| name | outcome-driven-innovation |
| description | Apply Anthony Ulwick's Outcome-Driven Innovation (ODI) — the eight-step method that turns "what customers want" from guesswork into a quantitative process. Capture **Jobs, Outcomes, and Constraints** instead of customer opinions; rank them with the **Opportunity Algorithm** (`importance + max(importance − satisfaction, 0)`); and score every concept on the Customer Scorecard. Use when user mentions "outcome-driven innovation", "ODI", "Ulwick", "opportunity algorithm", "outcome statement", "desired outcomes", "underserved / overserved", "outcome-based segmentation", "customer scorecard", "jobs to be done" *for product development specifically*, "what customers want", or "Strategyn". Also trigger when planning a new product or feature, auditing why an R&D pipeline is producing duds, designing a customer interview to extract metrics (not opinions), prioritising opportunities, segmenting a market that demographics won't split, repositioning a product whose true value is hidden, or evaluating concepts before build. |
| metadata | {"authors":"Anthony W. Ulwick","source":"What Customers Want — Using Outcome-Driven Innovation to Create Breakthrough Products and Services, McGraw-Hill, 2005","version":1} |
Outcome-Driven Innovation (ODI)
In any other business process — manufacturing, finance, sales — a 50–90% failure rate would be unacceptable. In innovation, it has been tolerated for decades. Ulwick's claim: that failure rate is not a creativity problem, it is a process problem. The customer-driven movement asks customers what they want and dutifully builds it — but customers volunteer solutions, specifications, needs, and benefits, none of which are useful inputs for innovation. ODI replaces them with three inputs that are — Jobs, Outcomes, Constraints — and ranks them with a single equation.
Core Principle
"Customers buy products and services to help them get jobs done." The job is the unit of analysis, not the customer. Customers use 50–150 outcomes (not a handful) to judge how well any job is getting done — they have them in their minds but seldom articulate them. The job to capture them belongs to the company, not the customer. Once captured, the Opportunity Algorithm mathematically reveals which outcomes are underserved (high importance, low satisfaction — targets for innovation) and which are overserved (low importance, high satisfaction — targets for cost reduction). Everything downstream — segmentation, targeting, messaging, prioritisation, ideation, concept evaluation — becomes derivative of these rankings.
The mantra: Silence the "Voice of the Customer." The literal voice of the customer sidetracks innovation because customers are not qualified to know what solutions are best — that is the job of the organisation. Listen instead for the metrics customers use to judge value.
The Eight-Step ODI Process
Apply in order; each step's output is the next step's input.
- Formulate Innovation Strategy — decide which of the four innovation types (Product/Service, New-Market, Operational, Disruptive), which of the four growth options, and which customer in the value chain to target.
- Capture Customer Inputs — collect three input types: Jobs (functional + emotional [personal + social]), Desired Outcomes (50–150 per job), Constraints (roadblocks to getting the job done). See references/jobs-outcomes-constraints.md.
- Identify Opportunities — score every outcome by importance + satisfaction; rank with the Opportunity Algorithm. See references/opportunity-algorithm.md.
- Segment the Market — group customers by unmet outcomes, not demographics. Reveals "segments of opportunity" demographics cannot find. See references/segmentation.md.
- Define Targeting Strategy — choose which underserved outcomes to attack (broad-market or segment-specific) and which overserved outcomes to strip cost from.
- Position Current Offerings — re-message existing products around the underserved outcomes they already satisfy (Cordis went from 1% → 5% market share in six months on messaging alone).
- Prioritise the Development Pipeline — re-rank in-flight projects by how many targeted opportunities they address; kill the rest.
- Define Breakthrough Concepts — focused brainstorming on remaining unmet outcomes; evaluate with the Customer Scorecard before build. See references/customer-scorecard.md.
The Opportunity Algorithm (the load-bearing equation)
Opportunity = Importance + max(Importance − Satisfaction, 0)
Importance and Satisfaction are each the % of customers rating that outcome 4 or 5 on a 1–5 scale, placed on a 10-point scale (75% → 7.5). Score ≥ 12 = ripe opportunity; 10–12 = worth pursuing; < 10 = ignore. Overserved outcomes (low importance, high satisfaction) become cost-reduction targets, not feature targets. This single formula replaces the entire debate about what to build next.
The Input Taxonomy (this is the whole reframe)
Customers volunteer five things when asked for "requirements"; only the bottom three are useful. Train yourself to translate up the stack.
| Input type | Example (razor) | Useful for ODI? |
|---|
| Solution | "Add a lubrication strip" | ❌ Customer-as-engineer; produces me-too products |
| Specification | "Lighter weight" | ❌ Locks in one implementation |
| Need | "Reliable, durable" | ❌ Adjective — un-measurable |
| Benefit | "Faster", "easier to use" | ❌ Vague — 21 definitions of "easy to use" in one Motorola study |
| Constraint | "Can't get a signal indoors" | ✅ Roadblock to job — innovation target |
| Job | "Remove facial hair" | ✅ The unit of analysis |
| Desired Outcome | "Minimise the number of nicks when shaving" | ✅ Measurable, stable over time, drives every downstream step |
Outcome statements have a strict format: Direction (Minimise / Increase) + Unit of Measure (time, number, frequency, likelihood) + Outcome Desired. Only "minimise" and "increase" — words like "reduce", "eliminate", "prevent" measurably skew importance ratings.
