The Context Makes the Irrational Rational: Bob Moesta's Guide to Jobs to Be Done

Bob Moesta describes himself as an electrical engineer who has spent almost 30 years building products, and he says the lie he was raised on was "build it and they will come." The habitual approach is technological: build the thing, then ask who wants it. Jobs to Be Done starts from a different premise — that people hire products rather than buy them, and they hire them to make progress in their life. Once you can see the context someone is in and the outcome they want, the same market looks completely different.

Snickers and Milky Way are the illustrative pair: both candy bars, both bought in the checkout aisle, made from nearly the same ingredients except for peanuts. A product benchmark leaves you comparing softness, hardness, and calories. But ask people about the last time they ate each one, and Snickers turns out to be a missed meal, too much work to do, and a need to get energy back fast enough to return to the task. Milky Way is eaten alone after an emotional experience, positive or negative, as a way of regrouping.

That difference changes the competitive set entirely: a Snickers competes with a protein drink, a Red Bull, or a coffee, while a Milky Way competes with a glass of wine, a brownie, and even a run. Moesta says the framework exposes the true competitive set from what he calls the demand side of the world rather than the supply side defined by technology or business model. It is a way to see what customers actually want instead of trying to figure out how to sell things to people. That reframing is the difference between benchmarking features and understanding why someone chose anything at all.

Supply and demand, he argues, are not as connected as everyone assumes: a product does not create demand, a struggling moment does, and sometimes that moment exists for years with nobody solving it. At Southern New Hampshire University in 2010, Paul LeBlanc noticed 50 or 60 anomalies — students paying full price but never coming to class, watching everything online. Studying them revealed a different job from the typical 18-to-24-year-old: these students were older, had already earned a degree or had tried college and failed, and now had responsibility and needed something new. The university had not built a product for them; when it asked how many people wanted to go back to school but could not, the answer was not a thousand or 10,000 but over 200,000 students.

Founders insist they already look for pain, but Moesta says business school taught everyone the wrong frame: it is not pain and gain, it is context and outcome. The fact that he skipped lunch, still has work to do, and has a podcast running is what makes a particular solution valuable in that moment. Framed that way, the need becomes a vector of progress or intention that technology can be wrapped around, rather than a generic problem to solve.

He also argues that a kick-ass half is better than a half-ass whole, and points to QuickBooks, which he says has half the features and double the price. The point is to meet customers where they are rather than trying to wow or convince them; they convince themselves to make the progress. Adding more product is not the same as making progress easier.

People do not randomly do anything, so the heart of the method is causation: what made someone say today is the day I have to do something different? The push they feel — force one — usually has nothing to do with the new product, and it is the only reason they would leave the old one. Without a push, they cannot even see your product, because people are creatures of habit. Once there is a push, some idea of what is possible creates force two: the pull toward a new outcome or a new state.

Every new thing also creates anxiety — force three — and every old thing has the habit of the present — force four. If F1 and F2 are not greater than F3 and F4, people do not move, which is why Moesta frames a market as a system of behavior. Adding features increases pull but also increases anxiety ("can it really do all those things?"), while reducing friction can change behavior without touching the product at all. He illustrates this with a condo project for downsizers: buyers moving from 3,000-square-foot homes kept canceling six weeks after signing because they did not know how to get rid of all their stuff. Moesta raised the condo price and included moving plus two years of storage in the deal, and sales rose more than 30%.

Autobooks, a Detroit company that lets banks offer invoicing through things like Apple Pay instead of Square or PayPal, had to sell both small businesses and banks. When Moesta's team studied the banks, they found three very different reasons a bank would want the product. The company's process, however, was built around its own timeline: talk about the struggling moment, get the prospect to a demo, then close. What was missing was where the buyer actually was in their own buying process.

Moesta describes distinct phases: first thought, passive looking (problem aware but solution unaware), active looking (problem and solution aware, trying to frame a solution), and deciding (making trade-offs). When he asked the team where the customer was in their timeline of buying, they realized they were trying to close people who were still passively looking. They broke the demo apart and asked people where they were before deciding what kind of demo to give.

The result was three different demos: one that told stories and gave background on the problem, one that showed all the alternatives, and one that offered choices between ways to move forward. It sounds like it would lengthen the sales cycle, but Moesta says it nearly halved it and produced roughly 4x conversion, because the process met buyers where they were instead of where the seller wanted them to be. The seller's stages and the buyer's stages are not the same thing.

Moesta expands the buying timeline to six phases: first thought, passive looking, active looking, deciding, first use, and ongoing use — how the new habit gets built. He contrasts what people say before a purchase with what they actually do: 93% of prospective home buyers told him they wanted an Energy Star-compliant house, which cost about $30,000 to add, but nobody bought it; they bought the finished basement instead. This is why the method resembles criminal and intelligence interrogation rather than traditional market research — you ask people to tell the story of how they decided today was the day. It is also why he distrusts sales processes built on probability funnels: the real question is how many people are actually ready for your product.

