Ada’s playbook for finding Product-Market Fit

← Back to homepage

I listened to this podcast 5 times, extracted critical insights and structured them into a 16-step playbook to find PMF.

Watch the whole interview here→

The 16 steps to find PMF based on ADA’s journey

1- While working on improving Volley, they hit a constraint while growing customer support

2- The constraint made it difficult to deliver on a key company value: great customer care

3- A fork in the road: decide to follow the traditional customer care playbook by others or invent

4- Had a contrarian insight and felt a missed opportunity, so they followed their curiosity

5- Research mode: talked to other managers to understand their behaviour

6- Validated existence of paradigm and uncovered first-order limitations that lead VPs to bad strategies

7- Researched the problem deeply to uncover root-cause conditions for those behaviours

8- Identified that AI technology has changed conditions and opened up opportunity

9- Defined the operational-problem to solve that will unlock the opportunity

10- Inject themselves into the current solution and experience the pain personally

11- Executed manually while mapping the workflows to automate

12- Uncovered deeper mechanics, insights and constraints

13- Measured results and prove the impact of their manual solution

14- Built software that replicated their manual solution, and made it easy to use

15- Replaced manual work with their software

16- Packaged it, messaged it and created new customers

-

1- While working on improving Volley, they hit a constraint while growing customer support

"David and I built this product on iOS (Volley) and on the web that was growing really quickly at one point quickly out of the gate and we felt like we had sort of lightning in a bottle. It was very exciting. We encountered this customer service problem pretty early on. The problem was how do we scale our customer service operations in line with our user growth."

2- The constraint made it difficult to deliver on a key company value: great customer care

"Everything valuable had come from a really tight feedback loop between our users and our product engineering teams. We knew our customers' names, we had a relationship with them. We went from treating our customers as people to then treating them as numbers that we were trying to keep at bay."

3- A fork in the road: decide to follow the traditional customer care playbook by others or invent

"It was pretty sad to see us sort of pursue the conventional customer service playbook, which is essentially figure out how to talk to your customers less the bigger you get. That almost every other business in the world was running."

4- Had a contrarian insight and felt a missed opportunity, so they followed their curiosity

"When we saw that, I became pretty curious about why it was that we were investing in figuring out ways of talking to our customers less. Because it struck me that the bigger we got, the more valuable customer data would become to us. In other words, this whole paradigm felt backwards."

5- Research mode: talked to other managers to understand their behaviour

"So I picked up the phone and called as many VPs of customer experience as I could find. David and I interviewed a bunch of people and we asked them "Why is it that you're speaking to your customers less and not more as you're growing?"

6- Validated existence of paradigm and uncovered reasons that lead managers to bad strategies

"Virtually everyone agreed with the premise that it was sort of backwards to be pursuing the strategy. We first confirmed that there was a pattern, that the same problem was manifesting across industries.

They also echoed each other in saying that customer service is a cost center. This is how we are structured, this is how we've designed things, this is my job. And while I know more about the customer than anyone else in the company, we can't afford to increase our level of interaction.

Having a 10th phone call in a row and people are saying almost the exact same phrases when they're describing the problem and they're reacting to it the same way. You almost feel like an anthropologist - you just sort of get to a point where you can kind of predict what someone's going to say in response to one of your questions."

7- Researched the problem deeply to uncover root-cause conditions for those behaviors

"Understanding of that problem became the sort of groundwork. It's important to understand it really intimately, like really really deeply.

Why do you want to minimize the amount of customer interaction? Because it's more profitable. And the customers reaching out to you too much might mean that your product's not working well, you should just make the product more user friendly.

But the paradigm of talking to your customers less was really born of an era where people really weren't making as many data-informed decisions; because a lot of this information wasn't digitized to begin with; it was very difficult to mine that data, extract insights and inform your decision-making."

8- Identified that AI technology has changed conditions and opened up opportunity

"However, in the digital era where we can actually start to make use of this information. You can do more with it, it becomes much more actionable, much more useful. And so I felt as though the technology was actually quite a bit ahead of the strategy and that seemed like an opportunity for us."