Eight Sources of Variability (what ODI fixes)
The reasons 50–90% of innovation initiatives fail, per Ulwick — every one mapped to an ODI step:
| Failure mode | Fixed by step |
|---|
| Ill-conceived growth strategies | 1. Formulate Strategy |
| Faulty data collection (Voice of the Customer) | 2. Capture Inputs |
| Missed opportunities | 3. Identify Opportunities |
| Poor market segmentation | 4. Segment |
| Wrong growth targets | 5. Target |
| Unfocused marketing / messaging / branding | 6. Position |
| Poorly prioritised development | 7. Prioritise pipeline |
| Scattershot idea generation | 8. Define Concepts |
Common Mistakes
| Mistake | Why It Fails | Fix |
|---|
| Asking customers what they want | Customers offer solutions and benefits, not the metrics they use to judge value. | Ask "how do you measure success at this job?" — extract 20–30 outcomes per interview. |
| Capturing < 20 outcomes per job | A handful of outcomes hides the underserved ones — every job has 50–150. | Keep extracting until no new outcomes surface across 20–30 interviews. |
| Brainstorming first, validating later | You generate hundreds of ideas, then test them on customers — variability everywhere. | Identify underserved outcomes first; generate a handful of ideas aimed at them; evaluate on the Customer Scorecard before build. |
| Segmenting by demographics, role, or price point | Age and company size don't predict unmet outcomes — wastes the segmentation. | Segment by outcome patterns using cluster analysis on importance/satisfaction data. |
| Letting the sales force speak for customers | Salespeople capture inputs as "solutions" and "specs" — solutions in, solutions out. | Let trained researchers go to end users directly; let salespeople sell. |
| Treating "easy to use" as a requirement | One Motorola study found 21 distinct desired outcomes hiding behind "easy to use." | Pin every adjective to a measurable outcome statement. |
| Mixing "eliminate" / "prevent" / "minimise" in outcome statements | Word choice shifts the importance rating; introduces statistical noise. | Use only "minimise" or "increase" — nothing else. |
| Q-F-D / House of Quality for innovation | A manufacturing-reliability tool retrofitted to NPD; people fill out the matrix instead of doing the work. | Use outcome statements directly; the matrix becomes unnecessary. |
| Skipping Step 6 (Positioning) | The fastest revenue lift in ODI is re-messaging existing products against newly discovered underserved outcomes — Cordis 1% → 5% in 6 months. | Always run positioning before touching the pipeline. |
| Treating discovered opportunities as suggestions | Teams revert to favourite projects; the ranked list collects dust. | "Treat the discovered opportunities as sacred" — reward employees for satisfying ranked outcomes, not for ideas. |
Quick Diagnostic
Audit any product, roadmap, or research effort. Goal: every row Yes.
| Question | If No | Action |
|---|
| Can you state the customer's job to be done in one sentence with a functional verb? | You're describing a product, not a job. | Re-frame: "When [circumstance], the customer is trying to [verb + object]." |
Do you have 50+ outcome statements for that job, in min/increase + unit + outcome format? | You have opinions, not metrics. | Extract outcomes from 20–30 interviews; see references/jobs-outcomes-constraints.md. |
| Have customers rated every outcome on importance and satisfaction (1–5)? | You don't know what's underserved. | Quantitative survey (web/phone); ≥ 180 respondents per segment. |
| Have you ranked outcomes with the Opportunity Algorithm? | You'll target popular ideas instead of underserved ones. | Importance + max(Importance − Satisfaction, 0). |
| Did you segment by outcome patterns, not demographics? | Your "segments" are marketing personas, not innovation targets. | Cluster customers on the outcome-importance matrix. |
| Did you score concepts on the Customer Scorecard before building? | You'll discover failure after the spend. | Rate each concept's projected impact on every outcome; sum the weighted score. |
Reference Files
- jobs-outcomes-constraints.md: the three input types, job dissection, outcome statement format, the 1-hour outcome interview.
- opportunity-algorithm.md: the equation, the 10-point scale, the importance/satisfaction matrix, underserved-vs-overserved decisions.
- segmentation.md: outcome-based vs job-based segmentation, factor analysis + cluster analysis, why "segments of opportunity" beat personas.
- customer-scorecard.md: focused brainstorming, the five ideation guidelines, evaluating concepts on the full outcome set, the 5–10% vs 20%+ threshold for "breakthrough."
Further Reading
This skill is based on What Customers Want — Using Outcome-Driven Innovation to Create Breakthrough Products and Services by Anthony W. Ulwick (McGraw-Hill, 2005, ISBN 0-07-140867-3). See sources.md for the full extraction manifest. For the deeper methodology and Ulwick's later work:
- Ulwick, A. (2002). "Turn Customer Input into Innovation." Harvard Business Review, January 2002.
- Ulwick, A. (2016). Jobs to Be Done: Theory to Practice — the successor volume that formalises ODI under the JTBD banner.
- Strategyn (Ulwick's firm):
strategyn.com — case studies, the Outcome-Based Brand library, training in ODI.
- Christensen, C. & Raynor, M. (2003). The Innovator's Solution — Ch 3 references Ulwick's outcome-based segmentation directly.
About the Author
Anthony W. Ulwick is the founder and CEO of Strategyn, the consultancy that originated Outcome-Driven Innovation. He began developing ODI in 1984 after the failure of IBM's PCjr (his project at the time) convinced him that the customer-driven approach was structurally broken. Over the next two decades he refined the method across 100+ engagements with Microsoft, Bosch, Johnson & Johnson, Motorola, Pratt & Whitney, Cordis, and others. Harvard's Clayton Christensen called ODI "the most disciplined and predictable approach to innovation I have seen."