The Idea That Reframed Innovation
Clayton Christensen spent his career studying why great companies fail. His 1997 book The Innovator's Dilemma introduced "disruptive innovation" and made him, in the words of The Economist, the most influential management thinker of his time. He taught at Harvard Business School until his death in January 2020, and the theory he's best remembered for outside academia isn't disruption at all. It's Jobs to Be Done.
The idea went public in a 2005 Harvard Business Review article, "Marketing Malpractice: The Cause and the Cure," written with Intuit founder Scott Cook and researcher Taddy Hall. Christensen's frustration was simple. Companies pour money into customer research, slice their market into ever-finer demographic segments, and still launch products nobody wants. Something in the standard approach was broken.
His fix started with a borrowed line. "The great Harvard marketing professor Theodore Levitt used to tell his students," Christensen wrote, "'People don't want to buy a quarter-inch drill. They want a quarter-inch hole.'" The line has a longer history than that. Levitt himself credited it to a man named Leo McGivena, and the earliest known version showed up in a 1942 advertisement. Christensen didn't coin it, but he did the most with it. He turned a clever aphorism into a working theory of demand.
The core reframe: customers don't buy products. When a need shows up in their lives, they "hire" something to make progress, and they "fire" it the moment a better option appears. Your product isn't competing in a category. It's competing for a job.
What a Job to Be Done Actually Means
A "job" in this theory isn't a task on a to-do list. Christensen defined it precisely in his 2016 HBR article "Know Your Customers' Jobs to Be Done": a job is "the progress that a person is trying to make in a particular circumstance." Two words carry the weight. Progress, because people are trying to move from one state to a better one. Circumstance, because the same person hires different things in different situations.
That framing sounds abstract until you add the three dimensions every job carries:
- Functional: the practical, measurable task. Get to work. Fill an empty stomach. Learn a new skill.
- Emotional: how the person wants to feel. Confident, calm, in control, less anxious.
- Social: how the person wants to be perceived by others. Competent, generous, ahead of the curve.
Most companies optimize the functional layer and forget the other two. That's why a product can be objectively better on paper and still lose. If a customer hires your app partly to feel organized and partly to look serious about their work, a faster load time doesn't touch the reasons they actually chose it.
The "hire" and "fire" language matters more than it looks. It forces you to ask what the customer would have used instead. A note-taking app doesn't compete only with other note-taking apps. It competes with a notebook, a photo of a whiteboard, a Slack message to yourself, and doing nothing at all. The set of things that could get the same job done is your real competition, and it's almost always wider than your category.
This is the connection to Christensen's other big idea. Understanding the job is how you spot the moment a "good enough" cheaper option can be hired for a job that expensive incumbents overserve, which is the seed of the disruption he wrote about in Peter Thiel's argument about monopoly and competition.
The Milkshake That Taught Marketers to Ask Why
The story that made Jobs to Be Done famous starts with a fast-food chain trying to sell more milkshakes. (The case is often told as McDonald's, though Christensen and the Christensen Institute usually describe it generically, so treat the specific chain as unconfirmed.) The company had already done the obvious things. They surveyed milkshake buyers, asked whether the shakes should be thicker, cheaper, chunkier, or more chocolatey, tuned the product to match the answers, and saw no change in sales.
Bob Moesta, a researcher working with Christensen, took a different tack. Instead of asking who bought milkshakes, his team stood in the restaurant and watched when people bought them and what was going on in their lives at that moment. The data surprised everyone. Nearly 40% of milkshakes were sold early in the morning, bought by lone commuters, taken to go, and almost always the only thing the customer purchased.
So the team interviewed those morning buyers. The job turned out to have nothing to do with dessert. These people faced a long, dull drive to work. They weren't hungry yet, but they would be by mid-morning. They had one free hand and needed something to make the commute less boring. A bagel was dry and left crumbs and needed two hands. A banana was gone in two minutes and didn't fight off hunger. A doughnut was gone almost as fast. The thick milkshake, sucked slowly through a thin straw, took 20 minutes to finish, kept one hand busy, rode neatly in the cup holder, and quieted hunger until lunch.
