The Uncomfortable Arithmetic of Customer-Centric Innovation
Every company claims to be customer-centric. Every strategy presentation includes a slide titled “The Customer at the Center.” Every roadmap meeting features the phrase “we need to get closer to the customer.”
And yet, depending on which study you cite, somewhere between 70 and 95 percent of new products fail in the market.
If everyone is already customer-centric, why do so many innovations fail?
The answer is uncomfortable: most companies confuse proximity to customers with understanding of customers. They talk to customers, run surveys, organize focus groups, create personas and journey maps. They are busy with customer research. But they are asking the wrong questions — and drawing the wrong conclusions from the answers they receive.
This article examines why the standard toolkit of customer research methods fails systematically at customer-centric innovation — and what actually works.
Method 1: Customer Surveys — the Illusion of Data
The Structural Problem
Customer surveys are the default instrument of market research. They feel objective, quantifiable, and scalable. But they contain a fundamental design flaw: they ask customers about solutions, not about needs.
“What features would you like to see?” generates a list of solution proposals based on what customers already know. A customer in 2005 would not have requested “a smartphone with an app store.” They would have said: “a phone with a better keyboard.”
The Henry Ford attribution — “If I had asked people what they wanted, they would have said faster horses” — is probably apocryphal. But the underlying logic is sound: customers can articulate their needs, but they cannot invent solutions they have not yet encountered. That is supposed to be the job of the product team.
The Evidence
A Bain & Company study found that 80 percent of companies believe they deliver a superior customer experience. Only 8 percent of customers agree. This gap does not arise from bad intentions. It arises from companies collecting the wrong data and then drawing strategic conclusions from it.
When surveys ask about solution preferences, they produce responses filtered through customers’ awareness of what is technically possible. This means surveys systematically miss needs that no current solution addresses — which is precisely where the highest-value innovation opportunities live.
What Works Instead
Instead of asking about desired features, ask about goals and success criteria: “What are you trying to accomplish when you [do the job]?” and “How do you know when that went well?” These questions surface desired outcomes — solution-independent, measurable statements of customer need. The Jobs to Be Done method provides the structured framework for capturing and analyzing these outcomes systematically.
Method 2: Focus Groups — the Social Distortion Machine
The Structural Problem
Focus groups have been a standard market research instrument since the 1940s. Eight to twelve people sit around a table, a moderator asks questions, observers behind a one-way mirror take notes. The setup feels rigorous. It is not.
Focus groups produce social dynamics that distort individual needs. Several mechanisms guarantee this:
- Dominance effect: One or two opinionated participants steer the discussion. Quieter participants adapt.
- Conformity pressure: In group settings, people tend to align with the emerging majority view — even when their actual experience differs.
- Performance pressure: Participants try to give intelligent or creative answers rather than honest accounts of their functional struggles.
- Moderator effect: The framing of questions shapes responses. A moderator with an implicit preference distorts the entire session without realizing it.
A Real Example
A manufacturer of professional cleaning equipment ran focus groups with facility managers. The dominant group view: “We need quieter machines.” Everyone nodded. The company invested in noise reduction.
A subsequent quantitative JTBD study told a different story. Noise was indeed a concern — but it was already a well-served need. Satisfaction scores were adequate. The genuinely underserved outcomes sat in planning cleaning routes efficiently and documenting compliance for clients. The focus group had made the loudest complaint literally the loudest topic in the room.
The company spent eighteen months and significant engineering budget on a problem that was not particularly important to its customers, while the actual growth opportunities went unaddressed.
Focus groups measure what customers say in a social situation. Not what they actually need. Those are two fundamentally different things. In over twenty years of innovation consulting, I have not once seen a focus group identify the genuinely critical, underserved needs in a market. They find the visible problems — the complaints customers are comfortable voicing publicly. The real opportunities are usually invisible in that setting.
Method 3: Personas — the Narrative Construction
The Structural Problem
Personas are fictional customer profiles based on qualitative research. “Thomas, 47, production manager at a mid-sized supplier, two children, tech-savvy, reads trade journals.” They feel useful. They give the abstract customer a face. Teams can ask: “Would Thomas need this?”
But personas carry three fundamental weaknesses that make them poor tools for customer-centric innovation:
First: they are demographic, not need-based. Two production managers, both 47, both with families, can have radically different needs — because they operate in different contexts or are trying to accomplish different sub-jobs with the same class of equipment.
Second: they are not prioritizable. A persona tells you nothing about which need is more important than another. “Thomas wants efficiency” — but which kind of efficiency? At which step of the job? How important is that compared to other outcomes?
Third: they are confirmation-biased by construction. Teams build personas that reflect their existing assumptions. The customer who likes the current solution gets prominent persona treatment. The frustrated user who has developed workarounds falls through the cracks.
What Works Instead
Instead of segmenting customers by demographic similarity, segment by patterns of unmet needs. A JTBD-based segmentation might show: there is a segment of production managers massively underserved on changeover speed, and another segment struggling with compliance documentation. These segments cut across demographic categories — and they are strategically far more actionable than any persona. The product features that address underserved outcomes for a specific segment also identify the marketing message that will resonate with it.
Method 4: Net Promoter Score — the Dangerous Simplification
The Structural Problem
The NPS (“How likely are you to recommend us?”) has become the primary customer KPI for many organizations. It is easy to collect, easy to communicate, and feels like a clear action signal.
The problem: NPS measures satisfaction with the current relationship — not the presence or absence of unmet needs.
An NPS of 70 says: “Our existing customers are satisfied.” It does not say: “There are no underserved needs in this market.” Kodak’s NPS was presumably high until digital photography made film cameras irrelevant.
