This is the 1st Episode of Subscription League, a podcast brought to you by Purchasely.
User retention is the lifeblood of every app business. But where do you start?
Andy Carvell, Partner and Co-founder of Phiture looks at this question from the product development point of view and talks about how the effort to increase retention and engagement has to be built-in in applications.
For noteworthy quotes and key takeaways from the episode, read the article - All about Retention with Andy Carvell.
Episode Topics at a Glance
- Why is retention the key indicator of success for a mobile app?
- What working on retention means, and how to do it in the early stages of app development
- Andy’s top tools to boost retention
- Friction or keeping it short for the onboarding process
- SoundCloud’s famous hard sign-up model - did it work and what’s the lesson?
- Examples of apps that help users to build a new habit
- Analytics Stack
- How to reduce churn
More about Andy Carvell
Andy has over 20 years of experience in the mobile industry. His company Phiture, is a Berlin-based mobile growth consultancy and agency co-founded in 2016 with Moritz Daan with over 110 employees. Phiture consults clients around the globe on mobile growth topics, helping them develop and implement holistic mobile growth strategies.
Prior to Phiture, Andy led the user retention team at SoundCloud in Berlin. Andy published the Mobile Growth Stack in 2015, widely adopted as an essential framework for mobile marketers, which he continues to develop.
Andy Carvell’s Links
[0:39] Introducing Andy and his background in mobile apps
[0:37] Andy’s creation at Nokia - Space Impact game
[2:50] Retention - the key indicator of success for a mobile app
[4:35] Working on retention from the early stages of app development
[6:07] How Phiture helps companies to grow their apps
[8:00] Andy’s top tools to boost retention
[11:28] Should you add friction to the onboarding process?
[13:30] Advantages of having a hard sign-up wall in the onboarding process
[17:06] Apps that help users to build a new habit
[19:15] The best tools for building new features in a subscription app
[23:15] Benefits of predictive analytics
[00:00:22.410] - Olivier Destrebecq
Welcome, everybody. Today, I have Nicolas Tissier, CEO of Purchasely, with me to interview Andy Carvell. How are you guys today?
[00:00:31.950] - Nicolas Tissier
[00:00:33.240] - Andy Carvell
Hi, Olivier. Hi, Nicolas. Great to be here. Thanks for having me on the show.
[00:00:36.840] - Olivier Destrebecq
Oh, you're welcome.
[00:00:37.490] - Nicolas Tissier
Yeah, it's very exciting.
[00:00:39.820] - Olivier Destrebecq
We decided to invite you, Andy, today because you have something like 20 years of experience in the mobile industry. You founded Phiture, well, co-founded, I should say, with Moritz Daan, Phiture in 2016. Ever since, you've been consulting with clients around the globe on mobile gross topics, which makes you more than qualified to talk with us today.
[00:01:00.000] - Olivier Destrebecq
But on top of that, before Phiture, you also worked at SoundCloud on the user retention team, so really a lot of things that you've done that will make you an expert today. A document that you've been maintaining... Actually, you'll be able to tell us when you started that effort on that document. But the Mobile Growth Stack, which is used by a lot of marketers today to improve their user acquisition, retention, and all that good stuff, you're maintaining and publishing that today.
[00:01:23.530] - Olivier Destrebecq
But what really came to heart for me and why I really wanted to talk to you today is because you were the developer of the classic arcade shooter Space Impact on the Nokia phone back in 2000 or something like that. Can you, in 30 seconds, tell us what it was like?
[00:01:37.910] - Andy Carvell
Yeah, sure. Space Impact was a game that I developed back at Nokia back in 1999. It was my first job out of university, and I was working at an R&D facility of Nokia in the UK. This was like working in the future. It was like this space-age technology lab full of engineers and high-tech gadgets. I got to work on mobile devices that weren't going to hit the market for 6-12 months. It was really a dream job.
[00:02:05.360] - Andy Carvell
It was a fun challenge squeezing an arcade, shoot them up with scrolling backgrounds, and aliens, and sound effects, and even vibration effects, like a lot of firsts on a mobile game and trying to get that into squeezing performance to do that. Also, squeezing it into 16 kilobytes for all the code and data was quite a nice optimisation challenge.
