Innovation is one of the most-used, least-understood concepts in business. Executives call it a priority. Companies hire Chief Innovation Officers. Books, conferences, and consultancies sell frameworks for generating it. And still, the majority of new products fail. The most sophisticated market research, the biggest R&D budgets, the most rigorous product development processes — all of it still produces results that disappoint customers more often than it delights them.
Why Innovation Keeps Failing
Christensen’s diagnosis is precise, and a little disturbing: companies are almost universally working from the wrong causal model. They correlate demographics, purchasing patterns, and stated preferences with product choices, then use those correlations to predict what customers will want next. Sophisticated. Expensive. Fundamentally misleading.
The problem is correlation without causation. That a 35-to-50-year-old man with household income above $75,000 bought a specific product tells you almost nothing about why he bought it. The why — the specific functional, social, and emotional job the product was hired to do — is what determines what else he’ll buy, what would make him switch, and which innovations will actually matter to him. Without the causal model, you’re managing the symptoms of customer behavior instead of its sources.
Jobs to Be Done (JTBD) theory is Christensen’s proposed alternative. Core idea: customers “hire” products and services to make progress toward a specific goal in a specific life situation. The “job” is that progress — the functional, social, and emotional outcomes they’re actually after. When a product understands the job and serves it well, customers hire it reliably. When a product serves the job poorly, or doesn’t even understand what job it’s being hired for, innovation misses the mark. Consistently.
“Customers don’t buy products. They hire them to make progress in their lives.”
The Framework: Jobs to Be Done
The JTBD framework rests on several interconnected ideas, each worth pulling apart on its own, because each one changes how you see customer decisions:
- The job is the unit of analysis, not the customer. “Who is our customer?” is a less useful question than “What job is our customer trying to get done?” The same customer hires different products for different jobs. A 40-year-old professional hires one product for “impress a client at dinner” and a different one for “feed my kids on a Tuesday night.” Understanding the job, rather than profiling the customer, produces actionable insight customer segmentation can’t.
- Jobs have functional, social, and emotional dimensions. Most innovation programs focus on functional performance — features, specs, price. But customers often hire products primarily for social or emotional reasons. A gym membership may be hired not for physical fitness (functional) but for self-identity maintenance (“I’m the kind of person who has a gym membership”) or social belonging. Products optimizing purely for functional performance while ignoring the other two dimensions consistently underperform.
- The competition for a job is not always obvious. What else is being hired for the same job? Counterintuitive answers show up here. A hotel isn’t competing with other hotels for “give me a quiet place to do uninterrupted work” — it’s competing with the home office, the coffee shop, the airplane seat. An innovation team that defines competition only inside their own product category is designing for the wrong competitive context.
- The circumstances matter as much as the customer. “A 35-year-old male” isn’t a useful description of a customer situation. “A 35-year-old male in an airport with two hours before a flight and anxiety about tomorrow’s presentation” is the beginning of a useful one. The circumstance determines which job is active, and therefore which product would get hired. Same person, different circumstance, completely different need.
The Milkshake Study
The most famous example in the book is the milkshake study, which Christensen has been teaching for decades and which remains the clearest illustration of JTBD in action.
A fast food chain hired a consultant to help improve milkshake sales. Conventional approach: customer surveys on flavor preferences, focus groups on thickness and size, demographic profiles of milkshake buyers. The chain tried all of it. Barely moved the needle.
The JTBD approach went differently — spend a day in the restaurant and watch what job customers are actually hiring the milkshake for. What came out of it changed the whole analysis. Roughly half of milkshake purchases happened before 8 a.m. The buyers were mostly commuters, buying alone, rarely buying anything else. Asked why they picked a milkshake, their answers were telling: it kept them occupied during a long commute; it was filling enough to last until lunch; it fit in the cup holder; it was easier to handle than a bagel while driving.