The interview process starts by framing the question: what causes someone to say today is the day they want a new set of windows, or a vacation? Then you find people who recently bought — your product or someone else's — and ask what in the world happened. If you have not built anything yet, you ask what people will stop using when they hire you. The stories yield the pushes, pulls, anxieties, habits, trade-offs, and hire-and-fire criteria.

Rather than extracting themes across interviews, Moesta clusters them into pathways, because reasons come in sets: particular pushes pair with particular pulls, and the absence of a push makes a pull irrelevant. Most products, he says, are hired for three, four, or five different jobs that conflict with one another — one customer wants speed, another wants thoroughness. Intercom found four distinct reasons people hired it, then turned off the features that were irrelevant to each pathway: its acquire job competed with HubSpot, its support job competed with Zendesk, and pricing was set to match the progress being made. Moesta credits that segmentation with helping Intercom grow to a valuation over $2 billion.

In the middle of the conversation Moesta discloses that he has had three closed-head brain injuries before the age of seven and cannot read or write. He sees the spaces between words and the left edge of words first; his mother taught him to circle the five longest words on a page and work out what they have in common. The underlying problem is retrieval speed — by the time he decodes a word, he has forgotten the words before it. He says the condition gave him unusual pattern recognition: he can remember the first five words of a book's first paragraph and the last five of the last paragraph, and comprehend as well as anyone after three or four passes.

He writes through Scribe Media: ten two-hour recorded conversations, transcribed into a book that sounds like him talking, produced in about three and a half to four months. He now teaches as an adjunct at Kellogg, guest lectures at business schools, and works with Techstars and Y Combinator — a path that began with being told in high school to become a baggage handler or construction worker. He has extended the framework to the jobs of jobs: employees hire companies more than companies hire employees, and the most common phrase people use about how they got their job is "I was so lucky." When he unpacks the story, luck disappears into preparation, pushes, pulls, and anxieties — and more than 50% of people who took a new job did not get more money, because money often stands in for respect, responsibility, or a different measure of progress.

Moesta's first interviewing tip is to read Never Split the Difference by Chris Voss. One technique he uses constantly is playing something back incorrectly so the person says no and corrects him — once someone says yes, there is nothing more to say. His second rule is to talk only to people who have already tried to make the progress, because intentions are unreliable. When a large social media company noticed transactions happening on its platform without having built anything, his team studied eBay, Etsy, and Craigslist to learn what made someone say today is the day; that work mapped the jobs of both sellers and buyers and preceded what Moesta says is now almost a $3 billion marketplace. Lenny guessed Facebook Marketplace; Moesta declined to confirm.

On sample size, Moesta says the sets of pushes, pulls, anxieties, and habits start repeating after about seven or eight interviews. He usually does ten and never more than twelve, and he would rather run two rounds of twelve than one round of twenty-four. The influence here is Dr. Deming, the father of lean and quality systems, who pushed him to do things faster and smaller. Statistical significance, he argues, matters when you sample randomly; if you know the range of your market — say, half above 30 and half below — you can sample for representation and use design experiments instead of brute-force volume.

His third tip is counterintuitive: do not use a discussion guide. Everyone wants to ask the same set of questions, but that prevents you from following the thread with the most meaningful information in it. He uses pushes, pulls, anxieties, and habits as the frame and lets the rest be a conversation, avoiding "why, why, why" in favor of "tell me more about that" and "give me an example." When the person reaches the edge of their language, he brackets — "was it more about this or more about that?" — knowing neither is quite right, because the correction forces elaboration. He deliberately plays things back wrong for the same reason: the person will explain why it is not that.

For a lightweight first pass, Moesta sends people to ten recent buyers and tells them to ask not about the product but about why they bought it: what was going on, what they hoped for, what they worried about, what they had to give up, and how they convinced someone else. He listens for three sources of energy that have to exist for anyone to act: functional energy (time, space, effort, knowledge), emotional energy (how they feel and want to feel), and social energy (how others perceive them or how they want to be perceived). The same story usually contains all three, and the contradictions between them are where the insight lives.

If the product is established, he also talks to people who churned, because churn is not a dead end — the customer is still making progress, just somewhere else. They can describe the moment when the product stopped working for them, which is often a change in context rather than a missing feature. Nobody wants to change, Moesta says, which is exactly why studying why people change is the easiest route into the job.

His warning about stated demand is memorable: "Bitchin' ain't switchin'." People complaining about a missing feature does not mean they will act on it, which is what Basecamp learned when users demanded Gantt charts and resource allocation but would not have left over them. Following your best users can also pull a product upmarket into complexity and destroy the simplicity that brought everyone else in. Moesta's rule is to study people who already did something, even if they failed, rather than people who say they will.