[00:02:25.470] - Olivier Destrebecq
I'm sure. Well, let's get back on the topic for us, which is more around subscription, acquisition, and all that good stuff, not necessarily on video games.
[00:02:31.670] - Andy Carvell
[00:02:31.960] - Olivier Destrebecq
First, is there anything you want to add to the intro that I made just 30 seconds ago?
[00:02:36.440] - Andy Carvell
No, Olivier, that was a very nice intro, very kind words there, and I think you captured the salient points. Thanks a lot.
[00:02:44.620] - Olivier Destrebecq
Awesome. Again, today, we're going to focus on retention with you. Can you tell us why is retention so important?
[00:02:51.390] - Andy Carvell
Yeah, sure. Retention, this is keeping users sticking around. It's essentially lots of different ways to measure different ways to look at retention. But ultimately, if we boil it down, it's just, what proportion of your user base is coming back to use the product again and hopefully again and again and again.
[00:03:11.890] - Andy Carvell
This is really, for me, the measure of the value of the product. Even before you're monetising it, maybe before you've even figured out your monetisation model, you really need to have users coming back over the long term. If they're coming back time and time again, it shows that they're getting some value out of the product. For me, it's a really key measure of product-market fit.
[00:03:33.120] - Andy Carvell
I think focusing on retention from the very early stage is a real measure of does your product provide value to at least some proportion of your user base? From that, if you can answer that question with a yes, I think you're on a really good track to build out a really killer product and to go for real growth.
[00:03:49.720] - Olivier Destrebecq
Yeah. If users don't stay, then your business model is an issue, I guess.
[00:03:53.050] - Andy Carvell
Exactly. That's the most succinct way to put it.
[00:03:56.190] - Nicolas Tissier
It's more important to focus on retention at the beginning that I'm actually trying to figure out what business model you could put in place.
[00:04:05.160] - Andy Carvell
I would argue yes, definitely no harm in coming to the table with an idea of how you're going to monetise. But in any case, in order to be in a position to do that in any sustainable way, you need to figure out retention and you need to get to some level of product-market fit, which is a whole topic by itself. But I find that retention is a pretty good proxy for that.
[00:04:26.620] - Olivier Destrebecq
You've done a lot of work on retention with clients through Phiture, what needs to be in place to be able to work on improving retention?
[00:04:34.000] - Andy Carvell
Great question, Olivier. As we just discussed, I think you can and need to work on retention right from the very start. However, I think at different phases of the product growth and different phases of a company growth, working on retention maybe means slightly different thingsso. In the early days, the main way you're going to work on retention is just by building a solid product and a solid foundation, and you need to also measure that retention.
[00:04:58.150] - Andy Carvell
You're going to need to have some basic analytics in there right from the start. I would argue, in this day and age with some great analytics platforms out there, there's not really any point in waiting until later to get a good foundation of analytics set up. It's going to just pay dividends all the way through. But yeah, get some analytics in there so you can at least start to see what your retention looks like.
[00:05:18.370] - Andy Carvell
But in the early days, you're not going to be able to run AB tests very effectively because you won't have the cohorts of users required to get to statistical significance on these tests. You're going to have to do a lot more just by good design, a bit of gut feel, which is not popular in data-driven circles, but it's essential when you're early on. You don't have so much data to play with. But what you can do to bring to the table some data, which is still very valuable, even if not quantitative, and that's user testing, user interviews, getting that qualitative understanding of how users are using your product.
[00:05:52.000] - Andy Carvell
You can even do that when it's still wireframes and you can do it while it's still in alpha and beta. You should be constantly getting that qualitative feedback because you're not going to be able to do the data-driven testing, which you can do later. Later on, the retention work where, for example, Phiture, my company would get involved, we typically don't work with companies pre-product-market fit, but we're working with later stage companies that already have some traction, already have decent retention. Then we're helping them to get it to that next level.
[00:06:20.500] - Andy Carvell
Then there's a whole different set of tools and techniques and methods which come into play, which makes sense more at scale. Things like CRM, particularly, so mobile customer engagement, things like push notifications, email marketing, mobile Internet messages, gamification, these kinds of things which are great boosters for retention, if you get these initiatives right and can build programs around iterating on them in a data-driven way. But they take time, they take resources, they take additional tools, and particularly, they need large cohorts to play with because some of the optin funnels for things like push notifications and email can be pretty steep.
