Retention and engagement go hand in hand—but they aren’t one in the same. In a recent episode of Retention Masterclass, John Koetsier and I talked to “the Godfather of the Mobile Growth Stack” Andy Carvell, partner at Phiture, a growth consultancy based in Berlin. He puts the difference in simple terms: “So the way that I always kind of break it down is like, engagement is what the user is doing within the app, it’s the actions that they’re taking. Whereas retention is whether they’re coming back at all.”
Together with CleverTap and Phiture, Andy helped design the AIC framework to enable fast-growing consumer app companies to better understand the depth and intensity of user interaction. The framework categorizes user activity into one of the three types of engagement layers: Acknowledgement, Interest, and Conversion — with Conversion being the highest-value action a user can take within an app. This allows brand marketers to recognize which in-app activities contribute to revenue growth, thereby improving overall user retention strategies. In other words, Andy helped define a framework that allows app marketers to understand which engagement activities lead to better retention.
Listen to the podcast:
“So retention, in its classic measurement, won’t tell you what the user’s doing when they come back, just whether they’ve come back or not in a given time period,” says Andy. “Whereas engagement is really [about] app-specific metrics and measurement around like what’s the user actually doing within the app? Are they doing something valuable? Are they just literally opening the app and then bouncing out again?”
When you take the time to understand AIC and how each section fluctuates over time, not only will you have a better understanding your users, but you’ll be able to better drive retention.
AIC from A to Z
We all know that our data tells the story of user engagement, but making sense of the data—especially as we accrue more and more of it—can be a challenge. Frameworks can help us make sense of the ever-increasing amount of data available. “So what AIC gives you in its purest state is essentially a snapshot of your active users in any given period, and allows you to kind of break that down to a better level of visibility of what’s really going on,” Andy says. “So just as I mentioned earlier, your active users or your retained users in any given period, let’s say monthly active, only really tells you whether they were seen within the app or not. Whereas, if you actually want to know if they’re doing something valuable, it’s helpful to kind of maybe stratify that a little further. So AIC stratifies your active users into three buckets.”
- Acknowledgement—Imagine an AIC pyramid, and at the bottom is the least valuable engagement. “You’re always going to have the biggest chunk of users who are at least present in the acknowledgement layer, which means that they’ve either opened the app or clicked on a push message, they’ve basically just been seen within the app. So your acknowledgement layer basically includes all of your active users if they’ve been in the app, they’ve acknowledged that your app exists.”
- Interest—The second layer of the pyramid is the interest layer. This pool is narrower, and the actions these users take show more intent. Perhaps they’ve put an item in their cart but then left the app before competing the purpose.
- Conversions—And at the tip of the pyramid are that small, coveted pool of users who convert. If we stick with the e-commerce example, this would be a purchase. “But it could also be, many other apps that are not monetizing or not selling something, it could be posting something, for example, in a social media app,” says Andy.
If there’s an event tracked for each of those, then you can look at a graph of which of those events are happening the most often. He says to look at a histogram and at your top events to understand which actions users taking most often. “That will generally give you an idea of which are the most valuable events within the app,” Andy advises. “I’ll qualify that, somewhat. Depending on how you’re tracking, there may be events which are what we call ‘filler events.’”
Once you’ve refined your tracking and understand what users are doing the most, ask yourself, what are the key things which my app is doing? What is the key problem that it’s solving? What is the key value that it’s bringing to the user? And what events or what actions does the user take which unlocks that value? And how do you unlock that value to retain more users?
Ultimately, a good framework doesn’t tell you exactly what to do. “It’s not a shortcut to thinking,” Andy says, “but what it does do is provide, hopefully, a way to kind of guide your thinking and guide your activities based on some kind of some rules or some thinking which helps to kind of demystify what might be a complex topic.”
To learn more about AIC and other frameworks, read the full transcript below.
John Koetsier: Hello and welcome to Retention Masterclass. My name is John Koetsier. I am one of your hosts, along with Peggy Anne Salz.
Today we are chatting with a very interesting person. He is the inventor of the mobile growth stack, and we’re going to dive deep into some metrics of engagement, AIC, RRF RFM, and maybe even talk about a few mobile growth disasters.