9- Defined the operational-problem to solve that will unlock the opportunity

"We knew our success had come from really rapid feedback loops between our users and our product teams. The question was why - how do we, if that works at a small scale, it should work at least as well at a big scale, maybe even better if we had more data.

So how do we figure out how to solve that problem?"

10- Injected themselves into the current solution and experienced the pain personally

"Our belief was we've identified there's this problem across industries, now let's understand it more deeply. The way to understand it more deeply in our view was to experience the pain firsthand and as manually as possible.

We committed to understanding this problem at a much deeper level. That commitment meant going back to these VPs and asking if we could join their team as agents and become frontline customer service reps - and seven of them said yes.

We were learning how difficult and painful it can be to provide to scale customer service from the ground up as a frontline employee. We were doing manual customer support and investing in trying to learn as much as possible through that experience."

11- Executed manually while mapping the workflows to automate

"We really fought the temptation to write a single line of code. I think there's a lot of founders out there who think that they measure progress in code, to build an MVP. I think core to our success was actually a rejection of that premise.

Our goal was actually to go as long as possible without building an MVP. the reason for that is because we wanted to understand the manual processes that our software would eventually automate better than anyone else. This puts you in a position to build software that is truly excellent."

12- Uncovered deeper mechanics, insights and constraints

"We learned three core things in the course of being agents:

1. 30% or more of the inquiries we were responding to were repetitive and mundane - in some cases was upward of 80%
2. The agent experience that we had was like far from ideal. It was super negative. No one was waking up out of bed in the morning going "I'm so excited to spend more time in this Enterprise customer service software" - it was just a painful product experience.
3. They wanted to talk to their customers like they talked to their friends. Why can't I message this person, why can't I text them? And this we heard over and over again and that idea was always rejected, because of the conventional customer service paradigm which is reduce customer contact - messaging was thought of as likely increasing it.

13- Measure their results and proved the impact of their manual solution

"We became amongst the most productive agents on each one of these teams and we essentially - we just worked really hard. And we saw the impact of that work:

1. We saw that customers were less likely to leave
2. We learned that our colleagues liked their jobs more and were far less likely to leave when we were handling as many of the repetitive inquiries as possible.
1. That was a big deal because we attrition rates in contact centers = 45% annually
2. If you remove the repetitiveness of the job, what you're left with is something that's way more intellectually stimulating - people just loved working with us as a result.
3. And the third thing we witnessed was that the data that we were privy to, these customer conversations, they felt like a gold mine.

We'd proven the value of the software before it existed - proved it manually - we still to this day at Ada say "do it manually first" when you're solving a problem."

14- Built software that replicated their manual solution, and made it easy to use

"That it was then that we set out to build the software that replicated that manual behavior we had engaged in.

And the solution became - it sort of appeared right, it was so painful that we were going through that the software almost became medicine - like take the pain away. And we took an ML approach because we had access to so much data, and then we focused on making those ML techniques that we were developing as easy to use as possible because all our colleagues were non-technical."

15- Replaced manual work with your software

"The reason we knew Ada worked in the early days or was going to work is because we didn't get fired. We ran this software and our managers, they didn't care - they love the fact that response times went down, they couldn't tell the difference between our responses and our software's responses.

Pablo: You substituted yourself with software while still working as agents?

Mike: That's right. It's a classic experiment - we held one variable fixed, we manipulated another and we saw that the results were the same and we knew that the software was validated."

16- Package it, message it and created new customers

"The next problem to solve is how do we get people to recognize - how do we identify people who also have this problem and help them recognize that we can potentially solve it for them.

And that we did through initially I'd say two things: one was a lot of outreach to friends and friends of friends companies, and the other was through a lot of cold emailing and messaging iteration.

Our friends all responded to our email and they said "I'd be happy to introduce you to our head of customer service". They didn't give us a deal but they gave us a bat, an opportunity to get in front of the decision makers within their companies.

And that was huge - it was the learning from those that really informed the second thing, the messaging which allowed us to acquire customers that we didn't have any connection to."

—

Created by Javier Rincon (Fractional Head of Product →)