[00:06:59.280] - Andy Carvell
If you just have 1,000 users, you're not going to get any significant results from any real signal that you can work with to optimise these programs.
[00:07:08.130] - Olivier Destrebecq
What would you say is, I don't want to say, minimum size, but what user base do you start getting good results on your AB test and all that other stuff?
[00:07:16.820] - Andy Carvell
Yeah. At Phiture, when we look at this from a customer qualification point of view, we typically say on our retention service, 500,000 MAU is a good proxy for a company that's got a decent level of product-market fit and a decent level of cohorts to experiment with. But what just looking at the MAU number doesn't tell you is what the new user acquisition is like. Particularly if you're going to work on early funnel, stuff like activation onboarding, which are often very great inputs for improving retention over the long term, then you also need to be acquiring users at a decent rate, so some tens of thousands of new users per month at least.
[00:07:55.680] - Olivier Destrebecq
Nice. What are the big levers users use to increase retention?
[00:08:01.530] - Andy Carvell
In the Mobile Growth Stack framework, which I published, and just back to your intro. Yeah, originally published, that when I was still at SoundCloud running the retention team there. It's something that we've continued to develop and publish while at Phiture.
[00:08:13.690] - Andy Carvell
But yeah, we have an engagement and retention layer. One of the key elements of that is the product because I always say engagement and retention starts with product. If you don't have an engaging product there, which comes back to this whole question of product-market fit, but if you don't have an engaging product, then it doesn't really matter what else you do.
[00:08:32.530] - Andy Carvell
You might be able to trick some users into coming back a couple of times by sending them a push notification, or a summary email, or something, but you're not going to keep them coming back over the long term. You're not going to achieve long-term retention if the product is not solving a problem or delivering value in some way for the users. It all starts there.
[00:08:50.040] - Andy Carvell
But assuming that you have a valuable product, then there's a lot of other stuff you should work on, so onboarding and activation being really key steps. I think the attention span of a typical mobile user or basically typical person in this day and age is very low. The barrier to downloading and installing an app is typically very close to zero.
[00:09:11.970] - Andy Carvell
It's a situation where somebody can just decide to install your app in an instant. Maybe they've seen an advert or they were just browsing, maybe just download it on impulse. They're interested, but they are going to very quickly get distracted. You've got essentially maybe 30 seconds, maybe a minute while they're in that app to give a good impression and to get them ideally through an onboarding step. If there are set up steps like account creation and things, then ideally you want to get them to a point where they're already seeing the value of the app within that first session because many of them won't come back even for a second session.
[00:09:46.160] - Andy Carvell
Typically, because there's such a steep drop-off in that part of the funnel, I'd say onboarding and activation is typically where we see the most upside in working. Then in terms of ongoing engagement, things like gamification or behavioural science-type stuff to try to build loops and try and generate habits can become super interesting. Things like run streaks, for example, or weekly stats, summaries, and things like that just like an exercise app where you're getting summaries, and scores, or maybe it's comparing you against other users on a leaderboard telling you you're 35th fastest at the 10-kilometer race in your area or whatever it is.
[00:10:24.820] - Andy Carvell
Yeah. These stats and quantification stuff works pretty well. In terms of channels, I think I already mentioned, things like push notifications that are still great, still super valuable. Email marketing also still works. I think email is getting tougher and tougher as a channel, but there's still value there. In-app messages, I'd say, are in terms of messaging channels, like anything you can deliver to your active users within the app, either as a popup or a full screen interstitial or as part of the app UX in a feed or depends how you expose it. But in-app interactions with users are incredibly impactful.
[00:11:00.640] - Andy Carvell
It really then becomes about how you leverage these channels to get across a message, which is adding value for the user rather than just distracting them or annoying them.
[00:11:09.680] - Olivier Destrebecq
One thing that struck me is talking about the onboarding and how you want to get the user as soon, as you can, to a point where they see the value of the app. I think I've worked on so many projects where that just wasn't there. What you said is resonating with me to make sure that you get the users to the value to the a-ha moment in a way, as soon as possible.
