The ODI Process: 6 Steps to Systematic Innovation

Six Steps That Change How You Innovate

Most innovation processes are either too vague (“empathize, ideate, prototype, test”) or too rigid (“fill out Gate 2 form 14B”). Outcome-Driven Innovation sits in the space between — structured enough to be repeatable, specific enough to produce actionable output at every step.

The ODI process has six steps. They are sequential: each step produces deliverables that the next step requires. Skipping a step, or doing them out of order, breaks the logic chain and undermines the results. This is not a buffet where you pick your favorites — it is an engineering process where each stage builds on the last.

In this article, we walk through each step in detail. For each one, we cover what happens, what it produces, common pitfalls, and a concrete example. For the complete framework context, see our Outcome-Driven Innovation pillar guide.

Step 1: Define the Market Around the Job to Be Done

What Happens

You identify and precisely scope the core functional job that your target customers are trying to get done. This is the foundation of the entire project. Get the job wrong and everything downstream — the outcomes, the survey, the strategy — will be wrong too.

A job is not a task (too small), not a goal (too vague), and not an activity (too solution-dependent). It is a fundamental process that the customer initiates to achieve a specific objective. It follows a standard syntax: verb + object + contextual clarifier.

Examples:

• “Maintain the sterility of a surgical field during an operating procedure”
• “Secure heavy loads to a flatbed trailer for highway transport”
• “Monitor the structural integrity of a bridge over its operational lifetime”

What It Produces

• A precisely defined core functional job
• Related jobs (adjacent jobs the customer also needs to do)
• Emotional jobs (how the customer wants to feel while executing the job)
• Social jobs (how the customer wants to be perceived)
• A consumption chain job map (how the customer acquires, uses, and disposes of solutions)

Common Pitfalls

Defining the job too narrowly. “Drill a hole” is too narrow — it limits your solution space to drilling technology. “Create a hole of the desired diameter and depth in a surface” opens the space to include lasers, punching, water jets, and technologies that do not exist yet.

Defining the job too broadly. “Manage a construction site” is too broad — you will end up with 500 outcomes and no actionable focus. The right altitude captures a complete process that a single job executor can describe from beginning to end.

Confusing jobs with solutions. “Use a CRM system” is a solution, not a job. The job might be “Track and manage customer relationships throughout the sales cycle.” The test: if the statement mentions a product or technology, it is not a job.

Example

A manufacturer of surgical instruments initially defined their market as “surgical suturing.” After proper ODI job definition, the job became: “Repair damaged tissue to restore its structural integrity.” This broader framing opened the competitive frame to include surgical adhesives, stapling devices, and energy-based tissue welding — all of which competed for the same job. It also revealed that 30% of the outcomes their customers cared about had nothing to do with the suturing act itself, but with preparation and verification steps surrounding it.

Step 2: Discover Customer Needs Through Qualitative Research

What Happens

You conduct in-depth qualitative interviews with job executors — the people who actually do the job. The interviews are structured around the Job Map, a universal framework that breaks any job into its component process steps.

The standard Job Map has eight stages:

1. Define — Determine the objectives and plan the approach
2. Locate — Gather the items and information needed
3. Prepare — Set up the environment and resources
4. Confirm — Verify readiness before executing
5. Execute — Perform the core activity
6. Monitor — Track progress and results during execution
7. Modify — Make adjustments as needed
8. Conclude — Finish the job and verify the outcome

For each stage, the interviewer probes for desired outcomes: “What are you trying to achieve at this step? What makes this step difficult? How do you know you’ve done it well? What could go wrong?”

The interviewer’s discipline is critical. Customers naturally describe solutions (“I wish the handle were bigger”). The interviewer must translate these into outcome statements (“Minimize the likelihood that the instrument slips during use”). This translation is the hardest skill in the ODI process.

