ABOUT LISA: LinkedIn | SYSTM | Maven

ABOUT ADIKTEEV: Website | LinkedIn  | Retention Club 

ABOUT ROCKETSHIP HQ: Website | LinkedIn  | Twitter | YouTube


KEY HIGHLIGHTS

📌 Personalize retention strategies to align with specific user needs, avoiding generic approaches.

🔊 Implement A/B testing to determine the most effective retention tactics, such as targeted discounts.

♣️ Collect and analyze data using surveys and AI tools to understand customer departure reasons.

🪐 Segment users by behavior and geographical location to refine marketing strategies.

🌊 Identify whether issues are related to user retention or activation to address the correct problems.

🖥 Actively gather feedback, including during cancellation processes, to extract valuable insights.

🗡 Engage with users regularly to maintain brand presence and swiftly resolve any arising issues.

🛒 Avoid overused communication tactics like “we miss you” messages, opting for relevant and meaningful interactions.

FULL TRANSCRIPT BELOW

I’m excited to welcome Lisa Kennelly back to the Mobile User Acquisition Show. Lisa, welcome back.

LISA 

Thanks so much for having me back.

SHAMANTH

I’m excited to have you because last time we spoke, you described a very unconventional and interesting playbook. You described strategies and tactics that I think few people have tried.

I encourage folks to check out your past episode to find out more. Today, we are tackling a very different topic and theme. Based on our pre-recording call, you’ve shared notes for today’s call, and I’m very impressed. I’m excited to dive in today.

LISA 

Thanks. 

SHAMANTH

There is a lot of generic advice about retention. “This is finally the year of retention.”  “Retention is the new acquisition and involves sending emails on specific days”  and sending PMs. 

There’s a lot of advice that is relevant, but when you look closely, it’s just generic. You shared last time that some of the more powerful opportunities are specific to a product rather than something that can be generalized.

So, talk to me about this and perhaps give me an example from your work so it’s clearer.

LISA 

I get why people want generic advice. You want to be able to implement it across a bunch of different businesses. If it needs to be specific to your product, then you have to go out and do the work.

But if you’re trying to get people to start using your product again after they’ve stopped for whatever reason, hopefully, you’ve uncovered why. The most helpful thing you can do is remind yourself of the context in which they used you in the first place.

There are definitely retention strategies that revolve around broadening the use case because they’ve tapped out one.

But don’t forget the first time they used you, and remind them to use you that way again, assuming you delivered value on that the first time. For example, at Klarna, we have a team working on retention insurance. They ran a test targeting people who last purchased five months ago rather than just being like hey, remember we exist or Hey, come back to Klarna; they ran an AB test. 

The test was: Here’s the place where you bought from Klarna the last time five months ago. If you shop there again, you’ll get 15 percent off. We know they’re used to purchasing from this merchant. If we give them a little incentive, will they remember to do it again or stop at the same place?

Remember you bought from H&M last time. Wasn’t that great? You can also use Klarna at these other different merchants. For both of those, it was testing, like, is it incentivizing them at the same place, or is it letting them know they can use the same habit differently? The second one performed better, which is also a good reminder that you don’t always need to use incentives to get people to come back, which is another tactic a lot of people use, like throwing money at the problem.

If you have the money, you can earn it back. That’s great, but it’s not always the solution.

SHAMANTH

That’s an excellent point about incentives because they seem to be a very easy and convenient fix to many retention problems. Get 50 coins or get 10 percent off, for example, seems to be an easy fix, but you’re right that it isn’t always a lasting solution.

That’s a great call out as well. In order to understand even what product-specific experiments to run, you have to have a very nuanced level of insight about the users. Even just to formulate the two hypotheses that you talked about, you would have to understand very clearly which kinds of hypotheses you could actually test.

What are some of the ways that you and your team get these insights that folks could look at and learn from?

LISA 

There are a couple of different ways. One is that we run an ongoing survey of people. If you haven’t purchased in a certain amount of time, like I said, it’s around five months, and it’s long enough to say, okay, this person is lapsing.

.

We ask them why. And we aggregate those insights. Those are on automatically. Qualitative customer conversations are also amazing at every single point of the funnel. And I use this all the time.

