Jobs to Be Done (JTBD)
"People hire products, they don't buy them, they hire them to make progress in their life." - Bob Moesta
What It Is
Jobs to Be Done is a framework for understanding customer behavior by focusing on the progress people are trying to make in their lives, rather than the products they're buying. Co-created by Bob Moesta and Clay Christensen, it flips the traditional product-centric view to a demand-side, customer-centric perspective.
The key insight is that supply and demand are not as connected as most people think. Most people believe creating a product creates demand, but actually a struggling moment causes demand. Products don't have jobs—people have jobs, and they "hire" products to help them make progress.
The biggest misconception about JTBD is thinking it's about "pain and gain." It's actually about context and outcome. The context someone is in determines what they value and what progress they're trying to make. Context makes the irrational rational—when a decision seems illogical, you likely don't have the full story of the context that drove it.
How It Works
The Core Concept: Competing Against Non-Obvious Alternatives
Consider Snickers vs. Milky Way—both candy bars with similar ingredients, sold in the same aisle. Traditional competitive analysis would benchmark features: calories, texture, price.
But when you talk to people about when they actually eat each:
Snickers is typically eaten when someone:
- Missed their last meal
- Has a lot of work to do
- Is running out of energy
- Wants to get back to tasks quickly
The "job" is meal replacement—Snickers competes with protein drinks, Red Bull, coffee, and sandwiches.
Milky Way is typically eaten when someone:
- Just had an emotional experience (positive or negative)
- Is alone
- Needs time to regroup
The "job" is emotional comfort—Milky Way competes with a glass of wine, a brownie, or even a run.
Struggling Moments as the Seed of Innovation
Struggling moments exist before products do. They are the real source of demand and should drive your roadmap—not features.
Example: Southern New Hampshire University found 50-60 "anomalies"—students paying full price but never coming to class, watching everything online. Rather than dismissing them, they studied them and discovered:
- They were older than typical students
- They had tried college before and it didn't work
- They had new responsibilities requiring them to change
This insight led to an online program now serving over 200,000 students—one of the largest universities in the world.
The Progress Vector
Value is not just about the outcome—it's about where you start AND where you end up. The same outcome is valued differently depending on the starting point.
If someone starts here → and ends here, they value the progress X amount. If someone starts lower → and ends at the same point, they value it much more.
This is why some customers want a "price discount" when you're giving them more than they need—you're overshooting their progress vector.
How to Apply It
Step 1: Find People Who Made the Switch
Don't talk to people who "want to" do something—talk to people who actually did it. Study behavior change, not intentions.
- If you have a product: Find 10 people who recently bought it
- If building something new: Study what people would "fire" when they hire you (e.g., study Craigslist/eBay/Etsy users if building a marketplace)
Step 2: Interview for the Story, Not Opinions
The method is based on criminal and intelligence interrogation, not traditional market research. You're not asking what people want—you're reconstructing what happened.
Ask about:
- What was going on in their life when they decided to make a change?
- What were they hoping for?
- What were they worried about?
- What did they have to give up?
- How did they convince others?
Step 3: Cluster, Don't Segment
Instead of looking for themes across all interviews, cluster the patterns. Find the "pathways"—sets of pushes that work with specific pulls.
Most products are hired for 3-5 different jobs, often in conflict:
- One person wants speed
- Another wants thoroughness
Understanding these conflicts helps you decide what to build, what to turn off, and how to position.
Sample Size
From a causal mechanism and set theory perspective, patterns start repeating around 7-8 interviews. Do 10-12 max. Two rounds of 12 is better than one round of 24.
When to Use It
Best for:
- Zero-to-one product development
- Understanding why customers switch
- Identifying true competitive alternatives
- Building roadmaps based on struggling moments
- Reducing churn by understanding what job your product isn't fulfilling
Not ideal for:
- Markets with no real customer choice (e.g., employer-provided health insurance)
- Deeply habitual behaviors where people can't recall decisions (e.g., buying gum)
- When companies want to "find jobs" that validate their existing product assumptions
Common Mistakes
- Doing it in a conference room - You must talk to real customers who made real decisions
- Focusing only on outcomes - Context is equally important; it's context AND outcome together
- Treating it as hypothesis testing - JTBD is hypothesis building research; you don't know the answer going in
- Following best customers - They'll take you "upmarket" and destroy why others hire you
- Adding more features - More features create anxiety; sometimes reducing friction beats adding features
Source
- Guest: Bob Moesta
- Episode: "How to find work you love | Bob Moesta (Jobs-to-be-Done co-creator, author of "Job Moves")"
- Key Discussion: (00:05:07) - Introduction to JTBD and the Snickers/Milky Way example
- YouTube: Watch on YouTube
Related Frameworks
- Four Forces of Progress - The forces that drive switching behavior
- Six Phases of Buying - The timeline of how customers buy
- Layers of Language - Getting past surface-level customer responses
Further Reading
- Demand Side Sales by Bob Moesta - The tactical method for applying JTBD
- Competing Against Luck by Clay Christensen et al. - The theory behind JTBD
- Never Split The Difference by Chris Voss - Interviewing techniques that complement JTBD