What It Produces

• A completed Job Map with 8-15 process steps
• 100 to 150 desired outcome statements, organized by Job Map stage
• A clear picture of the job’s scope and complexity

Common Pitfalls

Stopping too early. Most teams conduct 5-8 interviews and assume they have captured everything. ODI requires 15-30 interviews to reach saturation — the point where new interviews produce fewer than 2-3 new outcomes. In our experience, interviews 12-20 are where the most differentiated insights emerge, because these conversations tend to include edge-case users and unusual contexts.

Capturing solutions instead of outcomes. This is the most common error. “I need a lighter device” is a solution. “Minimize the physical effort required to operate the device for extended periods” is an outcome. Every captured statement should follow the format: Direction + Metric + Object of Control. See How to Write Outcome Statements for the complete guide.

Interviewing the wrong people. The job executor is the person who actually does the job — not the buyer, not the specifier, not the department head. In B2B contexts, this distinction matters enormously. The surgeon, not the hospital procurement officer, is the job executor for “repair damaged tissue.”

Example

In an industrial equipment project, we mapped the job of “lifting and positioning heavy loads at a construction site.” The Job Map revealed 12 process steps. The “Define” stage alone produced 14 outcome statements — things like “minimize the time it takes to determine the optimal lifting approach for the load’s weight distribution” and “minimize the likelihood that the planned lift path conflicts with other site activities.” By the end of 24 interviews, we had captured 137 desired outcomes. Only 40% related to the core “Execute” stage (actually lifting the load). The other 60% addressed planning, preparation, monitoring, and post-lift verification — stages that most competitors’ product development had ignored entirely.

The moment when a product manager sees the full Job Map for the first time is always eye-opening. They realize that 60% of what their customers care about happens before and after the core product usage. That’s where the unaddressed opportunities hide.

Step 3: Quantify Over- and Underserved Needs

What Happens

You take the 50–150 outcome statements from Step 2 and field a quantitative survey to a statistically representative sample of job executors. Each respondent rates each outcome on two dimensions:

1. Importance: “When [executing the job], how important is it to you to [outcome statement]?” (1-5 scale, where 1 = not at all important, 5 = extremely important)

2. Satisfaction: “When [executing the job], how satisfied are you with your ability to [outcome statement] using your current solution?” (1-5 scale, where 1 = not at all satisfied, 5 = extremely satisfied)

The sample size depends on the market. For consumer markets, 300-600 respondents are typical. For B2B markets, 180-300 respondents are sufficient. The key is statistical confidence — you need enough responses to detect meaningful differences between outcomes.

The data feeds into the Opportunity Algorithm:

Opportunity Score = Importance + max(Importance - Satisfaction, 0)

Scores range from 0 to 20 (on a rescaled basis). Outcomes with scores above 10 are underserved — representing real market opportunities. Scores above 15 indicate severe unmet needs.

What It Produces

• Importance and satisfaction scores for every outcome
• Opportunity scores ranking all outcomes from most to least underserved
• Statistical confidence intervals for each score
• Raw data for segmentation analysis in Step 4

Common Pitfalls

Using a convenience sample. Surveying your existing customers is easy but biased. They chose your product for a reason — they may have a systematically different needs profile than the broader market. Include competitor users and non-consumers for a complete picture.

Surveying too few outcomes. Some teams try to shorten the survey by pre-selecting “the most important” outcomes. This defeats the purpose. The whole point of quantification is to let the data tell you what matters — not your pre-existing assumptions.

Ignoring satisfaction variance. An average satisfaction score of 3.5 might mask a bimodal distribution — half the market is very satisfied, half is very dissatisfied. This is a segmentation signal, not a single-point estimate.

Example

In a medical device project, we surveyed 240 surgeons on 128 outcome statements related to “repairing soft tissue damage.” The results revealed that 22 outcomes had opportunity scores above 12 — indicating significant unmet needs. Strikingly, 15 of those 22 outcomes related to pre-procedural and post-procedural steps (tissue assessment, preparation, and verification), not the repair procedure itself. The client’s existing product roadmap had zero features addressing any of them. That single finding redirected an $18 million product development investment.