SHAMANTH

It definitely makes sense to talk to customers. When we talked last, you shared how important it is to categorize or segment users because, as you said, someone who has churned would give you completely different insights from somebody who’s currently a very heavy user.

What are some of the ways you recommend categorizing or segmenting users to get these insights?

LISA 

There are a couple of different ways. At Klarna, when we looked at that churn survey and saw the reasons people churned, we found quite a few differences depending on the market.

So, is this consistent across all markets? For example, in the UK and the US, which are slightly newer markets, more people are not retained because they need help understanding the process.

In other markets, Germany and Sweden, they do understand it, but it’s, you know, they didn’t know there were use cases, but on their initial one, or they have concerns about customer service, right?

Some of it’s segmenting along. Are there market differences you need to understand? Then, you can create messaging or campaigns to address those concerns before the other customers who are in those buckets churn. But I mean, you can also look at these people who were active for a long time and then stopped. So they’re at risk of churning. Are they the people who have already totally quit? They’re also going to have different insights that you need to address and understand why.

SHAMANTH

Certainly, geo-specific segmentation and the other criteria you describe can be hugely important. Let’s say somebody’s listening to this interview, and they’re like, yeah, we want to go out and talk to my customers; depending on the size of the companies, that could mean very different things. How would you recommend that this customer research or insight process be different depending on the size of the company that someone is at?

LISA 

If you have the resources to set up an automatic survey, that’s one piece. The hardest part is analyzing the data you get. So it’s great if you’ve got this thing going around in the background, but it will add no value for you if no one’s looking at it.

This is where AI tools are great because you can combine and analyze that data and get a report on it. You’re able to send this out automatically, and you’re collecting some data. You’re taking the insights, and we’re feeding them back into ourselves, right?

It used to be harder, but now, some tools make it easier. One tool that I know several companies are using and that we’re looking at is Outset AI. I don’t know the price point, but that’s one of the similar tools that is looking at how to take all this customer research and extract insights from it.

There are also cancellation steps. It’s like a win-back approach where you try to cancel a product, and they’re like, are you sure you don’t want to deal? Are you sure you don’t want to deal? You know, they do all the things, and when they’re really, really sure, they’re like, no, they can’t win you back.

They have a mandatory question. Why are you canceling? And you cannot cancel until you answer that question. And, like, in a worst-case scenario, you lost the customer, but you got an insight out of it. I feel like ClassPass does that. Headspace has done that before. That’s another tool in your cancellation flow where you are collecting that insight.

SHAMANTH

Yes. The sort of lapsed ones are people who used to be active, but they stopped. If you look at your data and say, here’s this segment of people, we can see that we’re getting value, they aren’t anymore; those people are going to be. Hopefully, they really loved you at some point, and hopefully, they’re willing if you reach out, whether it’s from even the founder level, like, Hey, I saw you use our product, and you stopped. I’d really like to understand why.

I think you’d have a decent response rate on those. Even if it’s just a segment, like let’s pick five of these and have somebody pretty high up in the company do the outreach.

AI makes a lot of this much easier. In the past, a lot of this was very painstaking, and now, again, a lot of my work’s not on the retention side, but something we’ve just done is take a bunch of user reviews. Upload them to GPT and just say What are the big messages? Talking points that we can use in ads on the acquisition side. This was something that was always on our list in the past. We’re like, we should do this. We should do it, but who’s going to review each of these reviews? You’re right that AI just makes it so much easier.

LISA 

Exactly. Another thing you can do with some things is look at your customer support tickets and the things people are writing into customer support, what they are complaining about, aggregating that in the same way, and analyzing what are the things that people are upset about the most.

What are their continued frustrations? And are we really addressing those?

SHAMANTH

Certainly. It’s not just that you have to create something new to collect this; you are probably collecting all of this information, and you don’t know it yet.

LISA 

Yes, and now you have the tools to process them.

SHAMANTH

If nothing else, just upload everything to Chat GPT and ask it to analyze. Certainly, you could use more specialized tools like Outset and the stuff you mentioned. Certainly, you could do those as well. In terms of running these analyses and user research programs like these, what kind of cadence or frequency would you typically recommend?

LISA 

I believe you should talk to your customers on a weekly basis. How often you speak to them about the retention piece of the funnel or later on in the journey depends on how critical it is for your business right now.

