Clayton M. Christensen

Christensen blames hit-or-miss innovation on data that shows correlation, not cause. In a McKinsey poll, 84 percent of executives called innovation extremely important to growth, yet 94 percent were unsatisfied with their results. In 2015 one thousand public companies spent $680 billion on R&D.

Stage-gates, rapid iterations and six-sigma rules look scientific but create the feeling of progress without causing it. A process built on the belief that innovation cannot be understood will stay good at describing and bad at predicting.

He uses his own profile to show data cannot explain a purchase. He is sixty-four, six feet eight, wears a size sixteen shoe and drives a Honda minivan in a Boston suburb. None of that makes him buy the New York Times.

He might hire the paper to skip dull lobby magazines before a doctor's visit, follow March Madness, or look informed to neighbors. No ZIP code or income band would surface those reasons.

He cites fields that became reliable once they found the causal mechanism. Pasteur's germ theory ended about two thousand years of trial-and-error medicine. Deming and Juran taught Japanese carmakers to trace each defect to its root cause while Detroit kept building lemons.

At Toyota each step is done the same way every time, so each repetition works as an experiment. He sees innovation as standing where carmaking stood before the quality revolution.

Christensen treats failed launches as anomalies with causes, not bad luck. He endorses discovery-driven planning from Ian MacMillan and Rita McGrath: list the assumptions a strategy rests on and test them before heavy investment.

Do not approve a project on its first projections and later bend them to fit what happens. He says theory should teach how to think, not what to think.

His disk drive research showed well-run firms losing to cheaper entrants. The work came from doctoral research at Harvard Business School. Managers doing everything right took their firms to the top and then off a cliff.

Digital Equipment Corporation collapsed around 1988 and almost every other minicomputer maker failed at the same time. Because the whole industry fell together, the bad-manager explanation failed.

Disruption theory predicts how incumbents respond, not what to build. Even a firm setting out to disrupt has under a 25 percent chance of building exactly the right product.

The idea later stretched to cover anything clever and new. He says it stays useful for spotting threats and predicting incumbent behavior.

He asked the reverse question: how do successful firms know where to grow? His answer is what makes customers pull a product into their lives. The Theory of Jobs to Be Done fills the gap disruption leaves.

The book opens with a fast-food chain whose shake sales did not move after it made shakes cheaper, chunkier or chocolatier.

The circumstance, not the customer, decides what gets hired. Eighteen hours of watching showed many shakes sold before 9:00 a.m. to lone commuters who wanted something to last a boring drive and fit the cup holder. The shake beat bananas, doughnuts, bagels and a Snickers bar because it lasted about twenty minutes.

In the afternoon the same buyers hired a shake for a different job: a parent wanting to say yes to a child. There the rivals were a toy store stop or a game of catch.

Every job has functional, social and emotional sides, and the last two can win. A smoker's break delivers nicotine, calm and time outside with friends. A Facebook break delivers rest and a virtual water cooler, so Facebook partly competes with cigarettes.

P&G's cheap Pampers sold weakly in China as containment. Reframed around better sleep, backed by research showing babies slept thirty more minutes, they reached about $1.6 billion in sales and roughly 30 percent share by 2013.

He insists a job be written in verbs and nouns, never adjectives. "A chocolate shake in a twelve-ounce cup" is a spec, because every answer is a shake. Staying occupied and full until a 10:00 a.m. meeting is a job, and bananas, bagels and coffee all compete for it.

If every option sits in one product class, move up a level. A spec for moisture-proof house wrap leaves Tyvek or nothing. Comfort in every season at low energy cost also admits fiberglass, sweaters and fans.

A job is narrower than a need and outlasts any one solution. Needs like eating are too generic and themes like being a good husband too abstract. Segway answered a general need for personal transport and by most measures flopped.

Jobs are discovered, not created, and the rule that follows is to love the job, not your solution to it.

He shows industry categories hiding the real rivals. Unilever held about 70 percent of US margarine and its roughly twenty-one shelf brands differed mainly on fat percentage. Shoppers hired spreads to moisten bread or keep food from burning, so cream cheese, olive oil and nonstick spray competed too.

Butter overtook margarine by the mid-2000s and Unilever said at the end of 2014 it would spin spreads into a stand-alone company. He notes trans fat worries partly drove butter's rise.