The competition wasn't other milkshakes. It was bananas, bagels, doughnuts, and boredom. Once you see the job, the product improvements write themselves. Make the morning shake even thicker so it lasts longer. Add small chunks of fruit to give a little surprise on the drive. Move the machine to the front counter so rushed commuters can grab and go.
Then the team noticed the same product got hired for a completely different job in the afternoon. Parents bought milkshakes as a small treat for their kids after school, and there a thick shake was a problem, because the child took forever to finish it while the parent waited. One product, two jobs, two ideal designs. No demographic survey would ever have surfaced that, because the buyers looked nothing alike and the same buyer even switched jobs by time of day.
One honest caveat: the widely repeated claim that this insight produced a specific sales jump has no solid source behind it. The milkshake story is valuable as a lesson in how to look, not as a documented revenue result.
Why Demographics Lie and Jobs Don't
Traditional segmentation builds a customer profile: age, income, education, location, a stock-photo persona with a name. It feels rigorous, and Christensen argued it quietly misleads. His blunt version: "The fact that you're 18 to 35 years old with a college degree does not cause you to buy a product." Demographics correlate with buying. They don't cause it.
The distinction between correlation and causation is the whole point. A profile can be statistically accurate and still useless for building a product, because it describes the buyer rather than the buyer's motivation. The morning commuter and the afternoon parent might be the exact same 40-year-old person, hiring the same milkshake for two unrelated jobs a few hours apart. Segment by person and you see one customer. Segment by job and you see two markets.
Here's the practical difference the two lenses produce:
| Question | Persona / demographic lens | Jobs to Be Done lens |
|---|---|---|
| Who is the customer? | 35-year-old suburban parent, mid income | Anyone with a boring commute and one free hand |
| What do they want? | A better milkshake | Progress: stay occupied and not hungry until noon |
| Who is the competition? | Other milkshake brands | Bagels, bananas, doughnuts, boredom |
| How do you win? | Tweak flavor and price | Redesign for the actual circumstance of use |
| What does research look like? | Survey existing buyers | Watch the moment of purchase and struggle |
The jobs lens keeps pointing you back to a circumstance and a struggle, which is where the money is. That's also why it pairs so well with how founders find opportunities in the first place. The best startup ideas come from noticing a job people are doing badly with the tools they have now, which is the throughline of Paul Graham's advice on getting startup ideas and of navigating the idea maze.
The Four Forces That Move a Customer
Knowing the job is half the picture. The other half is understanding why customers stick with a worse option even when a better one exists. Bob Moesta, working with Chris Spiek at the Re-Wired Group, gave the JTBD community its sharpest tool for this: the four forces of progress.
Every decision to switch from an old solution to a new one is a tug-of-war between two forces pulling toward change and two forces pulling back toward the status quo.
| Force | Direction | What it is |
|---|---|---|
| Push of the situation | Toward change | The frustration with the current way that starts the search |
| Pull of the new solution | Toward change | The appeal of the new option and the progress it promises |
| Anxiety of the new | Away from change | Fear that the new thing won't work, is hard to learn, or is a mistake |
| Habit of the present | Away from change | The comfort and inertia of the way things are done now |
The rule is simple arithmetic with a hard lesson inside it: a switch happens only when Push plus Pull outweighs Anxiety plus Habit. Most product teams spend all their energy pumping up the Pull, adding features and sharpening the pitch. They ignore the two forces holding the customer back.
That's a mistake, because Anxiety and Habit are often the bigger numbers. A customer who is intrigued by your product but afraid of the migration, unsure it will work, and comfortable with what they already have will not switch, no matter how good your feature list is. Reducing anxiety (free trials, guarantees, hand-holding onboarding, social proof) and easing out of old habits (import tools, familiar defaults) frequently moves more customers than another feature. The four forces explain the awkward truth that a genuinely better product can still fail to get hired.
Jobs to Be Done in the Wild
The theory earns its keep in real companies. Three examples show the range, from software to education to a candy bar.