So if your retention or return is not really hurting you so bad or is at the normal level expected, then maybe that’s not what you prioritize. So that kind of depends on your business. Otherwise, I just think you’re keeping an eye on it monthly. If the numbers are fluctuating a lot, then we need to maybe dig into the ongoing insights we’re getting. Do we need to do a deeper dive? Do we need to look at, like, we’re seeing a spike, and why is that happening?

SHAMANTH

You’re right, depending on how serious that issue is. I think that can govern your frequency. In our pre-recording briefing, you talked about instances where companies think that they have a retention problem, but they actually have an activation problem and not a retention problem.

Can you talk about this and elaborate on what you mean by that?

LISA

Absolutely. I think this is actually the case most of the time, or a lot of the time when people never get activated in the first place. And my colleague Matt Lerner has a great Twitter thread on this, which is how they tackled sharing at PayPal. It’s basically, if you go on a first date with someone and then they ghost you, that’s not a divorce. That’s an activation problem. So that’s the same thing. Like these people, they’re not divorcing you. It’s like you never really clicked in the first place. So, if people use your product one time, that’s not a retention issue. That’s like activation, that’s onboarding.

Depending on the product, of course, it’s not one; it’s probably several times, and it’s until they get habituated. If you’re not getting to that step in the first place, your retention is less of a problem than your actual activation and people seeing the value you’re providing.

So that’s why the segmentation: Let’s separate the people who are not activated. Do we need to fix things there? And then, okay, these are the people who are activated. Are we retaining them effectively? Why or why not? Those would be totally different approaches to fixing those things.

SHAMANTH

Going back to where we started in terms of one-size-fits-all solutions, some classic examples would be “we miss you.” Everyone’s seen that, and a lot of us assume that. It’s ubiquitous because it works. And you shared that that may not always be optimal, if only because people are kind of burnt out by it.

Can you elaborate on why that’s not effective and what you would recommend that people do instead?

LISA

It probably worked back when push notifications were first introduced. People probably thought it was novel, but the problem with that, I think, is that it’s supply-side thinking. So it’s basically inside out. You’re thinking from the perspective of your company, which misses the person. So they’re like, hey, and the customer does not care. Your end user is like, Yeah, who cares if you miss me?

I signed up for you once, but it didn’t work, or I forgot something. You, as a company, missing the customer doesn’t do anything, doesn’t solve the problem, and doesn’t address any of the reasons they want to use you. So, if you know the reason they actually signed up in the first place, or if you know the reason they stopped using you, then communicate that.

Give them a good reason to use you rather than just you, like whining and banging on their door. One example we’re testing now is that in the U.S. and a lot of other countries, it’s tax refund season, you know, so people may have a little more extra money in their wallets so they can come and Buy something with Klarna because we help you buy things.

We’ve been testing messaging that is contextually relevant. Hey, it’s tax refund season. Check out what we have to buy. And that’s because if you have a little extra money to go shopping, you might think about how you want to pay for it, as opposed to just, hey, remember, we exist.

So that’s general. Think about it from the following: What’s the message? If you were the customer that you would receive, that would make you think, Oh yeah, that was a great experience using it before I should try it again.

SHAMANTH

I think your example about the first date would also just be very relevant in this context because you don’t want to send them a, I miss you after the first date. It would be a little bit weird. Yeah. Right. You can maybe say I miss you if it’s someone who has been using it for a long time and then stopped. Like, there’s a time and a place for it if it’s contextually relevant, but I see it used pretty broadly.

Lisa, this has been incredibly insightful. Thank you for sharing natural examples of natural strategies. Much like last time, I really appreciate how concrete a lot of your examples and recommendations are. This is perhaps a good place for us to wrap. But before we do that, can you tell folks how they can find out more about you and everything you do?

LISA

Thanks again for having me. It was really fun to do the research for this and check in with our retention team and various people. Thank you for giving me an excuse to brush up on my retention tactics. You can find me on LinkedIn.

I am also a coach in a program called SYSTM, so you can check that out. Sometimes, we also teach courses on maven.com, so you can find me there as well.

SHAMANTH

Thank you so much for being on the show. 

LISA 

Thanks again.