More problematically, NPS tells you nothing directional about innovation. When it drops, you know something is wrong — but not what. When it rises, you know something is working — but not what to do next. For a product roadmap meeting, NPS is almost useless.
What Works Instead
Measure satisfaction and importance for each individual desired outcome. This gives you not a single number but a map: here we are strong, here we are weak, here is the largest opportunity. That map is what drives customer-centric product decisions. For a deeper treatment of innovation metrics, see the related article on Innovation Metrics and Measurement.
Method 5: Voice of the Customer — the Mistranslation Problem
The Structural Problem
VOC programs collect customer feedback from multiple sources: complaints, support tickets, sales reports, trade shows, social media. The idea is sound — aggregate all customer signals and analyze systematically.
The problem lies in the translation step. Customer voices are solution requests, complaints, and emotional expressions. Collecting them systematically does not transform them into needs. It creates a systematic collection of symptoms.
A customer who says “the software is slow” is describing a symptom. The underlying need could be: “Minimize the time between entering a search query and displaying relevant results.” Or it could be: “Minimize the likelihood that the software becomes unresponsive during a critical workflow step.” These two outcomes point to entirely different engineering priorities.
VOC programs collect “the software is slow” and count how often it appears. But complaint frequency is not a reliable proxy for strategic priority. Customers complain about what bothers them most visibly — not about the needs that no current solution addresses.
The Core Problem: Inputs Versus Outcomes
All five methods share a single structural flaw: they collect inputs and treat them as outcomes.
- Inputs are what customers say, request, complain about, and wish for.
- Outcomes are the measurable criteria customers use to judge success at each step of the job.
The confusion is understandable. When a surgeon says “I need a better endoscope,” it sounds like a clear statement of need. It is actually a solution request. The outcomes lie deeper: “Minimize the likelihood that anatomical structures in the surgical field are misidentified.” “Minimize the time required to adjust the optics to varying light conditions.”
The difference between collecting inputs and capturing outcomes is not an academic distinction. It is the difference between product development that is systematically guided by what customers are trying to achieve and product development that is systematically misguided by what customers think they want.
Info
What Actually Works: The Outcome-Driven Approach
The alternative to the methods described above is Outcome-Driven Innovation (ODI). ODI addresses the structural problems of traditional customer research through three mechanisms:
1. Systematic Need Identification
Instead of asking customers for opinions, wishes, or complaints, ODI decomposes the customer’s job-to-be-done into 100 to 150 desired outcomes. These are solution-independent, measurable, and stable — they describe what customers want to achieve at each step of the job, regardless of which technology or product might achieve it.
2. Quantitative Prioritization
Each outcome is rated by a representative sample of customers on two dimensions: importance and satisfaction with current solutions. The resulting opportunity score — calculated as Importance + (Importance − Satisfaction) — shows with statistical confidence which needs are underserved and how large the growth potential is. This turns the roadmap debate from an opinion contest into a data discussion.
3. Need-Based Segmentation
Instead of segmenting customers by demographics (which has limited strategic value), ODI segments by patterns of unmet needs. The result is customer groups defined by what they need — groups that can be specifically targeted with new offerings and positioned against in competitive messaging.
The decisive advantage of ODI over traditional methods is that we do not rely on what customers say. We measure what they need. That sounds like a small difference. In practice, it is fundamental. It is the difference between a product strategy built on evidence and one built on articulate advocacy.
Why Organizations Stay With Methods That Do Not Work
If the weaknesses are clear, why do most organizations not change their approach?
Reason 1: Confirmation Bias
Traditional methods deliver data that confirms existing beliefs. That is comfortable. JTBD data frequently shows that the actual picture is different from what was assumed. That is uncomfortable — and it creates organizational resistance before any methodology conversation even begins.
Reason 2: Methodological Inertia
“We’ve done it this way for twenty years” is not an argument, but it is a powerful brake. The market research department has established processes, supplier relationships, and budget lines for proven methods. Changing anything requires organizational will and executive sponsorship.
Reason 3: Apparent Speed
A customer survey with ten questions can be completed in a week. A full JTBD study takes three to six months. What organizations often fail to account for: the survey produces ten data points of questionable strategic relevance. The JTBD study produces 150 prioritized data points that can underpin a product strategy for three to five years.
Reason 4: Missing Competence
JTBD and ODI require specific capabilities: the ability to define jobs at the right level of abstraction, formulate outcomes correctly, design valid surveys, and interpret results strategically. These capabilities are still not widely available in most organizations, making reliance on familiar methods the path of least resistance.
The Pragmatic Path to Better Customer Research
Nobody expects you to discard every existing customer research method overnight. The pragmatic path:
Phase 1: Acknowledge the gap. Accept that your current methods tell you something about customer opinions — but not necessarily about customer needs. The NPS, the VOC program, the focus group results — they are inputs, not outcomes.
Phase 2: Run a pilot. Choose a strategically important market and conduct a JTBD/ODI study. Compare the results with what your existing methods would have indicated. This comparison is usually the moment of clarity. Teams frequently discover that the needs they thought were critical are already well-served, while the genuine growth opportunities were invisible in their current research.
Phase 3: Integrate. Build JTBD-based need identification into your innovation process. Keep what works — NPS for monitoring satisfaction trends, VOC for capturing frontline signals — but replace opinion-based methods with outcome-based ones for strategic product decisions.
Frequently Asked Questions
Audit Your Customer Research Methods
Book a complimentary discovery call to explore how these ideas apply to your organization.