Step 4: Discover Hidden Growth Opportunities

What Happens

The quantified outcomes from Step 3 are plotted on the Opportunity Landscape — a scatter plot with importance on the y-axis and satisfaction on the x-axis. Every outcome becomes a data point. The resulting visual reveals the strategic structure of the market at a glance.

Three zones matter:

• Upper left (underserved): High importance, low satisfaction. These are the growth opportunities — customers care about these outcomes and current solutions fail to deliver. Clusters of underserved outcomes define “opportunity spaces.”

• Diagonal (appropriately served): Importance roughly equals satisfaction. These are table stakes. Your solution must satisfy them, but they will not differentiate you.

• Lower right (overserved): Low importance, high satisfaction. Current solutions exceed what customers need. This signals potential for a simpler, cheaper alternative — the classic disruptive strategy.

Beyond the Opportunity Landscape, this step includes outcome-based segmentation: using cluster analysis (k-means, latent class) to identify groups of customers who share similar patterns of unmet needs. These segments are far more actionable than demographic segments because they tell you what to build for each group.

What It Produces

• The Opportunity Landscape (visual + data)
• Identified opportunity spaces (clusters of underserved outcomes)
• Outcome-based segments with distinct opportunity profiles
• Segment sizing and attractiveness assessment

Common Pitfalls

Treating every underserved outcome equally. An outcome with a score of 11 and one with a score of 16 are both “underserved,” but they are not equal opportunities. Focus on clusters of related high-scoring outcomes — they define solvable problem spaces.

Ignoring the overserved zone. Companies pursuing premium strategies often dismiss overserved outcomes. But overserved outcomes in one segment paired with underserved outcomes in another segment can reveal a market bifurcation — an opportunity for both a premium and a value product.

Forcing demographic segments. The urge to cut the data by company size, geography, or industry is strong. Resist it. Outcome-based segments are more actionable because they are defined by what people need, not who they are. Two surgeons at the same hospital, doing the same procedure, can be in different outcome-based segments.

Example

For an agricultural equipment manufacturer, the Opportunity Landscape showed a clear bifurcation. One cluster of 35 outcomes related to precision and data integration was severely underserved (average opportunity score: 13.8). Another cluster of 28 outcomes related to basic operational reliability was overserved (average opportunity score: 6.2). Segmentation revealed that 40% of the market was in the “precision” segment (willing to pay more for data-driven capabilities) and 45% was in the “reliability” segment (overserved by current feature-rich products and price-sensitive). The remaining 15% was mixed. This data justified a dual-product strategy that the company had debated inconclusively for two years.

Step 5: Formulate a Growth Strategy

What Happens

Based on the Opportunity Landscape and segments, you select a growth strategy. ODI defines five strategic options, each suited to a different market pattern:

Differentiated strategy: Target a segment with a significant cluster of underserved outcomes and create a solution that addresses 10 or more of them. This is the most common ODI strategy — it produces premium products that customers will pay more for because they address needs that no current solution satisfies.

Dominant strategy: Address underserved outcomes across multiple segments simultaneously. This produces market-leading products but requires larger investment and broader capability.

Disruptive strategy: Target the overserved segment with a simpler, cheaper solution that strips out features serving overserved outcomes. This is Christensen’s disruption theory, quantified. Instead of guessing whether a market is “ready for disruption,” you have data showing exactly which outcomes are overserved and by how much.

Discrete strategy: Focus on a narrow cluster of 3-5 related underserved outcomes and solve them perfectly. This creates defensible niches — particularly valuable in fragmented B2B markets.

Sustaining strategy: Incrementally improve appropriately served outcomes. Low risk, moderate reward. Suitable for maintaining market position while developing more ambitious strategies.