Intercom may be the clearest software adopter. The customer-communication company built its product and marketing strategy on JTBD and published a book about it, Intercom on Jobs-to-be-Done, written by Des Traynor, Paul Adams, and Geoffrey Keating. Instead of asking which features to add, they framed decisions around the job a customer hired Intercom to do, which kept the roadmap anchored to progress rather than a wish list.
Southern New Hampshire University is the most rigorously documented case. President Paul LeBlanc suspected that online adult learners were hiring the university for a different job than traditional 18-year-olds on campus. Bob Moesta's team ran JTBD interviews and found the online-learner job was closer to "help me step it up," usually triggered by a career milestone, a layoff, or financial pressure. Those students needed four things: convenience, strong customer service, a credential that mattered, and speed to completion. SNHU rebuilt admissions and financial aid around that job, cutting response times from weeks to days. By the case's own reporting, online enrollment grew from under 500 students to more than 130,000 over roughly eight years. (The revenue and enrollment figures come from the university and its consultants, so read them as case-reported rather than independently audited.)
Snickers shows the idea in positioning. The "You're not you when you're hungry" campaign, launched at the 2010 Super Bowl, sells the bar against the job of fixing hunger and getting back to yourself, not against other chocolate. That reframes the competition as anything that satisfies hunger in the moment. Mars reported strong sales gains in the campaign's first year. Worth noting: this wasn't a Christensen case study, and the agency didn't use his vocabulary. It's an apt illustration of job-based competition, not part of the original research.
| Company | The job customers hire it for | What changed |
|---|---|---|
| Intercom | Talk to my customers where progress is happening | Roadmap organized around jobs, not features |
| SNHU | Help me step up my career, fast and with support | Rebuilt admissions and aid around convenience and speed |
| Snickers | Fix my hunger and make me myself again | Positioned against all hunger fixes, not just candy |
The pattern across all three: the winning move was to define the competition by the job, then design everything (product, process, message) to do that job better than the alternatives the customer would otherwise hire.
How to Find the Job Your Customers Are Hiring
The theory is only useful if you can uncover the job, and interviews are where it happens. Christensen and Moesta both emphasized studying the moment of purchase, especially the switch, when someone fired an old solution and hired a new one. Here's a practical sequence:
- Find recent switchers. Talk to people who just started using your product (or a rival's). The decision is fresh, so the real reasons haven't been rewritten by hindsight.
- Rebuild the timeline backward. Start from the purchase and walk back: when did they first realize the old way wasn't working? What happened that day? What did they do next? You're mapping the push.
- Hunt for the struggle. People rarely state their job directly. They describe frustrations and workarounds. The job hides inside the struggle, so listen for the moment things got bad enough to act.
- Name all three dimensions. Write the functional job, then push for the emotional and social ones. "I wanted to organize my notes" is functional. "I was embarrassed that I couldn't find things in meetings" is the real fuel.
- List what they almost hired. The alternatives they considered, including doing nothing, reveal the true competitive set and the four forces at play.
Notice that steps 2 through 4 are really about noticing and capturing detail before it fades, which is exactly the reading and research skill Glasp is built for. When you're studying a market, Glasp's web highlighter lets you highlight the exact phrases in customer reviews, forum threads, and support tickets where a struggle shows up, turning scattered evidence into a searchable record. A huge amount of customer and founder insight lives in interviews and talks on video, and YouTube Summary by Glasp pulls the key points out of an hour-long conversation in minutes so you can highlight the moments that reveal a job. Once you've gathered enough, Glasp's AI chat lets you interrogate everything you saved, asking what job keeps coming up across dozens of sources instead of trusting your memory of one loud interview.
This is the same discipline that separates real demand from imagined demand in finding product-market fit: you're looking for evidence of a job people already struggle with, not a feature you wish they wanted.
Jobs to Be Done in the AI Era
The theory has become newly relevant, because AI has made features cheap. When any team can bolt a chatbot onto a product in a weekend, the feature stops being a differentiator. What lasts is understanding the job well enough to know which capability actually helps.