Peggy Anne Salz: Absolutely. Thank you, John. I’m Peggy Anne Salz, as you said, and we have a great guest, because I’m going to rattle this off and you’ll say, wow, right? You said the growth stack. I would say he’s the… my term for it, “godfather” of the mobile growth stack. I was there many years ago when he brought this up and said here’s what it looks like, and we were all astounded because it was so overwhelming. Since then, he and his agency have gone on, they have been named the best app marketing agency globally in 2019, best retention campaign as well 2019, best personal contribution from our guest, keeps on going.
He built the AIC framework we’re going to be talking about, and he runs, as you said, the nightmare, the disaster Mobile Growth Nightmares podcast. I won’t go down that road… probably going to bring some interesting anecdotes as well based on that. Our guest is Andy Carvell, he’s partner at Phiture growth consultancy based in Berlin.
Andy, awesome to have you today!
Andy Carvell: What an intro, Peggy. Thanks for having me on, John and Peggy, I’m really happy to be on the show.
John Koetsier: I was very happy when I was hearing Peggy talk about nightmares because she didn’t mention podcast right off the top, I was hoping she didn’t mean your agency right there but… clearly not, clearly not.
Andy Carvell: Hopefully we don’t have too many nightmares at Phiture, but…
John Koetsier: It sounds like you don’t have any, to be honest, with all those accolades. So let’s start with some basics. Retention is the goal and it’s a big game changer, but can you define it for us? How’s it different from engagement? Those terms are sometimes getting blurred.
Andy Carvell: Yeah, they’re very closely connected of course. So the way that I always kind of break it down is like, engagement is what use is it doing within the app, it’s the actions that they’re taking. Whereas retention is whether they’re coming back at all. So retention, in its classic measurement, won’t tell you what the user’s doing when they come back, just whether they’ve come back or not in a given time period. Whereas engagement is really quite app specific, metrics and measurement around like what’s the user actually doing within the app? Are they doing something valuable? Are they just literally opening the app and then bouncing out again?
That’s something we’ll talk about, I guess with the AIC framework which tends to break that down a little further. But, yeah, that’s kind of how I would define it. Retention: Are they coming back? Engagement: What are they doing when they’re there?
Peggy Anne Salz: Okay, so when I’m a marketer, I’m looking at this and I’m saying, ‘Yes, I’m in on this. I want to be understanding retention and influencing user behavior.’ But then the question comes down to just as it was in acquisition, there are certain events that you want to understand. We don’t know what they’re calling them. I think you have a name for them as well, Andy, I’m hearing “critical events,” I’m hearing “habit forming events.” Those moments in the app journey where you’re saying, ‘Yes, this is where I am in the retention’ in your pyramid that we’ll get to, but can you give me some ideas about how to identify those because that’s where I’m hearing marketers are truly stumped.
Andy Carvell: Yeah, sure. I mean, you can kind of come at that from two directions, right? You can look at just the analytics and do it in a sort of a bottom up kind of way, and look at what … provided that you’re tracking all of the core actions that the user’s taking, so basically like what screens they’re visiting and anything that they can actually do within the app in terms of interacting with features.
If there’s an event tracked for each of those, then you can look at sort of just a graph of which of those events are happening the most often. Like, look at a histogram and look at your top events and see which events are users taking the most? That will generally give you an idea of which are the most valuable events within the app. I’ll qualify that somewhat, there may be depending on how you’re tracking, there may be events which are just kind of like what are called “filler events,” things like if you’re tracking clicks on everything then you’re going to see a lot of click events.
So you might want to kind of refine it a little closer, but you can look at basically the numbers and see just objectively, what are users doing the most? And from that, then I would sort of apply a kind of a top down kind of view which you might also want to start with, which is like just what are the key things which my app is doing? What is the key problem that it’s solving? What is the key value that it’s bringing to the user? And what events or what actions does the user take which unlocks that value, right?
So in an eCommerce app, that’s likely to be things like product views and purchases. If it’s a meditation app, it’s probably going to be doing meditation sessions. So the specific actions and events will be very specific to your app and your category, but the way to identify which ones are most useful, I think is ultimately down to your product proposition and the problem that your app is trying to solve.
