Outcome-Driven Innovation (ODI)

What it is

A quantitative research method, developed by Tony Ulwick and brought to a wide audience by his 2002 Harvard Business Review article "Turn Customer Input into Innovation", that measures which parts of a customer's job are badly served. It starts from the job the customer is trying to get done, breaks that job into steps, and writes the measures customers use to judge success at each step as desired outcome statements. A survey then asks a large sample of customers to rate every outcome on how important it is and how satisfied they are with current solutions. A simple formula turns those two ratings into an opportunity score, and a cluster analysis groups customers by which outcomes are underserved for them. Outcome-Driven Innovation is a registered trademark of Ulwick's consultancy, Strategyn, which sells it as a service; this entry teaches the method as Ulwick has published it.

ODI sits beside several entries in this knowledge base and answers a different question from each:

• JTBD Switch Interview Method finds, qualitatively, why people switched; ODI measures which outcomes are unmet and for whom.
• Jobs-to-be-Done Positioning Framework uses Ulwick's job statement to position a product; ODI supplies the measured evidence.
• Kano Model classifies features by their effect on satisfaction; ODI rates outcomes, which outlast features.
• MaxDiff Analysis ranks relative importance only; ODI adds satisfaction, so it shows the gap.
• Dan Olsen's Importance vs Satisfaction Framework (The Lean Product Playbook, 2015) is a lightweight variant of the scoring for product teams. ODI's outcome-based segmentation, and its use in targeting and positioning, make it a PMM method rather than a backlog tool.

When to use it

• Choosing which segment to target in STP, when firmographic segments all look similar and you need to know which buyers have the most unmet needs.
• Positioning keeps leading with features competitors match, and you need evidence of which outcomes customers rate as important and poorly satisfied.
• Entering a crowded market, where finding an underserved segment is the only credible way in.
• Planning a low-cost or simplified offer, and you need proof that a segment is overserved and would accept less for a lower price.
• Product and PMM disagree about priorities, and a quantified customer read would settle it faster than more debate.

How to run it

1. Define the job executor and the core job. Name who is trying to get the job done and write the job as a verb plus an object, free of any solution: "agree and sign a commercial contract", not "use contract software".
2. Build a job map. Break the job into its steps, using Ulwick's universal sequence as a checklist: define, locate, prepare, confirm, execute, monitor, modify and conclude. Interview 15 to 30 executors to fill it in.
3. Write desired outcome statements. Each has four parts: a direction (minimise or increase), a metric (time, likelihood, number), an object (what is being controlled) and a clarifier (the context). Example: "Minimise the time it takes to find the approved version of a clause during a negotiation." Strategyn reports that a market typically yields 50 to 150 statements. Remove any statement that names a feature or a solution.
4. Survey a large sample. Ask each respondent to rate every outcome for importance and for current satisfaction on a five-point scale. Strategyn cites samples of 180 to 3,000 respondents; plan for at least 100 per segment you expect to find, or the segmentation step will not hold.
5. Calculate opportunity scores. Convert each rating to a 0 to 10 score (the share of respondents giving the top two points, divided by ten). Then apply Ulwick's formula: opportunity = importance + max(importance minus satisfaction, 0). The floor at zero stops a well-satisfied outcome from scoring below its importance. Scores run from 0 to 20.
6. Read the scores. Published sources do not fix one set of cut-offs. Practitioner summaries of Ulwick's work commonly treat scores above 10 as worth pursuing, above 12 as strong and above 15 as extreme. An outcome is overserved where satisfaction is higher than importance.
7. Segment by unmet outcomes. Cluster respondents on their opportunity scores, not demographics, to find groups with distinct sets of underserved outcomes. Profile each segment by size, firmographics and current solution so sales can find it.
8. Turn findings into decisions. PMM uses the result three ways: lead positioning and messaging with the target segment's top underserved outcomes; choose which outcome-based segment to target; and flag overserved segments for a simpler, cheaper offer, the opening a low-cost or disruptive entrant would use.

Cadence & ownership

At scale, product or strategy usually owns the innovation and roadmap decisions ODI informs, while PMM or a customer insights team commonly runs the study, owns the segmentation output, and owns the positioning that follows. A founding or solo PMM may run it end to end, often with a survey vendor for fieldwork and a freelance analyst for the clustering. A full study takes 8 to 12 weeks. Run one major study every one to three years per market, and refresh it after a major market shift: a new category entrant, a pricing model change, or a move into a new segment.

Maturity stages

• Lightweight: Olsen-style importance and satisfaction scores for 20 to 30 outcomes. Useful for prioritising; too small to segment.
• Full study: a complete job map, 50 or more outcome statements, a sample large enough to cluster, and named outcome-based segments.
• Embedded: segments carried into CRM and ICP scoring, and the study refreshed on a set cycle.