Real rivals are often other categories, or doing nothing at all. Reed Hastings named video games, wine and board games as Netflix rivals. BMW redefined its job as mobility and found new rivals in Tesla, Uber and Zipcar. Chip Conley says 40 percent of Airbnb guests would not have traveled, or would have stayed with family.

John Doerr had asked whether Netflix competed with Amazon, and Hastings answered with everything people do to relax.

Counting nonconsumers reveals markets far bigger than rivals' sales. Small business owners jury-rigged Quicken to keep books. QuickBooks offered half the functionality at twice the price and became the global leader, and Intuit reached $4 billion in revenue.

Its real rivals were bookkeepers, extra hours, spreadsheets and shoeboxes of receipts. Owners wanted confidence that invoices went out and bills got paid, not accounting software.

Customers who look identical can hire you for unrelated jobs. SNHU split campus coming-of-age students from online adults needing speed, service and credentials. It cut its reply time from a 24-hour email to a call within ten minutes, and online revenue grew from $32 million to nearly $535 million.

It moved the online team two miles away to a former mill yard, free of legacy rules. About twenty people charted admissions on a whiteboard and circled every hurdle.

He says perspective, not a new research tool, finds a job. Personas, ethnography and focus groups work when read through a jobs lens. He names five sources: your own frustrations, nonconsumers, workarounds, chores people dread, and odd uses of products.

Sal Khan's cheaper and crappier math videos for his cousin won. Kimberly-Clark's Depend Silhouette earned $60 million in year one, and ZzzQuil came from people taking NyQuil just to sleep.

He tells teams to drill deep on one vein before going wide. Pick one source of struggle, then widen sideways to capture the full context. Record functional, social and emotional dimensions before designing anything.

Intermountain Healthcare's Todd Dunn felt the emotional score of a sterile exam room as a patient. Intermountain then adopted a design method that unpacks all three dimensions before any innovation brief.

He interviews recent buyers as a timeline back to the first thought. Brian Walker looked like an impulse buyer of a $699 Costco mattress. His $2,000 pillow top had given him a year of pain, he had researched for three months, and Costco's easy returns removed his last fear.

The interview skips feature questions. He says: "Great innovation insights have more to do with depth than breadth."

He trusts what people do in a circumstance over what they say. Sony's research said buyers would reject a player that could not record, until Akio Morita overruled marketing. The Walkman sold over 330 million units. American Girl's focus group rejected the idea that became a $700 million sale to Mattel.

Sales data mostly records the purchase and can be misread, so he asks for the full story of the struggle.

He maps push, pull, habit and anxiety, not just the pull of the new. A switch happens only when push and pull outweigh habit and anxiety. Losses weigh about twice as much as gains, so the new must be far more appealing, not slightly better.

Health clubs found annual contracts created so much anxiety that people did not join. ING Direct's cafés handle no teller cash transactions and exist mainly to show the virtual bank is real.

Every job involves two hires and a firing. The Big Hire is the purchase and the Little Hire is each later use. Only repeated Little Hires show the job is getting done.

Hiring a product means firing something, even doing nothing: mobile phones fired wristwatches and Depend buyers fired staying at home. Airbnb storyboarded forty-five emotional moments before launch.

A condo builder removed buyers' anxieties and cut cancellations. Interviews found no demographic pattern but a recurring worry about the dining room table. The builder shrank the second bedroom 20 percent, added moving help and storage, and cut options to three.

Cancellations fell from five or six a month to one and a $3,500 price rise held. In 2007 it grew 25 percent while industry sales fell 49 percent.

He wants a written job spec before any design starts. It lists the three dimensions, the tradeoffs customers accept, every competitor including nonconsumption, and the obstacles and anxieties to remove.

Medtronic's better pacemakers sold poorly in India. Interviews with over a hundred stakeholders found four barriers, and free screening, two-day bridge loans and patient counselors followed. About 167,000 patients were screened and about 15,000 treated through December 2015.

He says lasting advantage lies in experiences rivals struggle to copy. Rowland's belly band added two cents and twenty-seven seconds to American Girl packing, and she refused to drop it because the wait makes unboxing exciting. Cheaper lookalikes from Toys"R"Us, Walmart and Disney made no dent.