John Koetsier: It’s funny because you mentioned meditation. I was thinking of Muse. I have a Muse by the way, I have the app and I’m kind of stressed out because I haven’t been using it recently, so …
Peggy Anne Salz: I could never tell that John, you don’t look stressed to me.
John Koetsier: Exactly. Let’s get into some of the alphabet soup. You talk about AIC, RRF, RFM, some of that is stuff that many marketers may not have heard of. What are they? How do you use them? And where do you use them in different segments or different areas of what you’re doing as a marketer?
How do they help mobile marketers?
Andy Carvell: Yeah, so I’m a big fan of frameworks and I’ve developed some myself. You’ve mentioned a couple, we also like to use frameworks in our work at Phiture, and when we’re working with clients I find, first of all, that well developed frameworks that are based on data and expertise really help to kind of put a context around the work that you’re doing in marketing, or in any other discipline really.
And so a good framework doesn’t tell you exactly what to do, it’s not a shortcut to thinking, but what it does do is provide hopefully a way to kind of guide your thinking and guide your activities based on some kind of some rules or some thinking which helps to kind of demystify what might be a complex topic. So that’s just where I would start with my kind of love of frameworks.
Secondly, so you mentioned three there: AIC, RFM and RRF. I guess I’ll tackle them in that order because I think they’re kind of a useful prism.
So AIC, something that I developed last year and published, yeah, I can’t remember exactly what date last year, but it’s got some pretty good feedback and we use it at Phiture quite a lot with our clients to help them to get a better picture, kind of a deeper picture of engagement, engagement and by definition retention. So what AIC gives you in its purest state is essentially a snapshot of your active users in any given period, and allows you to kind of break that down to a better level of kind of visibility of what’s really going on. So just as I mentioned earlier, your active users or your retained users in any given period, let’s say monthly active, only really tells you whether they were seen within the app or not. Whereas, if you actually want to know if they’re doing something valuable, it’s helpful to kind of maybe stratify that a little further. So AIC stratifies your active users into three buckets.
So the A, sense of acknowledgement, which is kind of – you can imagine this like a pyramid, I’m going to do it with my hands – so you get the broadest layer at the bottom. You’re always going to have the biggest chunk of users who are at least present in the acknowledgement layer, which means that they’ve either opened the app or clicked on a push message, they’ve basically just been seen within the app. So your acknowledgement layer basically includes all of your active users if they’ve been in the app, they’ve acknowledged that your app exists.
The second layer, interest layer, we just talked about kind of defining those key actions, which are kind of more important and more indicative of genuine user engagement. And so I’m going to come back to that one first because it’s easier to define the interest layer by the conversion layer, which is the top of the pyramid. You’ve got usually a few users who are, hopefully more than a few, who are taking what we call the core action, or it could literally be a conversion in the case of, you know, if there’s a purchase, if it’s an eCommerce app or something else where the key activity is buying, then it might be buying behavior. But it could also be, many other apps that are not monetizing or not selling something, it could be posting something, for example, in a social media app.
That’s often the core action that you want users to take and the representation of the ultimate value that they’re getting from the app, but also that they’re contributing to the platform.
So typically an app will only have like one core action. With Pinterest, maybe it’s ‘pinning a pin’ I think is the terminology, but, you know, pinning a piece of content. With Airbnb, it’s making a booking, so that is really a classic conversion action. And so typically there’s one action that, ultimately, you hope that all users might do at some point, and that will be your core action.
So when you look at this pyramid, it’s not mutually exclusive, right? So anybody who’s in the top layer would also be in the acknowledgement layer because they would have come into the app and then they would have done the core action and probably they’ll have done some other actions during that session as well. And that’s what you see in the interest layer, which is basically you filter out all of the kind of filler actions that I just mentioned and really select all of the other actions the user could take which are indicative of engagement and maybe intent, but it’s not the core action that you want users to take.
So we’re looking, basically their like interest layer actions in a social media app might be that they’re browsing, commenting on other people’s posts, liking, commenting, replying if it’s a social app. If it’s a content app, then maybe it’s just kind of clicking around exploring content. And I think that’s kind of an intent-driven and sort of deliberate action by the user that’s interacting with some of the core functionality and features rather than, for example, just like going into the menus and changing their settings, that wouldn’t typically be something that you would count as interest. So anyway…
John Koetsier: It seems like a really core chunk there actually, really…
Peggy Anne Salz: It is.
