Transcript: One Metric to Align the Org: How AppsFlyer Turned Product Adoption into Retention

Rachel Frenkel BenHanoch: [00:00] And then I said, we have to look at the same goal and have all the wheels pointed forward, and that's exactly what we did with CS — or with CS and with marketing and with product — when we looked at this Product Adoption. And let me take one step back and say. Beyond the fact that we started with one small group — obviously, any move inside an organization has to start with a small group that's your design partner, both for the process and for the results and to show that it works; only then can you of course roll it out to the rest of the product groups. But CS people came in too, and they were incredibly excited that we showed them we'd be able to create early indicators. It's really, really crazy — let's talk for a moment about how this Product Adoption worked. I sat down with the product manager and the analyst of each product organization, and we said: pull me data that reflects depth of usage — not just how many times they log in, but which features they activate, what configuration they set up, which persona within... ...the customer account is using this thing. All of these things are very indicative of depth of usage and of the value they extract, and we built this compound KPI that's made up of weights for a series of actions the account performs inside the product. And we got some value from 0 to 1 — like 0.2 and so on. Now, when we looked at the customers, at this value of theirs over time, we saw fluctuations, but it was still hard to understand from those fluctuations what they meant — if a customer went from 0.4 to 0.42. Is that significant enough? Is it meaningful, do we need to address it, do we need to do something about it — we didn't know. So we decided to create buckets, to look at the picture a bit from above, a bit from afar. We said, let's split this into three buckets: a low bucket — low, or low usage — like they touched it but didn't use it, they logged in but didn't do anything with it. Medium — they engage, that's something — and those who are fully adopted, well, highly adopted. What is highly adopted? They use the advanced features, they get into it, they extract value from it, they pull reports, and so on. Now, what I asked the analyst for — and this, I think... ...is essential, and a bit of a game changer in the value this thing provides — we asked for there to be an adoption range. Say, in one product it was from 74 to 100, in another product it was from 89 to 100; each was a different number, but one that would be correlated... ...to the customers' chance of retention. Looking at past data — and then the survivorship, how you can do that — but we looked at past data and said: customers we saw over time who were in this range of the Product Adoption value stayed with us, renewed, did renewals for us over the years. And then it was a game changer, because CS said, we get it — if we have a customer who isn't... In short, [the conversation] went backwards — let's explain this slowly, so they don't have to rush through it at speed. Correct me: this is basically what you did.

Eyal David: [02:49] You looked historically, according to the criterion that basically segmented into different buckets; you took certain users over a certain time frame who behaved appropriately, or in the way you want — right? You looked back at what they did and knew how to project that onto others. Exactly right. And you predicted — well, I wasn't there anymore at one point, but I can... I think its strength was on several fronts. One front: CS said, wow, now we understand our goal, we want...

Rachel Frenkel BenHanoch: [03:19] ...to make sure they reach a level of usage above this threshold, and then I know that someone is with us for the long term. And it also helped CS prioritize their time, because they say... ...those who are in the high spot, I'll reach out to them once a quarter; those who are on the threshold of moving up to high, let's push them to cross over to high; those who are in the low bucket, maybe we'll reinforce the goal for them more, because they have a higher chance of churn. That is, they knew how to prioritize how to work with their customers, because CS... ...a certain way of reaching all the customers, and otherwise I couldn't have moved them there, or said the right thing to them — to CS, right? So you also basically need — it's not that you work with the sales org, with the specific person — you need to go top-down again and explain why it's worth it, and suddenly you have a really good case in hand. Right. And you built a product.

Eyal David: [04:04] Right. I'll just close the loop on the two aspects. One aspect: the segmentation capability was very, very deep, meaning you could segment by...

Rachel Frenkel BenHanoch: [04:15] ...geography, industry, time zone, and so on and so forth — in order to find exactly the similar and different parameters between a customer who's low and a customer who's high, and to push with a very specific campaign that touches exactly the point where they need to improve in order to move up to high, and to convert them to high. And we did that — I mean, there were really jaw-dropping results. In one of our products, we took and looked at the cohort of this medium bucket and said, let's reach out to them, let's do... ...let's prepare very, very targeted marketing material, because we know where they can be better compared to those who already were. Through CS we set up meetings; CS sat down with them and explained the value to them. After a few such meetings, over a span of three months, sixty-eight percent of the customers who were approached with these precise marketing materials moved up to the high bucket — meaning we secured that they'd stay with us for the long term. It was a crazy success. Because this was, essentially, an episode from Product Builder — [if you] want to hear the...

Eyal David: [05:15] ...full conversation, search for the feed, and don't forget to follow us.