The most common failure mode in AI products right now is building from the model outward: "Look what this can do, who might want it?" Christensen's whole argument runs the other way. Start from a job customers already struggle to do, then ask whether AI does that job better than what they hire today. A well-designed AI feature reduces the Anxiety and Habit forces (it removes tedious work, it fits an existing routine) rather than adding a novelty that demos well and gets fired in a week.
The framing also clarifies what customers hire an AI tool for. Functional jobs like summarizing a report, drafting an email, or analyzing a spreadsheet are the obvious ones. The emotional jobs matter just as much: feeling less overwhelmed, trusting that nothing important slipped through, looking capable in front of a boss. Tools that win the emotional job tend to keep the customer even when a rival matches them on raw capability. This is commentary rather than canon, since Christensen died in 2020, but it's a faithful extension of the theory, and it explains why so many technically impressive AI features never get hired twice.
For anyone building in this space, the jobs lens is a filter against hype. It turns "what can the model do" into "what progress is the customer stuck on," which is the only question that reliably points at something people will pay for. The founders who build durable knowledge about their users, not just their technology, are the ones described in how founders build knowledge systems.
Frequently Asked Questions
What is the Jobs to Be Done theory in simple terms?
It's the idea that customers don't buy products for their own sake. They "hire" a product to make progress on a job in a specific situation, and they "fire" it when something does the job better. Clayton Christensen defined a job as "the progress that a person is trying to make in a particular circumstance." The practical payoff is that you design and market around the customer's real goal, not around your product's features or the customer's demographics.
Who created Jobs to Be Done?
Clayton Christensen, a Harvard Business School professor, popularized it, most publicly in a 2005 HBR article and his 2016 book Competing Against Luck. The intellectual roots go back further, to Theodore Levitt's "quarter-inch hole" line and Peter Drucker's writing on what the customer considers value. The four forces of progress, a key extension, came from Bob Moesta and Chris Spiek at the Re-Wired Group, not from Christensen himself.
What is the milkshake example about?
A fast-food chain wanted to sell more milkshakes and got nowhere by asking customers how to improve the product. Bob Moesta's team instead watched when shakes were bought and found nearly 40% went to lone morning commuters using them to survive a boring drive and stave off hunger until lunch. The real competitors were bagels, bananas, and doughnuts, not other milkshakes. It's the classic demonstration that the job, not the product category, defines the competition.
What are the four forces of Jobs to Be Done?
They're the forces that decide whether a customer switches solutions: the Push of the current situation and the Pull of the new option drive change, while the Anxiety about the new thing and the Habit of the present hold it back. A switch happens only when Push plus Pull is greater than Anxiety plus Habit. The lesson is that reducing anxiety and habit often moves more customers than adding features.
How is Jobs to Be Done different from buyer personas?
Personas describe who the customer is (age, income, role), which correlates with buying but doesn't cause it. Jobs to Be Done describes why the customer buys, the progress they're trying to make in a moment. The same person can hire the same product for different jobs at different times, which personas miss entirely. Christensen's point was that the job, not the demographic, is what actually causes the purchase.
Conclusion: Hire the Right Tools for Your Own Learning
Jobs to Be Done is usually taught as a tool for building products, but the same lens sharpens how you read and learn. Every book you open, every video you watch, every note you take is something you've hired to do a job. Most of the time nobody names the job, which is why so much reading turns into passive collecting: highlights you never revisit, saved articles you never open, a library that grows without making you any smarter.
Name the job first. Are you hiring this material to make a decision, to learn a skill you'll use next week, to understand a debate well enough to have an opinion? The answer changes what you highlight and what you can safely ignore. Progress, not accumulation, is the goal, exactly as it is for the customer in the milkshake line.
That's the job Glasp is built to do. Hire Glasp's web highlighter to capture the specific passages that move you toward your goal instead of drowning in whole articles, use YouTube Summary to extract the progress-making points from long talks, and query everything you've saved with Glasp's AI chat so your reading compounds into something you can actually use. Then see what job others are hiring the same sources for through Glasp's community feed. Figure out the job, hire the right tool, and fire the ones that only made you feel productive.