Library
Jobs to be Done: Theory to Practice · 11 of 12
Jobs to be Done: Theory to Practice
Entrepreneurship HIGH

Outcome-Based Segmentation

Jobs to be Done: Theory to Practice Tony Ulwick
segmentation unmet-outcomes ODI strategy-selection value-proposition

Outcome-Based Segmentation

Key Principle

Segment customers by shared patterns of unmet desired outcomes, not by demographics, psychographics, or firmographics. Subsets of customers encounter different situational complexity in the same job-to-be-done, producing distinct clusters of unmet needs that demographic slicing cannot detect. Factor and cluster analysis on quantitative outcome ratings reveals these hidden groups and dictates which growth strategy is viable for each.

Why This Matters

  • Demographics are noise. Hundreds of segmentation studies across dozens of industries confirm that age, geography, and psychographic profiles fail to explain need variation (p. 96). Two people in the same demographic can face entirely different job complexity.
  • Averaging kills the signal. Bosch surveyed ~270 tradesmen on circular-saw outcomes. The aggregate data showed zero unmet needs. Only after outcome-based clustering did a segment of finish/advanced carpenters (>30% of users) emerge with 14 of ~85 outcomes unmet (p. 146).
  • Strategy selection depends on segment composition. The ratio of underserved to overserved segments determines whether a differentiated, disruptive, or dominant strategy is viable. Without this data, strategy is guesswork (p. 99).
  • Value propositions decay. Competitors converge on messaging dimensions that were once underserved but are now satisfied. Companies that do not re-measure segmentation continue "skating to where the puck had been" (p. 100).

Good Examples

Bosch CS20 Circular Saw (pp. 97, 146-149): Finish carpenters needed frequent blade-angle and blade-height adjustments -- added complexity invisible to demographics. ODI segmentation surfaced 14 unmet outcomes in that segment. The team designed features like the DirectConnect cord system (which simultaneously reduced cord-cutting risk, prevented plug snagging, and cut repair downtime while lowering product cost) and a table cutout for line-of-sight alignment. Satisfaction in unmet-need areas rose from 63% to 87% -- roughly 4x the typical new-product improvement of <10%. The CS20 became a top-selling, top-rated circular saw in North America.

Coloplast Wound Care (pp. 99-100): Every competitor messaged "We help wounds heal faster." Outcome-Based Segmentation revealed that for a key underserved nurse segment, 10 of the top 15 unmet outcomes had nothing to do with healing speed -- they related to preventing patients from worsening their wounds. Coloplast repositioned around complication prevention and achieved double-digit growth in under six months.

Abbott Medical Optics (pp. 149-150): Stuck in "me too" service delivery, AMO applied ODI to the job of replenishing ophthalmic lenses for cataract surgery. Job mapping captured ~100 desired outcomes and revealed that the traditional front-office/back-office split in materials management was an artificial boundary. Dissolving it opened new service territory.

Counterpoints

  • The method requires quantitative surveys of 180-3,000 customers per study (p. 98), making it expensive and slow for very early-stage ventures that lack an established customer base.
  • Factor and cluster analysis can produce different segment solutions depending on analyst choices (number of clusters, rotation method). The segments are data-driven but not uniquely determined.
  • In markets with very homogeneous jobs (minimal situational complexity), outcome-based segmentation may converge on a single segment, adding cost without strategic insight.

Key Quotes

"If you do not know what underserved and overserved segments and desired outcomes exist, you will not know which growth strategy to pursue. You will be guessing at innovation and competing on luck." (pp. 98-99)

"10 of their top 15 unmet desired outcomes related to 'making sure the wound doesn't get worse.'" (p. 100)

"Satisfying all of these outcomes at the same time is what made this a true innovation." (p. 147)

Rules of Thumb

  1. Never aggregate before segmenting. Market-level averages cancel out the unmet-need signal. Always run cluster analysis on outcome ratings first (p. 146).
  2. Three segment types, three strategy implications. Underserved segments support differentiated/dominant strategies; overserved segments support disruptive strategies; appropriately-served segments offer no growth lever (p. 99).
  3. Opportunity score thresholds. Importance + MAX(Importance - Satisfaction, 0). Scores >10 = solid opportunity, >12 = high, >15 = extreme (p. 148).
  4. Re-segment periodically. Value propositions decay as competitors invest in once-underserved outcomes. What differentiated last year may be table stakes now (p. 100).
  5. Profile the complexity driver. After clusters emerge, use profiling questions to identify the situational factors causing the complexity -- this makes segments actionable for targeting (p. 98).
  6. Design for outcome clusters. A feature that satisfies multiple unmet outcomes simultaneously is the hallmark of true innovation; single-outcome features are incremental (p. 147).

Process Steps (Quick Reference)

  1. Capture desired outcome statements for the JTBD using standardized syntax (p. 98).
  2. Run a quantitative survey (180-3,000 customers) measuring importance and satisfaction per outcome (p. 98).
  3. Apply factor and cluster analysis to let segments emerge from the outcome data (p. 98).
  4. Use profiling questions to identify what situational factors cause the complexity driving each segment's unmet needs (p. 98).

Related References