[00:11:28.420] - Andy Carvell
Yeah, I think it's supercritical, but just caveat that there are definitely different trains of thought and different approaches that both seem to work in different scenarios on how you get there. There's an argument for actually adding more friction in the onboarding process. You're going to trade off the number of people who complete the onboarding by adding extra steps.
[00:11:48.070] - Andy Carvell
Great example that I used in an article that I read recently was Fastic. It's an intermittent fasting app. They have something like 16 onboarding screens. They ask a lot of information, it's very playful, it's very nicely presented. I have to assume that they've optimised it a fair bit.
[00:12:05.560] - Andy Carvell
But the classic wisdom is keep it short, cut out screens, get them through the onboarding as quickly as possible. Fastic actually asks a lot of stuff. There's also a train of thought and data to back it up that the more that you show interest in the customer earlier on or in the user, you asked them a lot of questions, you get them to actually invest a little bit of themselves into that process, then they get the feeling, whether or not it's delivered on or not, that the subsequent experience is going to be more personal and more personalised to them, more tailored, maybe, also fitness, actually.
[00:12:42.070] - Andy Carvell
But fitness, diet, and fasting apps, things like this, they do promise to the user that they're going to create a tailored plan just for them. If the user thinks that they're going to get a higher quality experience out of investing more time in the onboarding, then sometimes they'll do it. Certainly, what we do see is that users who go through a slightly more high friction onboarding, even though less of them will complete it, the ones that come out the other end are more sticky and more highly retained.
[00:13:08.370] - Andy Carvell
It's not an easy decision as to whether to make it super fast or to add a bit more friction. But in any case, you want to try and get them through it in the first session.
[00:13:16.420] - Nicolas Tissier
I have a quick question regarding the account creation. Does it make sense to actually force the user to create an account during the onboarding? Or should companies focus on explaining the value of the product and getting the intent of the user?
[00:13:31.970] - Andy Carvell
Another great question there, Nicolas. I'm going to really give my classic answer, which I hate to give this answer, but it depends. I've seen firsthand. I've run some experiments with this at SoundCloud. Initially, we had a hard sign-up wall, as we call it. The first thing the user had to do was create an account.
[00:13:52.040] - Andy Carvell
Then actually, with a bit of pressure from Apple, at the time, Apple was really pushing everybody to make the experience accessible to everybody on apps rather than forcing them to sign up. We actually changed it.
[00:14:03.520] - Andy Carvell
We had, it wasn't exactly an AB test, but we had Android still with the hard sign-up wall, Apple with what we call anonymous experience, where we could basically force them to sign up later if they would keep using it, or at least heavily suggest to them repeatedly that they do so, which is typically what you have to do if there's value for you as a publisher to have users to create an account. Often, there is value because you can then reach them via email, for example, which is typically hard to do when they haven't created an account. Maybe you also get their phone number or other information which you can use. If there's value in that, then of course you want to have the highest number of registrations.
[00:14:40.210] - Andy Carvell
What we saw at SoundCloud back then and caveat, this was six years ago or so when we experimented with this, but we were never able to get sign-ups to the same level as having a hard sign-up wall when we asked people to sign up later. We didn't optimise those efforts enough, but we've seen it also in other apps since that we've worked with. If you want the user to create an account, it's generally good to get them to do it early because that's when they have the highest intent. Later, they're going to be less motivated to do it, particularly if they feel like they can access a lot for free, anyway.
[00:15:12.290] - Andy Carvell
This is quite similar, I guess, Nicolas, to the question of when you should ask the people to subscribe, which is also an it depends question, I think.
[00:15:20.600] - Olivier Destrebecq
You mentioned that the account sign-up never going to match the hard sign-up screen. But do you know from a business perspective if that actually had an impact on revenue and long-term value of a user, or was it a wash?
[00:15:33.930] - Andy Carvell
Yeah, there were some negative effects for us, for sure, particularly with SoundCloud. Your account is pretty critical, actually. It's where a lot of the drivers for retention, which SoundCloud was around that time, it was starting to roll out an attention-based models, in other words, advertising revenue as well as a subscription. They're monetising listening behaviour, which means that people need to come back and listen.