What It Produces

• A documented growth strategy with rationale tied to opportunity data
• Target segment definition
• Specific outcome targets (the underserved outcomes your product will address)
• Competitive positioning based on which outcomes you will serve better than alternatives

Common Pitfalls

Trying to address every underserved outcome. A product that tries to do everything does nothing well. Select 10-15 underserved outcomes that form a coherent cluster — related by Job Map stage, by customer workflow, or by technical feasibility.

Ignoring feasibility. ODI tells you where the opportunity is. It does not tell you whether you can solve it. Strategy formulation must integrate opportunity data with your technical capabilities, manufacturing capacity, and go-to-market strengths.

Defaulting to sustaining. It is always safer to improve what you already have. But sustaining strategies rarely produce growth. If the data shows significant underserved clusters, a differentiated or disruptive strategy will produce better returns — even accounting for the higher risk.

Example

A manufacturer of material handling equipment had opportunity data showing 18 underserved outcomes related to load visibility and spatial awareness, and 12 overserved outcomes related to manual control precision. The differentiated strategy was clear: invest in sensor fusion and augmented visibility systems while simplifying the manual control interface. The company allocated 70% of its development budget to the underserved cluster and 30% to cost-reducing the overserved features. The resulting product achieved 140% of its first-year revenue target and commanded a 15% price premium over the previous generation.

Step 6: Generate Ideas and Create Concepts

What Happens

With specific underserved outcomes identified and a strategy selected, ideation becomes targeted rather than open-ended. Instead of “How might we improve the customer experience?” the team works from specific briefs:

• “Generate concepts that minimize the time it takes to verify load stability before initiating a lift”
• “Generate concepts that minimize the likelihood of tissue damage during suture placement”
• “Generate concepts that minimize the number of adjustments needed to achieve the desired cutting depth”

Each brief is a measurable outcome. Concepts can be evaluated against the outcome target — does this idea plausibly address the outcome? How well, compared to competing concepts? — before any prototyping begins.

Ideation sessions in ODI involve cross-functional teams (R&D, marketing, product management, manufacturing) and typically produce 50-100 concepts across the targeted outcomes. These are filtered, combined, and refined into 3-5 product concepts that address clusters of underserved outcomes.

What It Produces

• Concept descriptions tied to specific underserved outcomes
• Concept evaluation matrices (how well does each concept address each target outcome?)
• 3-5 refined product concepts ready for feasibility assessment and prototyping
• A clear rationale for each concept, traceable to customer data

Common Pitfalls

Reverting to pet ideas. The temptation to revive a shelved idea from two years ago is strong. If the idea does not address the identified underserved outcomes, it does not belong in this round — no matter how elegant the engineering is.

Generating concepts for every outcome. Focus on the outcomes your strategy targets. You do not need to solve everything in one product generation.

Skipping concept evaluation. Some teams generate ideas and then jump straight to prototyping the one that “feels right.” Evaluate every concept against the target outcomes first. The concept that best addresses the highest-scoring underserved outcomes is the strongest candidate, regardless of internal enthusiasm.

Example

For a consumer healthcare product, the team had identified 8 underserved outcomes related to application accuracy and dosage confidence. The ideation session produced 67 concepts. After scoring each concept against the 8 target outcomes, three concepts clearly dominated — each addressing 5-7 of the 8 outcomes. These three were prototyped and tested. The winning concept addressed all 8 outcomes and became the next-generation product, achieving a 4.2-star average customer rating compared to 3.1 stars for the predecessor.

The Complete Timeline

A full ODI project typically runs 12 to 16 weeks:

Frequently Asked Questions

What To Read Next

• Outcome-Driven Innovation: The Definitive Guide
• What Is ODI? Outcome-Driven Innovation Explained
• How to Write Outcome Statements That Drive Product Decisions
• The Opportunity Algorithm: Finding Underserved Customer Needs