Discovering JTBDs | Michael Harrower

The scene

Michael Harrower

Head of Product Design · Disciple

Disciple was originally an agency that built bespoke apps for musicians to create memberships for their fanbases. In 2019 it began the transition towards off-the-shelf SaaS and by the end of the year it had collected customers of many types, such as religious groups, a large motorhome company, photography teachers, wellness coaches, charities, among ~380 others.

The team had some sense of why customers had chosen to join Disciple and why they stayed, but it hadn't been analysed properly. What we needed was to elevate the product experience by reminding the customer of their JTBD(s) and designing a product experience that aligned with why they joined.

The problem

The Disciple product boasted a long list of features, but it suffered from a lack of direction.

Without deep understanding of why our best customers chose Disciple, it was impossible to double down on this perceived value and weave it into the product. What were they doing before they discovered us? What was their value prop to their members? What did they see in Disciple that made them think it would be the catalyst for them to transform into who they wanted to be?

380

Active customers

Unclear

of their journey before Disciple

The work

I led a three-person team (myself, the Head of Growth and a Customer Success Manager) on what became the most impactful project of my time at Disciple.

Define successful customers

Interview 10 successful customers

Find trends between interviewees

Document findings

Create a survey

Send to successful customers

Validate interview findings

Create JTBD(s) definition

Phase 01

Defining Success

We started by defining what a 'successful' customer looked like. After debate, we landed on three criteria: 48%+ contributor rate, 60%+ retention rate, no payment issues in six months. This led to a list of customers that CS backed as being successful.

Example of the customer attributes used to define "success" (in Custify).

While the survey was running, I also looked at churn data to try to understand how our potential ICP stood up against the rest of the customer base when it came to retention.

When looking at all customers, there was a significant drop in retention at 12 months – most of our customers pay annually, so this meant they were churning before their first renewal. By the time the second renewal arrives, we're left with ~40% of customers ready to pay for their third year upfront.

I then looked at churn rates in the categories that spanned the "successful" customer group. The result was emphatic. These customers basically always decided to stay at their second renewal.

Lastly, I looked at the percentage of customers that were in this successful group vs the rest of the customer base.

Phase 02

Jobs-to-be-Done Interviews

The goal was to interview at least 10 customers in the customer group identified as successful. All three of us attended each call, so we could debate observations and try to reduce each other's biases.

When deciding the script questions, we kept in mind that understanding what the value of Disciple was in the moments they discovered and subscribed to Disciple instead of the competition would lead us to JTBD understanding that we could evangelise to the rest of the business and use to sharpen the product roadmap.

The interview script sections.

The specific questions for the section "Understanding the user's journey". This section gave us insights into what tools and setup the Disciple product seemed appropriate to replace, and whether they were making money out of the offering before they subscribed.

The specific questions for the section "Community value proposition". This section sought to understand the essence of why their community/business became successful.

Some of the raw notes for each participant.

Phase 03

Trends and observations

Collaborating on a sheet intended to mark trends in observations, the three of us debated each point. An observation would need the three of us to be in agreement before it could be marked as true for any given participant, to reduce the risk of false positives.

After each call, we added observations to this sheet. We marked against each participant that the observation was true for.

Once the interviews were complete, we were left with a shortlist of observations that were true for at least 50% of the interviewees.

We made a sheet of primary findings, summarising what we thought we'd learned beyond the observation trends.

Phase 04

Surveying to validate interview findings

We had a pretty strong hypothesis about our most successful customer's background and journey when choosing Disciple, so the next step was to validate it by asking a much larger group of around 50 customers in the original "successful customers" list.

The questions followed a similar structure to the interview script, with additions where we wanted to probe whether we were on the right track. For example, questions focusing on job-to-be-done used a multiple choice format with the choices being based on our observations and findings from the interviews.

An example of results for a survey question. This was one of the most important questions in the survey, because the answers comprised the JTBDs we thought we might have discovered in the interview process.

The survey results largely confirmed our learnings from the interviews, adding confidence to what we thought we had discovered.

Phase 05

Jobs-to-be-Done

Five jobs-to-be-done emerged from the research, validated by the survey. These became the lens through which the product team evaluated every roadmap decision going forward.

The JTBDs we settled on.

It was clear to the research team that we had discovered a customer group that were more successful and more reliable than the rest. I held a strong opinion that we should focus the product roadmap for the next few months on delivering improvements for them specifically, to see if we could increase the number of them that were subscribing. To make this work, the whole company needed to understand the journey we'd been on to get to this point, so they could internalise it for themselves and help contribute to shifting us in the right direction. Here is some of the material I used to describe ICP to team members who weren't involved.

Rather than using the typical user profile format, we opted to explain specific details about their firmographics and the content offering they presented to their members, along with examples of existing customers that we considered ICP. This was appropriate because the team were already very close to our customers, so it made more sense to reference examples they understood well than create fictional proxies.

We also presented some clear trends we thought were important.

It took meetings with each department independently to get the wider team onboard with the research findings enough to confidently use them in their own work. This part of the process was absolutely crucial to the success of the project. Without the whole team sharing the same beliefs, the value from the project would have been very limited.

The outcome

5

Jobs-to-be-Done established

3

ICP segments defined

A clear set of JTBDs gave the whole team motivation in understanding how our best customers were operating when they discovered Disciple. This informed the roadmap, the way we spoke to customers in marketing and the way we onboarded them.

What I learned

Research conducted together is better.

Interviewing with 3 colleagues together showed me how your memory and biases can deceive you in qualitative research. Debating the observations led to a higher level of diligence that was extremely beneficial for the project.

Evangelising your learnings takes repetition.

Being directly involved and learning about the project afterwards are two very different things. To get the wider team to adopt our findings, it took repetition and multiple smaller group sessions.

Next chapter

Improving the IA of a B2B dashboard

The dashboard had a very large number of features with very little thought applied to architecture or pattern design.

Read it →