[00:15:57.960] - Andy Carvell
Essentially, they're monetising retention on the advertising tier at least, and arguably also on the subscription tier because if they don't come back, they're going to cancel the subscription at some point. Creating an account on SoundCloud meant that you had somewhere to store your favourite tracks. You could like tracks. You could also repost them to other users, so it had some viral loop effects as well on the rest of the ecosystem.
[00:16:18.160] - Andy Carvell
But just thinking about user behaviour, a lot of SoundCloud listening behaviour was based on tracks that users had listened to already. Being able to like a track and store it in my Like Tracks, being able to create my own playlists that I could come back and listen to later, these were like actually fundamental things which were only possible with an account. Now, arguably, we could have created an anonymous account structure for them. But that was not something that we built out. We could have essentially given everybody an account, just not tied it to their real world identity or email address.
[00:16:49.590] - Olivier Destrebecq
Email. Yeah. Interesting. One of the other tools that you've mentioned when we talked last week was getting the user to build a habit around your application to help with retention. Do you have an example of an application that does a great job of helping your user build a habit?
[00:17:06.210] - Andy Carvell
Helping your user to build a habit around retention. Yeah. I mentioned run streaks earlier, so I can give a couple of examples from the lifestyle category. One of our customer's headspace, who we've been working with for a few years now, I think they do a fantastic job of helping users to build habits. They have this thing called run streak. If I meditate every morning, my run streak increases by one. If I miss a day, my run streak drops back to zero.
[00:17:31.740] - Andy Carvell
After a while, if I built up a 30-day or 100-day run streak, then there's a certain fear of losing that.
[00:17:39.040] - Olivier Destrebecq
Yes, that's pretty harsh. Back to zero.
[00:17:41.170] - Andy Carvell
You become attached to it. It's actually back to zero, back to square one. I think a lot of apps that are built around habit, exercise, and things like this, also use these dynamics, which are broadly lumped in a behavioural science techniques. I think also anything like this, where there's regular usage and regular activity like weekly stat summaries or maybe even daily stat summaries, depends on the cadence that people are using it, or it could be monthly, but a regular summary of your progress within a certain app.
[00:18:09.960] - Andy Carvell
I think, again, built on this same principle of progress, essentially. I had a little dog in the United Kingdom when I was living there, and she used this fitness tracker, which is a little device attached to her collar which would track her movement. The device and the app was called Whistle, a really nice app. They gave really good stats summaries, so I would get this cute little summary every week telling me if the dog had hit her exercise goals.
[00:18:36.280] - Andy Carvell
This was great. When I moved to Berlin, she stayed with my parents, which was very sad, actually, that I was not able to bring the dog. But what it meant was that I could keep track on whether my dog was getting taken for enough walks and I could call my parents up and say, "Hey, look, she didn't hit her fitness goal this week." So yeah, it was good for retention.
[00:18:53.070] - Olivier Destrebecq
Yeah, I'm sure. The other question is who would get the bone if she hit her fitness workout goals for the week?
[00:19:01.890] - Andy Carvell
Well, I hope the dog would get it.
[00:19:06.210] - Olivier Destrebecq
There's a lot of questions from startup when they get going on what tools should they use to work on their retention? Do you have any advice there?
[00:19:15.790] - Andy Carvell
Yeah. As I mentioned, you've got very early stage pre-product-market fit retention work, which is mostly in the product. Even there, I would recommend to have some decent analytics, behavioural analytics that you can see what users are doing. You can measure their progress through certain key funnels like activation, maybe purchase, things like this, to have deep links properly set up on your product. Assuming it's a mobile app, being able to link users to every part of the app is going to be super helpful, both within the product to be able to leverage those links, but also, particularly when you start adding on email push notifications, in-app messages, these things, to be able to actually drive users from a message to a particular section of the app becomes super important.
[00:20:00.400] - Andy Carvell
It's often something that's overlooked. When developers are building new features and screens, they sometimes forget to implement the deep links. You definitely have your deep links set up. Some capability to run productivity testing becomes more important as your cohorts grow. Some folks will build their own framework for that. There are also various solutions on the market that can help you with productivity testing.