The job was never “enjoy a delicious treat.” The job was “make my commute less boring and hold off hunger until lunch, without mess or the distraction of eating real food behind the wheel.” The milkshake’s real competition wasn’t other milkshakes. It was bananas (too quick), granola bars (crumbs), bagels (awkward one-handed), Snickers bars (too much guilt).
The innovation implications turned out totally different from what the original customer research suggested. Flavor optimization was almost irrelevant. Making the milkshake thicker, so it lasted longer through the commute, and adding small fruit chunks for variety — those were the high-value improvements. Speeding up the purchase process mattered a lot too, since commuters are time-pressured. None of this showed up in the traditional research, because the traditional research was asking about the product instead of the job.
The “Nonconsumer” Insight
One of the more valuable extensions of the JTBD framework is Christensen’s analysis of nonconsumers — people not buying any solution for a given job, not because they lack the job, but because nothing available is good enough to hire.
Nonconsumers are often the most important innovation opportunity in a market, and they get systematically overlooked by companies whose research focuses on existing customers. Makes sense why existing customers are the default research population — they’re accessible, they’ve got purchase history, they can give feedback on products they’ve already tried. But that’s exactly the limit: they can only describe variations on solutions they’ve already used. Nonconsumers describe the jobs nobody’s served well enough to earn their business — which is often where the biggest unmet need is hiding.
The disruptive innovation theory Christensen laid out in The Innovator’s Dilemma often starts right here: an innovation “good enough” to pull in nonconsumers at a price they’ll pay, even while it’s less capable than the existing products on the metrics everyone’s used to tracking. Established companies dismiss it, because their existing customers don’t want it. Meanwhile the newcomer builds a base among nonconsumers and keeps improving until it’s competing for the mainstream customers too. By the time incumbents notice the threat, the challenger’s already won the new segment.
Jobs to Be Done Applied to Strategy
JTBD has strategic implications beyond product development. Christensen argues a company’s purpose should be defined by the job it serves, not the product category it happens to sit in. “We make hamburgers” is a category definition. “We help people eat on the go, quickly and cheaply” is a job definition. The job definition produces better long-term strategy because it keeps attention on the actual value being created rather than whatever form that value currently takes.
Companies that define themselves by product category stay vulnerable to innovations that serve the same job in a different form. Companies that define themselves by the job they serve can respond to those innovations — or get there first — because they’re tracking the right variable.
Netflix versus Blockbuster is the canonical example. Blockbuster defined itself as a video rental company. Netflix initially defined itself as a DVD-by-mail service. But Netflix’s underlying orientation was toward the job — “help people watch what they want, easily and affordably” — which is exactly why the company could pivot from DVD-by-mail to streaming without an identity crisis. Blockbuster couldn’t pivot, because its entire operation was built around the product category, physical video rental, rather than the job underneath it.
The “Hiring” and “Firing” Metaphor
Christensen’s most useful rhetorical move is the hiring/firing metaphor for customer decisions. Customers don’t just buy — they hire. And when a product fails to do the job, they fire it and hire something else.
The firing decision is especially useful for understanding customer behavior conventional analysis tends to miss. When a customer stops buying your product, the conventional response asks “what did we do wrong?” Often the wrong question. The right one: “what job were they trying to get done, and what did they hire instead — and why did that alternative do the job better?”
The answer often reveals the customer never defected to a direct competitor at all. They hired a workaround, a totally different category of product, or a change in behavior that eliminated the job altogether. Understanding these “firing” decisions produces more actionable insight than understanding why customers pick you over a direct competitor.
“The causal mechanism behind a purchase—what causes a customer to buy a particular product or service—is that they have a job to be done.”
Where Competing Against Luck Falls Short
Two honest criticisms.
The JTBD framework, powerful as it is, isn’t as operationally specific as it needs to be. Christensen shows convincingly why the framework matters, with compelling examples. He’s a lot less precise about how to actually identify the jobs your customers are hiring for, how to validate a job hypothesis, or how to structure an organization to consistently develop this kind of insight. The book diagnoses better than it prescribes. Practitioners who actually want to implement JTBD need supplementary material — the work of Bob Moesta and Chris Spiek, who developed the “switch interview” methodology, in particular.
