How Jobs to be Done help you to segment based on causality, not correlation

Evelien Al
on
December 8, 2017

One of the ways in which Jobs to be Done theory has helped me a lot is by changing my focus from correlation* *to causality. As far as that’s possible with qualitative, often after-the-fact (of using, buying a product) research. Jobs to be Done focusses on why people chose to ‘hire’ (aka use or buy) a product or service. When interviewing you try to find out: what drove them to that particular hire? What small factors all contributed?

The job to be done includes someone’s goals, and the circumstances under which the person tries to achieve those goals. This will determine what someone ends up using.

Why marketing personas don’t always work

In marketing the focus is often on correlation*: *who’s more likely to use a product? Male vs female. City vs countryside. Age. Smartphone usage.

Useful if you know your target group, and understand their reasons for buying your product. But many startups and innovation teams are not at this stage yet. They first have to build an understanding of their target group and needs.

What I’ve seen with many innovation teams though is that they take this marketing-approach to segmenting. ‘Millennials’ must be the term we’ve heard most often when discussing segments. “But that’s not a segment, that’s a demographic!”

When starting out with a business you want to learn. You want to learn about the market you’re in and your potential buyers. To do this you’re going to do qualitative research in the form of interviews and observation, and run online experiments like landing pages. But if you’re targeting a group called ‘millennials’ you will never be able to draw conclusions from your experiments, except for “our target group was too broad”. Even if your experiment ends up being a success, you won’t know who was interested in your proposition, and why. It’s very hard to take next steps based on these kind of results.

It’s because in the group “millennials” there’s such a diversity of people. People with many many different goals, different circumstances, pains, gains. The reasons why they choose one product over another differs wildly from person to person. Even a hundred interviews would not be enough to draw good conclusions, and trust me, as a startup you won’t have the resources for that.

That’s one of the reasons why during customer development workshops we teach teams to start by segmenting narrowly, more than you’re comfortable with. “Millennials with full-time jobs” doesn’t cut it, it’s still too broad. “Millennials having a family, living in the city and working full-time jobs” neither. “Full-time working millennials living in Amsterdam, who had their first child the last year” is getting close.

Segmenting for startups & innovation teams

So how to start segmenting then? Take the classical marketing-approach, where you segment people based on ‘classic’ demographics like age, gender, and vocation?

The key is to focus on causality. If you’re doing your explorative research focus on what made people use or buy certain products. Their goals and circumstances, in other words: their jobs to be done. Then starting basing your target group (demographics) on this*. *For example: when you are making an app to get people into investing, you might discover that having studied a certain degree like economics or finance *does *make a difference in someone’s likelihood to start investing. More so than gender or age.

You can start broad, observe some patterns, and zoom in from there. One of the great benefits from using jobs to be done is that you focus on the real causes of hiring decisions, by focussing on just goals, and how someone’s circumstances influence how that person is trying to reach her goals. In that sense it’s an abstract approach, but that doesn’t mean you cannot connect it to a persona with demographics and characteristics. Just make sure your persona is grounded in characteristics that actually make her more likely to use your product, not what you’d expect a persona to look like.