A bank spends two years and a small fortune learning that its small-business customers don't want a better business loan. They want to not think about cash flow. The insight is correct, well-researched, and sits in a deck. Nothing ships. A year later a startup launches exactly that product, signs up the same customers, and the bank's team watches it happen from the inside.

The bank did the hard part. It discovered the job. It just couldn't serve it.

Jobs to be done is a discovery framework: it tells corporations how to find what customers are trying to get done, but not how to serve a new job their own structure was never built for, which is why startups serve those jobs first and co-creation is how corporations reach them.

That is the gap the jobs to be done framework doesn't close. JTBD is a discovery tool, and a very good one. It tells you how to find what your customer is really trying to get done. It says nothing about what happens when your own organization is structurally incapable of serving the job once you've found it. For most corporations, discovery is the easy part. Execution against a new job, one the company wasn't built to serve, is where the work goes to die.

What does the jobs to be done framework actually say?

Jobs to be done is the idea that customers don't buy products, they hire them to make progress on a job in their lives. The milkshake isn't a milkshake. It's a hire for a boring morning commute, and it competes with bananas and bagels, not with other milkshakes.

The canon here is deep and worth crediting directly. Clayton Christensen, Taddy Hall, Karen Dillon, and David Duncan laid out the modern version in Competing Against Luck and in their 2016 Harvard Business Review article, "Know Your Customers' Jobs to Be Done." Tony Ulwick built the rigorous measurement layer with Outcome-Driven Innovation in Jobs to Be Done: Theory to Practice. Bob Moesta and Chris Spiek developed the demand-side interview and the Forces of Progress. Alan Klement carried the thinking forward in When Coffee and Kale Compete.

They wrote the book on discovering jobs. This piece is about the part the framework was never meant to solve: serving a new job from inside a company structured around the old one. That isn't a critique of JTBD. It's the next problem in line after you use it well.

Why can't corporations serve the new jobs they discover?

Because every system inside a mature company is calibrated to serve the jobs it already serves, better and cheaper. The blindness isn't cultural. Nobody decides to ignore new demand. It's structural, built into the incentives, and the incentives are doing exactly what they were designed to do.

Walk the org and you can name the failure modes:

FunctionWhat it optimizes forWhy the new job losesCustomer research (NPS, CSAT)Satisfaction on the current jobMeasures how well you serve today's customer, not whether a different job existsProduct roadmapRequests from existing customersInherits the current customer's priorities, which are about the current jobSales incentivesUpsell and expansion into the installed baseRewards selling more of the current job, never the unproven new oneR&D and financeNear-term return against the core P&LCuts new-job work first, because it shows no revenue yetLegal, risk, complianceKnown risks of the known businessTreats the unfamiliar shape of a new job as a threat to manage, not an opportunity to chase

Each of these is a strength of the core business. Together they form what we and others have called the organizational immune system, and it always wins. A new job arrives looking like risk, distraction, and margin dilution, and the system does what it's built to do. It protects the host. Defending today's revenue beats building tomorrow's in every quarterly review, because today's revenue is real and tomorrow's is a hypothesis.

So the new job gets discovered, documented, and slowly starved. Not by bad people. By good structure pointed at the wrong target.

Why do startups catch the jobs corporations miss?

A startup carries none of the structural weight that traps the corporate build. It has no installed base whose current job must be defended, no roadmap inherited from existing customers, no sales force compensated to sell the old thing, no core P&L that makes the new job look small.

The founder's advantage isn't vision or nerve. It's the absence of drag. A founder serving one new job obsessively will out-execute a corporate team serving ten jobs at seventy percent intensity, every time, because the corporate team is structurally divided and the founder is structurally free. The startup can build the whole company around a single job, price for it, hire for it, and let the market correct it in public. Speed and focus are not personality traits. They are what's left when you remove the immune system.

This is the real shape of the corporate-startup relationship. It isn't a competition to manage. It's a mismatch to exploit. The corporation has what a startup burns years and tens of millions chasing: distribution, capital, regulatory standing, customer trust, category expertise. The startup has the one thing the corporation can't manufacture internally, which is structural freedom from its own model. Corporates can't, startups do, and the gap between them is the opportunity, not the problem.

How does co-creation bridge the gap?

The bridge is to stop asking the corporation to serve the new job from inside, and to build a separate venture designed to serve it from day one.

In the language of dual transformation, this is Capability C: the link that lets the core business and the new business each operate on their own terms while sharing the assets that matter. The corporation supplies the advantages it has earned. The venture supplies the freedom to serve the new job without the immune system fighting it. Done deliberately, you get something neither side could build alone: a company built for the new job, standing on the parent's distribution and credibility.

That is co-creation. Not an innovation lab inside the building, where the immune system still has reach, but a real venture with its own cap table, its own incentives, and a founder with something to lose. The corporation buys access to a job it can't serve internally. The venture gets a running start most startups never have.

What does jobs to be done look like in practice?

Elanco is a global animal health company. Through customer and market work, the new job was clear: livestock producers needed to turn verified emissions reductions into income, and food companies needed credible credits to hit their climate targets. Real job, real demand, real willingness to pay.

Elanco's organization was structurally blind to it. Every instinct was tuned to selling animal health products through an existing channel to existing customers. A carbon marketplace served a different customer and ran on different economics. It was a business model innovation the parent was structurally unfit to run, and it required something it could never provide from inside: neutrality. An industry marketplace owned outright by one animal health company is not a marketplace anyone else will join.

So we co-created Athian as its own company. It launched the first livestock carbon insetting marketplace and began paying farmers for verified reductions. The job Elanco discovered but could not serve internally is now being served by a venture built specifically to serve it, with Elanco's expertise and standing behind it. That is the whole pattern in one company.

The job you can already see

If you've used jobs to be done well, you probably already know a job your organization can't serve. Most leadership teams do. It sits in a deck, or in someone's head, waiting for a structure that will never form around it inside the current company.

The framework did its work. It found the job. The next decision isn't more discovery. It's where the job gets built. Corporates can't serve new jobs from inside the structure that defends the old ones. Startups can. Co-creation is how you get the founder's freedom and the corporation's advantages working on the same job at the same time, which is the only combination that reliably gets a new job served at scale.