge. The feedback that builds trust guide has templates for the closing message.
The one sentence that doubles submission quality
Before you publish, add this sentence somewhere visible:
We'll act on ideas that [insert your actual criteria here].
Fill in the blank honestly. "We'll act on ideas that can be tested in under 60 days without a budget increase." Forcing yourself to complete that sentence forces clarity on what you're really looking for. When people see it, they self-select - half-formed thoughts get sharpened or held back. Both are good outcomes.
Resist aspirational filler ("ideas that fundamentally transform how we work"). The disciplined version names a specific constraint: budget, timeline, scope, or technology.
Five ready-to-use templates
Pick the closest fit, fill in the bracketed sections, and adapt to your operating vocabulary.
- Process improvement. How might we reduce [specific problem] in our [process] without [key constraint]? Not looking for: ideas requiring new software or that we have already tried. Good = testable within [timeframe] with [resource constraint].
- New product or service. What new [product, service, feature] could we offer to [customer segment] to help them [job to be done]? Not looking for: incremental variations on what we sell. Good = could reach [revenue target] within [timeframe].
- Cost reduction. Where could we eliminate [waste, redundant spend] in [area] without affecting [key outcome]? Not looking for: shifting costs elsewhere. Good = savings of at least [threshold] per year, implementable within [timeframe].
- Customer experience. How might we make [touchpoint] easier or faster for [user type]? Not looking for: solutions that add steps. Good = testable with a small group within [timeframe].
- Safety or quality. What changes to [process, equipment, behaviour] could reduce [incident, defect] in [area]? Not looking for: ideas already in the improvement plan. Good = a clear mechanism for preventing recurrence.
Sector-specific patterns
The five components are universal, but framing differs by sector. Copying a manufacturing-style challenge into knowledge work usually produces a low submission rate.
| Sector | Framing that works | Framing that fails |
|---|---|---|
| Manufacturing & operations | "How might we reduce changeover time on Line 3 during the night shift, using existing tools?" | "How can we improve operational efficiency?" |
| Retail & customer-facing | "How might we reduce queue length at peak times in stores under 200m2?" | "How can we improve the customer experience?" |
| Healthcare & clinical operations | "What non-clinical operational changes would let us free up 15 minutes per shift on the ward?" | "How can we improve patient outcomes?" |
| Office & knowledge work | "How might we reduce approval steps in the expense process for travel under 500 EUR?" | "How can we improve productivity?" |
| Public sector & regulated | "What operational changes can a local team implement without a policy update to reduce citizen wait time at the service desk?" | "How can we modernise public services?" |
In manufacturing, reference a specific line, shift, or product family. In healthcare, separate clinical-workflow ideas (which need clinical sign-off) from operational ideas (which don't); mixing them stalls evaluation. In knowledge work, target a specific process, tool, or recurring meeting. In public sector, keep operational improvement separate from policy change; blurring the line produces submissions the platform cannot act on.
Warning signs to check before publishing
- The question has more than one "and" in it. Two questions = split into two campaigns. Combining ("reduce cycle time and improve quality and reduce cost?") feels comprehensive but produces less focused submissions.
- You haven't thought about what you'll do with the answers. If you can't describe what happens to submissions in 60 seconds, you're not ready to launch. The 2-hour triage method and 3 scoring models are the operational backbone of that plan.
- Only senior people would know the answer. The best Idea Challenges surface what someone on the front line knows that leadership doesn't. If only executives can answer, you're running a meeting, not a campaign.
What to do during the live window
The challenge text is necessary but not sufficient. The next two weeks are where most challenges drift, and the right cadence makes more difference than the original phrasing. The campaign communication templates have wording for each touchpoint.
- Day 1 - launch: reach the audience through the channel they actually use. Operators read shift-handover boards; retail colleagues read break-room posters; knowledge workers live in their messaging tool.
- Day 3 - midweek update: a short note acknowledging early submissions and naming an emerging theme signals attention.
- Day 5 - targeted reminder: nudge people who haven't submitted. Avoid blanket reminders that hit everyone.
- Day 7 - close and thank-you: name the number of submissions, the next step, and when submitters will hear back. The campaign momentum guide covers the cadence in more detail.
Frequently asked questions
How long should an idea challenge stay open?
Two weeks is the sweet spot for most internal campaigns. Shorter (5-7 days) works for urgent situations or already-engaged groups. Longer than three weeks risks losing momentum.
Should you offer incentives?
Build them in from the start and announce upfront ("top three ideas will be featured in next month's newsletter"). Avoid cash prizes - they attract people chasing money rather than solving real problems. Never add incentives mid-campaign; it signals the challenge is struggling.
What if you get very few submissions?
Low volume usually means one of three things: the challenge was too narrow or unclear, people didn't know about it, or they didn't trust that ideas would be read. The first two are fixable with better communication next time; the third requires publicly closing the loop on a previous campaign before launching another. The 20-question diagnostic covers how to identify your bottleneck.
How do you handle duplicate ideas?
Duplicates are a feature, not a bug - they tell you that multiple people care about the same problem. Group similar ideas during evaluation and note how many people submitted variations. Credit the person who submitted first, but acknowledge all contributors.
Can you run multiple challenges at once?
Yes, but carefully. Two simultaneous challenges dilute focus; three or more and engagement drops on all of them. Stagger by at least two weeks so each gets a clear communication cycle.
Should the question be written centrally or locally?
Almost always by local operational leaders. Central staff produce questions that read well in a strategy document but don't name the operational specifics the audience recognises. The strongest challenges are written by people closest to the work, lightly edited centrally for clarity.
What if the topic is sensitive (safety, culture, leadership)?
Lean into the sensitivity rather than around it. Acknowledge that the topic is sensitive, that submissions can be anonymous, and that the response will be specific. Honestly framed sensitive challenges produce more substantive submissions than sanitised ones.
How do anonymity, AI, and works councils factor in?
Safety, ergonomics, and culture-related challenges benefit from anonymity because employees are reluctant to attach their name to observations about someone else's behaviour. Operational challenges typically don't need it. AI is useful for clustering submissions, flagging duplicates, and surfacing themes, but should not autonomously decline ideas (GDPR Article 22 prohibits fully automated decisions with effect on individuals). In Germany (BetrVG Section 87), France, the Nordics, and several other European jurisdictions, employee representatives have legally protected co-determination rights over systems that touch performance or behavioural data; engage the works council before launch where relevant.
How do we know whether a challenge has worked?
Three signals. Submission rate above 25% of the invited audience (lower means the question or trust was the problem). Implementation of at least one idea inside the cycle (lower means evaluation is the bottleneck). Visible feedback to every submitter (lower means the loop is broken). If all three are positive, plan the next challenge inside 6-8 weeks. The measurement guide covers how to compute each signal.
Related guides
- Your first idea challenge: from question to decision in 10 days
- Who should you invite? The audience planning worksheet
- The communication template pack: 4 emails every campaign needs
- The idea submission template: what to ask
- How to measure your innovation programme
- How to triage 100+ ideas in 2 hours
- The idea scoring scorecard: 3 models for different situations