The examples skew heavily toward consumer products and large-company innovation. The application to B2B products, services businesses, smaller organizations is theoretically sound but needs adaptation the book never provides. JTBD does work for B2B — the jobs just tend to have more layers (individual user, organizational buyer, organizational goals), and the competitive analysis gets more complex.
For the broader context of how understanding your customer connects to organizational effectiveness, see our exploration of critical thinking skills — especially the section on causal versus correlational reasoning. The innovation failure mode Christensen describes has interesting parallels with the confirmation bias discussion in our piece on cognitive biases and decision making. And the strategic implications of JTBD connect to our deeper look at authentic leadership and understanding what you’re actually building.
The “Switch” Methodology: How to Actually Identify Jobs
The book describes the Jobs to Be Done framework with real clarity but stays vague on the methodology for actually identifying the jobs your customers hire for. The most effective methodology — developed by Bob Moesta and Chris Spiek, who worked closely with Christensen — is the “switch interview,” and understanding it is what makes the framework operational instead of merely aspirational.
The switch interview focuses on the moment of purchase, or the moment of switching away from a previous product, not general product satisfaction. The key questions look backward, situationally: What was happening in your life when you started looking for this kind of product? Walk me through the day you finally decided to buy. What were you hoping this product would do for you? What else did you consider? What almost stopped you from buying?
The method works because the switch moment is when the job is most visible — the customer made a specific decision in a specific context, to hire a specific product for specific progress, and the circumstances and tensions driving that decision can be reconstructed with surprising precision through careful interviewing. The job isn’t something a customer can just tell you (“I hired this because of job X”) — it has to be inferred from the story of the decision itself, including the forces pushing them toward the switch (what was dissatisfying about the old situation?), the forces pulling them toward the new product (what were they hoping it would do?), and the forces creating anxiety or resistance (what almost stopped them?).
A single switch interview takes about an hour and produces more actionable insight into why customers buy than most conventional market research produces in weeks of survey analysis. The reason is methodological. Surveys measure stated preferences, which are poor predictors of actual decisions. The switch interview reconstructs an actual decision in its actual context — which is where the real causal information lives.
The Innovation Failure Pattern: A Diagnostic
Christensen’s analysis of innovation failure follows a specific pattern that, once you’ve seen it, shows up constantly across industries and company types. Recognizing the pattern isn’t just intellectually satisfying — it’s diagnostic, helping identify which type of innovation failure you’re actually dealing with, and therefore which kind of intervention might address it.
Pattern one: feature-matching failure. The company adds features existing customers said they wanted, only to find the features don’t produce the predicted bump in adoption or satisfaction. Cause: the company listened to customers describe features they liked instead of understanding the jobs they were trying to get done. Features that sound good in a survey or focus group don’t always map to jobs that actually drive purchase decisions. The fix: shift the unit of analysis from “what features do customers want?” to “what jobs are customers trying to accomplish, and do our features serve those jobs better than the alternatives?”
Pattern two: adjacent market expansion failure. The company takes a product that succeeded in its original market into an adjacent one, and the adjacent market’s customers just don’t respond. Cause: the job in the adjacent market is different from the job in the original one, even when the product categories look similar. The milkshake that succeeds with commuters fails with parents bringing kids to the playground, because the job is entirely different. The fix: run job analysis in the new market independently, instead of assuming the job matches the one the product already serves.
Pattern three: disruption blindness. Leadership dismisses a new entrant’s product as inferior, relevant only to some segment they don’t compete for — and then watches the entrant improve and capture the mainstream market. Cause: leadership is evaluating the new product on the incumbent’s own metrics, which the new product performs poorly on by design, instead of on the jobs it’s actually being hired for by its initial customers. The fix: explicitly analyze what job the new entrant’s product is being hired for by its current users, then ask whether that job is evolving toward a direction that will eventually compete for your own customers’ jobs too.