American Girl has sold 29 million dolls, and a typical buyer spends over six hundred dollars. IKEA and Uber win the same way, on the experience rather than the product.

Nailing the job earns a premium customers are grateful to pay. Milwaukee's Sawzall and HOLE HAWG have held more than 80 percent of their job markets for decades against comparable rivals.

A premium forced through lock-in, like printer ink, breeds resentment. His test: "With jobs-based innovations, customers don't resent the price, they're grateful for the solution."

He tells builders to say who should not hire them. A restaurateur's early Yelp raves drew foodies hunting a hidden gem to a decent local place, and those visitors wrote bad reviews.

Airbnb's Chip Conley pictures an invisible report card on each guest's forehead, so listings avoid overselling.

He holds up Volvo as a purpose brand that lost its purpose. Volvo stood for safety from 1927. After Ford bought it in 1999 it chased luxury, was unprofitable by 2005 and was sold to Geely at a loss in 2010.

It grew again in 2015 after refocusing on safety, but he says it may have lost purpose-brand status for good.

He argues processes are the hardest asset to copy. Toyota let rival executives tour its plant twice a month and they still missed the bigger picture. Pixar's Ed Catmull published its processes and rivals have not matched its results.

SNHU gives tours to would-be competitors, and Paul LeBlanc says richer rivals cannot easily copy hundreds of linked processes.

He says to put someone in charge of the job, not redraw the org chart. In a 2010 Bain study, fewer than a third of 57 major reorgs produced meaningful improvement. At the Mayo Clinic in 2015, one coordinator booked his specialists and an MRI done by 11:30 ahead of a 2:00 visit.

Functional silos will likely stay. The change comes through processes, metrics and culture.

He wants customer-benefit metrics in place of internal efficiency. Intuit tested forcing QuickBooks trial users to phone in to register so sales could upsell them. Revenue rose, but users who only wanted to register hit busy lines and sales pitches.

Intuit now builds a benefit metric per product, such as time saved. Amazon opens each project with a mock press release stating the customer job, shown to every function at once.

Listening to real customer calls turned OnStar's job into peace of mind. OnStar began as cool features for luxury cars. CEO Chet Huber had all 300 staff spend an hour on call-center calls, where they heard crash victims in mid-accident.

OnStar built its own fast upgrade process, so GM cars ran a fifth-generation version while rivals ran version three. Ford scrapped its Wingcast rival two years after announcing it.

A short job phrase lets people decide without orders. TurboTax teams spent weeks ranking about 150 requested features until the job was reframed as taxes done. Before Hurricane Rita, OnStar gave every caller in the crisis area all services free after about fifteen seconds of approval.

He compares a clear job to commander's intent. American Girl staff could all recite "story, not stuff."

After launch, loud sales data pulls companies back to products. V8 pitched itself against peeling and boiling carrots and sales quadrupled in under a year. After that team moved on, the line sprawled into variants like Picante and V8 Splash! and went back to fighting other juices.

He says to name a champion for the job and gather passive data on purpose. He admits whether V8 truly lost the job is a suspicion based on partial knowledge.

He warns that selling more to existing customers blurs the job. DeWalt built its name on the best radial arm saw. After Black & Decker bought it, it copied drills and clamps from store shelves.

Deseret News went the other way. It dropped breaking news for next-morning analysis aimed at like-minded believers, about 56 percent of news consumers, and grew to over 100 million social followers.

New processes trigger an immune response, as the milk shake chain showed. The chain never implemented Bob Moesta's recommendations. Local managers found the changes to routines and resource allocation too hard.

No one killed the idea outright and it died quietly. Clear insight into the job does not survive on its own.

He names cases where jobs theory does not help. It does not help where there is no real struggle or current solutions are good enough, or where choices are pure math, such as commodities trading.

The theory explains why people choose, not whether you should offer the solution. Cigarettes do many jobs yet harm the customer, so ethics stay a separate question.

He treats anomalies as a reason to refine the theory, not abandon it. When a case does not fit, go back to the phenomena. Then refine the theory or mark a boundary past which it should not be used.

He warns that stretched too far a theory loses power, as happened with disruption, and says job risks the same fate if every motive gets the label.