He describes layers of language: the pablum layer ("it was good"), the fantasy-or-nightmare layer where people exaggerate, and then what actually happened. A car buyer will say he got a deal on a car he had been dreaming about, until the story reveals 280,000 miles, three large bills in four months, a noise, and a long trip coming up. One interviewee said she bought a $137 coat rack in a week; the debate about buying it had been going on for 18 months. This is also why Moesta insists value is relative to the starting point: if someone only wants to get a little way forward, overshooting the outcome makes them ask for a discount rather than pay for more.

Asked for a zero-to-one example of Jobs to Be Done producing broad adoption, Moesta points to the same social media marketplace and to Autobooks, and argues the framework's main value is deciding what not to build. Large organizations, he says, default to hypothesis-testing research, while Jobs to Be Done is hypothesis-building research: you genuinely do not know what you think you know. He cites Drucker's 1953 observation that what businesses think they are selling is not what customers are buying, and says it is still true. At Techstars, companies such as Nutrisense and Havoc Shield start with jobs work before they have a product; the framing question is what people will stop using when your product arrives, because there are no new jobs, only better ways to do existing ones.

Lenny notes that Twitter, where a previous guest had worked, tried a version of Jobs to Be Done and left people sour on it. Moesta distinguishes flavors: one is supply-side driven, taking the underlying technology and systematically scoring where it can be better, often with hundreds of prescriptive steps. His own method is qualitative and organic — a combination of process, practice, and skills — and he argues every company ends up with its own innovation process based on its people, customers, and technology. He also says Jack Dorsey was a fan who worked with Clay Christensen on the thinking rather than the method. The most dangerous version, in his view, is sitting in a room and hypothesizing the jobs, because you will be wrong.

Moesta clarifies the family tree: he collaborated with Clay Christensen, whom he met with once a quarter for 27 years, and helped on Competing Against Luck, but Christensen turned the work into a theory while Moesta kept it as a method. Tony Ulwick's approach comes from a different dataset and starts from functions — what jobs a product can do — whereas Moesta insists only people have jobs, not products, and not organizations; people inside organizations have jobs. He calls Ulwick's approach valuable, especially in regulated, high-risk, complicated systems, but notes it requires a disciplined organization to follow so many steps. Christensen's and Moesta's approaches, by contrast, derive from the same dataset.

Jobs to Be Done is the wrong tool when there is no real choice or when the purchase is too habitual to register. Moesta contrasts car insurance, which people actively shop, with employer-provided health insurance, where the choice is essentially good, better, best and only becomes real when someone is sick. He also points to chewing gum: people cannot remember buying a pack, but they can describe when they chew it, and Tide users who have never switched will invent a reason for their loyalty. In those cases he recommends ethnography, prototyping, and studying the "little hire" — how people use something rather than why they bought it. He also refuses engagements where a company wants him to prove the jobs it has already decided exist; the method only works if you accept how people see themselves.

Moesta says he deliberately kept the framework accessible rather than copyrighted or patented, so people could try it and adapt it. The misconception that frustrates him most is the reduction of Jobs to Be Done to pain and gain instead of context and outcome, or to outcomes alone without the starting context. The other repeated mistake is doing it in a conference room without talking to anyone, which guarantees you miss the contradictions and irrational parts. His core claim is that context makes the irrational rational: when a story sounds impossible, nine times out of ten it is because you do not have the rest of the story — and innovation happens where people change direction, not where they continue on momentum.

He closes with three takeaways. First, struggling moments are the key, they are everywhere, and only certain contexts make people act — so study them, because that is where innovation is needed most. Second, think about the progress people are trying to make against their standard, not yours: what is the context, what is the outcome? Third, choose what to suck at: every product makes trade-offs, and most failures come from a trade-off the customer did not agree with.

Asked for a small change with a big effect, Moesta cites Intercom's executive team doing the interviews themselves rather than delegating them; Paul Adams and Matt Hodges started at Intercom in his Detroit office. In large organizations, he looks for a small group that is struggling to ship or getting the wrong insights, gives it room to demonstrate the method, and lets case studies spread it. His favorite illustration is a Detroit condo project for downsizers: buyers insisted they did not want a dining room table and would not host holidays, yet many would not move until they knew where the table would go. Moesta sacrificed the second bedroom suite to build a symbolic place for the table — one you could never actually eat at — and sales rose 22%.

Above his desk, Moesta paints portraits of his mentors: Dr. Deming, Clay Christensen, Dr. Taguchi, and Dr. Willie Moore. Deming met him at 18 and took him to Japan; Moesta worked on the front lines at Ford helping cut product development cycle time from 72 months to 36, where he learned Toyota's tools and methods and Taguchi's approach to designed experiments. Willie Moore, his first boss at Ford and a PhD in particle physics, taught him empathetic perspective. Christensen came last: Moesta walked into his office because of a sign that said "anomalies wanted," introduced himself as an anomaly, and asked how he could help — the first person to do so in Christensen's year there — which began a 27-year relationship with four hours a quarter and no agenda.