JTBD and the Service Industry: Overlooked Applications
The book’s examples lean heavily toward physical products and consumer technology. Applying Jobs to Be Done to service industries — healthcare, education, financial services, professional services — is theoretically sound but gets almost no direct treatment. Worth addressing, because the framework is arguably even more valuable in service contexts, where the gap between what companies think they’re providing and what customers are actually hiring them for tends to run wider.
Take healthcare. Most providers think they’re in the business of treating disease. But patients hire healthcare providers for a much broader set of jobs: reassurance that a concerning symptom isn’t serious, an explanation of a diagnosis in terms they can actually understand, help navigating a complex web of specialists and treatments, coordination of care across providers who don’t talk to each other, and — yes — treatment of specific conditions. “Treat my disease effectively” is one component of what patients are hiring for. Not always the most important one. Providers who design around disease treatment while neglecting the reassurance, explanation, and navigation jobs end up producing technically competent care and chronically dissatisfied patients.
Same analysis applies to financial services (what job is a client hiring their financial advisor for — investment returns, or mostly “give me confidence that I’m going to be okay”?), to education (are students hiring the institution for knowledge, or for credentials, networks, and the experience of adult formation?), and to professional services (are clients hiring the consulting firm for analysis, or for credible external cover to make a decision they’d already made?). In every case, understanding the actual job — instead of the functional service category — produces better design, better communication, better outcomes.
Causation vs. Correlation: The Intellectual Foundation
The deepest contribution of Competing Against Luck is methodological, not strategic. It makes an explicit case for causal analysis over correlational analysis in understanding customer behavior, and gives a compelling account of why correlational analysis has stuck around despite consistently failing to predict how customers respond to innovation.
Correlational analysis persists because it’s cheap, scalable, and produces confident-sounding numbers. Survey 1,000 customers on their product preferences and you get a dataset that slices by demographic, cross-tabulates with purchase behavior, and shows up in front of leadership as evidence of what customers want. That this evidence consistently fails to predict how customers actually respond to new products gets blamed on implementation failures, market timing, competitive dynamics — anything but the fundamental inadequacy of correlation as a causal model.
Causal analysis — understanding specifically why a customer made the decision they made, in the context they made it — is harder, slower, and produces less confident-sounding numbers. A dozen carefully conducted switch interviews yields rich, specific, actionable insight into the causal mechanisms driving decisions. It just can’t be crammed into a bar chart or presented as “87% of customers prefer X.” It requires interpretation, synthesis, and the discomfort of qualitative complexity over quantitative simplicity.
Organizations that consistently innovate well — introducing products customers genuinely adopt rather than try once and abandon — are almost always the ones that found a way to do the harder causal analysis. They understand their customers’ jobs. They’ve built the research capability and internal culture that let job-level insight actually drive product decisions. Expensive, relative to survey research. Cheap, relative to the cost of the innovation failures it prevents.
The Jobs Lens on Your Own Career
The Jobs to Be Done framework applies to personal career decisions with the same logic it applies to products, and the application is illuminating in ways conventional career advice tends to miss. Standard career advice asks: what are your skills? What are you good at? Where do you want to be in five years? The JTBD reframe asks: what jobs are employers hiring for, and do your capabilities serve those jobs in a way that’s differentiated from other candidates?
Different questions, different strategies. The first produces a capabilities inventory oriented around what you’ve already built. The second produces an analysis of where the market has unserved or underserved jobs your specific mix of capabilities is well positioned to fill. The JTBD approach surfaces opportunities the capabilities-first approach misses, because it starts from the demand side instead of the supply side.
Practically: identify three to five specific roles or organizations you’re targeting. For each, run the equivalent of a switch interview — talk to people already in those roles about what their actual day-to-day work looks like, what problems they find hardest, what capabilities they feel their team is missing, what they wish their organization did better. Then check the jobs you’re hearing about against your own capabilities. Where’s the genuine mismatch between what the market needs and what the standard candidate profile offers? That gap is where your differentiated value proposition lives.