[00:20:21.440] - Andy Carvell
Similarly, if you want to start doing AB testing and experimentation with messaging, you're going to want to layer on a customer engagement platform, there are some great tools out there that can help you with these channels, like push, in-app, and email. They do some basic analytics, but they're not going to give you the full picture, a product analytics package will have. It's possible to combine those data sources to get a richer view of how your experiments are working.
[00:20:44.420] - Andy Carvell
Then, of course, a subscription platform like Purchasely wouldn't go astray.
[00:20:49.670] - Nicolas Tissier
Yeah, it allows you to focus on your product, which is the most important part, actually. That's one of the main advantages.
[00:20:56.820] - Olivier Destrebecq
What evolution do you see coming over the horizon on all those tools that you just mentioned? Are there things coming down the pipe?
[00:21:03.710] - Andy Carvell
Yeah. I mean, the tool space is, I'd say, one of the most active and always rapidly evolving. There's new tools springing up all the time, either directly challenging some of the bigger players. So I'd say, for example, in the CRM space, there's folks like Salesforce that have been around forever, and they're gigantic multi-billion companies. You've got newer companies snapping at their heels that are more agile, more designed for the mobile space, folks like Brace, who just IPOed recently. They are, I think, really starting to eat the lunch of Salesforce.
[00:21:37.150] - Andy Carvell
But then you see also a lot of consolidation within the larger platforms, amplitudes building out, all kinds of stuff, into predictive analytics. They're building out, I think, stuff in their customer data platform direction. Then we see customer data platforms like [inaudible 00:21:53], acquiring analytics firms iteratively and starting to integrate those things. We saw Airship acquiring Apptimize, which brings productivity testing and customer engagement.
[00:22:04.460] - Andy Carvell
This consolidation, it's something that I've seen. I've been in the industry a long time. I see it happening all the time. Everyone's always trying to build the perfect suite where they can be the one-stop shop.
[00:22:14.020] - Andy Carvell
But it never works that way because there's always a new tools, which you need, which are actually new categories or subcategories. Things like subscription platforms, this is not a category three or four years ago, at least not to my knowledge. It's a pretty new space, and there's a whole new set of really helpful tools out there that can really help you with that.
[00:22:34.930] - Andy Carvell
Lately, I've just seen a conversation with a few very early-stage startups who are interested to get my opinion on whether their tools for really focusing on data taxonomy, quality, and validation. I spoke to, in the same week, two or three startups all focusing on this very specific challenge. Sometimes folks, I think, start off with what seems like a small niche, and then they build it out until they become a giant. Then they start acquiring other adjacent companies to try to create a monolith. Then they get disrupted and the whole thing starts over.
[00:23:08.790] - Olivier Destrebecq
In one of the trends that you mentioned earlier with predictive analytics. I'm curious about that. Can you tell us a bit more about that one?
[00:23:14.820] - Andy Carvell
Yeah. Actually, it's one of the things, which I devoted a fair bit of space to in the 2022 edition of the Mobile Growth Stack, which I just presented back in December. I think our position on it at Phiture is that the companies that are going to succeed over the next two, three years in this space... I don't want to predict much further out than that because it's so fast-moving. But I think the companies that will succeed, or at least a lot of them, will be assisted or powered to some greater or lesser extent by machine learning or other forms of predictive analytics that are going to help them to do smarter, more fine-grained segmentation and targeting of specific users to identify users who can then be interacted with in a certain way.
[00:23:58.260] - Andy Carvell
Ideally also, joining the loopback on the interaction with that user to see okay. Let me put it in concrete sense. Let's say, identifying users who are at risk of churn. This would be a classic predictive analytics challenge, one of many. But propensity scoring in general, propensity to purchase, propensity to churn, propensity to like a Britney Spears track, to put it into a content recommendation format.
[00:24:23.980] - Andy Carvell
Just understanding is somebody likely to churn is definitely helpful because if I can identify a segment of users to live constantly refreshing segment, these users are my at-risk segment, then I'm like, "Okay, right now, I have a chance to interact with them and hopefully reduce that churn." Of course, then I have to also figure out what's the right interaction there that can modify their behaviour.