More work than polishing a resume and applying broadly. Also considerably more likely to produce the kind of targeted positioning that actually gets you hired for the job that uses your real capabilities fully — instead of generic positioning that draws equally generic responses from the other side of the hiring table.
What Makes Innovation Teams Effective: The Christensen Prescription
Christensen’s research into why companies fail at innovation led him to a structural insight that complements JTBD: innovation fails not just because companies don’t understand customer jobs, but because the organizational structures and incentives that make a company good at running its current business make it systematically bad at identifying and pursuing new ones.
The core structural tension: the resources, processes, and values — RPV — that enable excellent execution of the current business model actively resist any innovation that would undermine it. Resources (talented people, distribution channels, customer relationships) are usually more valuable inside the current business than inside an adjacent opportunity. Processes (measurement systems, planning cycles, approval chains) are built to make the current business more efficient, and they systematically slow down exploration of anything different. Values — what gets funded, what gets rewarded, what counts as success — reflect the priorities of the current business, and quietly filter out opportunities that don’t fit them.
The prescription: organizations that want to sustain innovation capacity need real structural separation between their current business and their innovation efforts — not just physical separation, though that helps, but resource, process, and values separation too. An innovation team evaluated against the current business’s metrics will stop innovating. An innovation team staffed entirely from current-business senior leaders will bring the current business’s assumptions along with them. An innovation team that has to fight the current business’s resource allocation process for every dollar loses. Every time.
Expensive, organizationally difficult. It takes leadership commitment to shield the innovation effort from the current business’s immune system, which will attack any real innovation as a threat to existing positions and priorities. The organizations that sustain innovation over time — that keep creating new businesses instead of just polishing the old ones — are the ones that found structural solutions to this, not just cultural slogans about “being more innovative.” Culture without structure produces innovation theater. Actual structural separation is what makes the difference.
The Metrics Problem: When Measurement Kills the Job
A consistent pattern in failed innovation efforts is what Christensen calls “optimization against the wrong metrics” — measuring innovation success against the established business’s performance metrics instead of the metrics that actually matter for the job the innovation is being hired for. Sounds obvious. It’s structurally baked into how most organizations fund and evaluate innovation anyway.
A new product evaluated on gross margin percentage gets killed even when it’s perfectly designed for its job, if that job happens to serve a lower-margin market than the current business. A new service evaluated on average revenue per customer gets killed even while it’s growing a new segment that doesn’t compete with existing ones, if the average revenue there is lower than the current business’s average. A new distribution channel evaluated on cannibalization risk gets killed even when it reaches a completely different customer segment, because cannibalization metrics don’t distinguish customers who would’ve bought through the old channel from customers who weren’t buying anything at all.
The structural fix is establishing innovation-specific metrics that measure what actually determines whether an innovation is creating new value: job completion (are customers accomplishing what they set out to?), nonconsumer adoption (are we reaching people who weren’t buying before?), switching behavior (are customers switching to us from genuine alternatives, not from nothing?), and repeat engagement (are they finding enough value to come back?). Different story than margin and revenue per customer. And they need to get assessed against the innovation’s specific opportunity, not the current business’s performance.
The Verdict on Competing Against Luck
One of the more important business books of the last decade. The Jobs to Be Done framework resolves a problem that’s frustrated innovation management since Peter Drucker first described it: why do sophisticated companies with abundant resources consistently fail to innovate successfully? Because they’re using the wrong causal model. Managing correlations instead of causes. Asking “who” and “what” when the real question was “why.”
The framework isn’t complete — Christensen himself acknowledges JTBD is a lens, not a recipe, and a full innovation management system needs additional pieces (organizational structure, development process, market testing) the book doesn’t provide. But the lens itself is transformative. Once you’ve understood what job a product is being hired for, it’s hard to go back to the product-centric view that produced the wrong insights before.
Read it if you’re responsible for product development, marketing, or strategy. Read it especially if you’ve felt the frustration of rigorous customer research that still produced products customers didn’t want. The book explains exactly why that happens and offers a better starting point. For the personal resilience dimension of building businesses around genuine customer understanding, see our guide to grit and resilience — which explores what sustains builders through the extended uncertainty of customer discovery. The decision-making improvements JTBD enables connect to our piece on decision making under pressure and our exploration of how mental toughness supports the kind of rigorous, honest thinking the framework demands.
Common Questions About Competing Against Luck
What is Jobs to Be Done theory?
A framework for understanding why customers buy products and services. Core idea: customers hire products to make progress toward a specific goal in a specific circumstance. The “job” is that progress — functional, social, emotional. Understanding what job a customer is hiring your product for produces better innovation decisions than demographic or psychographic profiling.
How does Jobs to Be Done differ from traditional market research?
Traditional market research correlates customer characteristics — demographics, purchase history, stated preferences — with buying behavior. JTBD looks for the causal mechanism: the specific job that triggers a purchase and determines what would make a product better or worse at doing it. Correlational research gives you accurate descriptions of who buys. JTBD research gives you actionable insight into why.
What is the milkshake example about?
A fast food chain found that half its milkshakes were bought by morning commuters hiring the milkshake to occupy them during a long drive and hold off hunger until lunch — not for enjoyment as a treat. The innovation implications (thicker milkshake, chunks for variety, faster purchase process) were entirely different from what conventional flavor-focused research suggested.
Who is competing against luck?
Any company that innovates based on correlation rather than causation — essentially, any company building products off demographic profiles, purchase history, and stated preferences without understanding the causal mechanism (the job) actually driving purchase decisions. Innovation based on correlation is competing against luck, because it has no reliable causal model to build from.
How does JTBD apply to B2B companies?
It applies, with more complexity. B2B jobs often have multiple layers: the individual user’s job, the organizational buyer’s job, the organization’s strategic job. These can pull against each other. JTBD analysis in B2B needs to identify all three layers and understand how they interact, rather than focusing only on the individual user level.
What are “nonconsumers” and why do they matter?
People not buying any solution for a specific job — not because they lack the job, but because nothing available is good enough. Nonconsumers represent unserved demand, often the most significant innovation opportunity in a market, and they’re systematically overlooked by companies whose research focuses on existing customers.
How does Competing Against Luck relate to The Innovator’s Dilemma?
The Innovator’s Dilemma described the pattern of disruptive innovation — how smaller, simpler products unseat established leaders. Competing Against Luck supplies the causal mechanism: disruptive innovations succeed because they serve a job the established product was never designed for, typically the job of nonconsumers. JTBD is the underlying theory explaining why disruption follows the pattern Christensen described in his earlier work.
How do I identify the jobs my customers are hiring my product for?
The book is vaguer about this than it should be. The most effective methodologies to come out of Christensen’s work include the “switch interview” (interview customers at the moment they switched to or away from your product, focused on the circumstances and tensions that drove the decision) and systematic observation of actual usage in natural contexts, rather than focus groups or surveys.
What is the three-dimensional job concept?
Jobs have functional, social, and emotional dimensions. Functional: what the product physically helps you accomplish. Social: how it affects how others perceive you. Emotional: how it makes you feel. Products that optimize only for functional performance while ignoring the social and emotional dimensions consistently miss part of why customers hire them.
Can JTBD be applied to personal career decisions?
Yes, and it’s illuminating. Ask: what job are employers hiring me for? Not “what are my skills?” but “what progress do they need to make in their specific situation, and how does hiring me help them make it?” That reframe produces more effective job search strategies, better interview performance, and a clearer sense of what you need to develop to be hireable for the jobs you actually want.
Related: Team of Rivals Summary
References
Editorial StandardsCorrectionsMedical DisclaimerAbout Our ContentAffiliate DisclosureSite Map