[00:24:45.700] - Andy Carvell
I think this is the bit that predictive analytics often, particularly when it's just the analytics part, it doesn't solve. It can tell you, "Oh, this person is likely to churn. That's useful information. But if I can't modify that user's behaviour, then ultimately, it's not super valuable to me to know that they're going to churn ahead of time."
[00:25:01.970] - Andy Carvell
But combined with a conscious effort and the tools to interact with that user, maybe to bring them an offer, or just to remind them the product exists, or send them some new content, or whatever it is. That's also a lot of work, a lot of testing to figure out what is the right form of interaction for this user in this situation.
[00:25:19.960] - Andy Carvell
But if a company can join the dots and all of these things and get measurable uplift, which over time also then feeds the algorithm to get even better at it, then they start to build a very sustainable, competitive advantage versus apps and companies that are just doing it the old fashioned way.
[00:25:34.360] - Olivier Destrebecq
Interesting. You mentioned users churning. One of the tools that marketers use today are win-back offers so that when the user churns, you try to get them back. Do you have any advice on how to do that properly?
[00:25:46.780] - Andy Carvell
Yeah. I've seen various folks using different approaches to tackle this challenge of win-back. I would say the first thing before I would get into specifics on offers would be to question whether you need to actually give an offer to the user in the first place, like coming back to what I just said about figuring out the right interaction. Typically, I would suggest that you don't jump straight to offering a deep discount, because maybe that's not actually the thing which is going to move the needle in all cases.
[00:26:14.510] - Andy Carvell
Maybe sometimes the user just needs you to pay them a bit more attention. Maybe actually the problem was not with the price, but maybe they just couldn't find what they were looking for in the product. Maybe they didn't get the right content recommendations, for example. There can be many reasons why users churn, which again, coming back to the predictive analytics.
[00:26:31.210] - Andy Carvell
The predictive analytics might not tell you why the user is going to churn. It might just say, "Okay, this user is probably at risk." I think understanding churn reasons is both difficult to do because these users, by definition, are the ones, which are least engaged and the ones least likely to be interested in filling in surveys and things.
[00:26:49.200] - Andy Carvell
But I would encourage everyone to try to understand why users are churning, both on an individual level and more collectively, to understand what are the motivations of folks that are leaving. The discount is maybe a band-aid on a bigger problem with the product. Actually, yeah, maybe discounts in the short term, but you really also need to fix the underlying problem.
[00:27:10.640] - Andy Carvell
But then back to discounting specifically, then it becomes a question, "Okay, at what point do we offer discount? How much?" Again, that's not a testing, I think, really, to figure out the right thing. I've seen some folks using couponing or reward platforms like customer loyalty platforms to help with that process.
[00:27:27.490] - Andy Carvell
There tends to be the bigger companies that can afford to bring in a specific dedicated customer reward platform or loyalty platform on top of their stack. We don't generally see it on the smaller end. For smaller startups, I'd say, I guess where I started, like talk to users as much as possible, try to understand what is going to motivate them.
[00:27:44.900] - Andy Carvell
Sometimes that will be a price discount, particularly if you're seeing more competition in your category and there's a new entrant at a lower price, then maybe you have to match that price. But nobody wants to get into a price war if they can afford it.
[00:27:56.580] - Olivier Destrebecq
Definitely. Thank you very much, Andy, for all those answers. It was really awesome getting to hear all your perspectives on those questions. If our listeners want to learn more about you and about Phiture, where can they go?
[00:28:07.310] - Andy Carvell
They can reach us at Phiture.com. That's with a PH. We spell it slightly ambiguously, so it's P-H-I-T-U-R-E.
[00:28:15.250] - Olivier Destrebecq
We'll put it in the show notes.
[00:28:16.590] - Nicolas Tissier
Great. Thanks. Yeah. They can also check out our blogs. We have mobilegrowthstack.com and asostack.com, which are quite well-read blogs. But yeah, I wanted to say thank you so much for having me on the podcast. Really great talking to you and thanks.
[00:28:29.360] - Nicolas Tissier
Thank you very much indeed. It was so much fun. Thank you.
[00:28:32.460] - Andy Carvell