John Koetsier: …because at that base layer, you’ve got everybody who just popped in, right? But in that interest layer, you’ve got things that you can take a look at and say, ‘Hey, people who do that lead to conversions, or people who do that 10 times in a week lead to conversion.’
So you get some idea of how you can drive your core growth loop right?
Andy Carvell: Yeah, you’ve exactly got it, John. Yeah. So what’s interesting with AIC and the way that we typically apply it, is to look not just at that snapshot, which gives you a good general kind of pulse check on the general health of your active user base. But then to also look at things like the elasticities between these layers.
Like if we get more people in to at least acknowledge that the app exists and we get more people in for more sessions, does that have a good correlation to taking interest actions and does taking interest actions have a higher correlation to the conversion action? And if the answer is yes, and if those correlations are high, that can really inform your strategy in terms of where you want to drive users to and how aggressively maybe you want to do acquisition, for example. If you see a really good correlation between acknowledgement and conversion, then by all means, just turn on the taps and buy more users, because you know that a decent chunk of them are going to convert, and so looking at these ratios between the layers, what we call the elasticities, as well as just the trends over time as you kind of make various changes in the app, then you can really start to understand the kind of engagement dynamics that you’re seeing in the app.
Peggy Anne Salz: I mean, I love that and we’re going to get a little bit into RFM because there’s an interplay there as well. You know, this is a framework, this is a nuanced way, sort of a gradient way of looking at engagement.
But before I go there for a moment, just to clarify here, Andy, you know it’s interesting that you can look at this model and you can already start to move some levers is what I’m hearing. I’m hearing if there’s a lot of acknowledgment and they’re not moving through, you need to nudge. Or the other way around, if you have this massive conversion up there, it might say maybe you’re making it too simple. Or maybe it’s a great opportunity you’re missing to even make it. So give me an idea here, because we want in Retention Masterclass for people to have takeaways.
It’d be great to understand just within this framework, what can I be doing at different levels because it is indicating an action I can take.
Andy Carvell: Yeah, sure. So one kind of practical application I’d advise people to look at, you can look at your sort of, you know, once you’ve got this model calculated, you can graph it like looking back over your historical data. So maybe graph it back over the last six months if you have that data available. And essentially plot three lines, you know, one is the size of your conversion layer at any given point. The second would be the size of the interest layer, so size, I mean like how many people were in that layer in that time chunk. So you can look monthly, for example, or ideally at a higher resolution, like weekly or even daily, and then try to tie that back. You’ll see some trends, you’ll see that that’s not a straight line because there’ll be different stuff happening, more acquisition at certain times, maybe you’ve run special offers and things. And then you can start to sort of see how these dynamics were affected by certain key events in that last period.
So when I talk about events here, I’m talking about sort of marketing events, things that you might’ve done. Maybe you had a product launch or maybe you got featured in the app store, or maybe there was a press release and you got featured in TechCrunch, and a lot of extra users came into the app. You’ll see whether these events or these activities were worthwhile from a conversion perspective. You’ll see okay, yeah, quite often, for example, you get featured on the app store, your acquisition goes way up but these users don’t stick around.
And you know what an AIC kind of plot will show you is getting featured on the app store or something that we should really push for it because we can see here, actually we had a decent spike and yes, some of them did actually end up doing interest actions, and maybe some of them also converted. So on balance, this was a valuable activity, but sometimes you’ll see you had a huge spike in the acknowledgement layer and the other two are absolutely level, like no change at all, which might suggest that it’s not worth really pushing for these kinds of big featurings because they’re not valuable. Or you know, that your value prop is just not landing with these users for some reason.
John Koetsier: Exactly, exactly.
Andy Carvell: So it could be something that you could do.
John Koetsier: What about RFM? Can you talk a little bit about that? How does that work? How does that fit? And where do you use that?
Andy Carvell: Yeah, so RFM stands for recency, frequency, and monetization. And there’s a few variations of it, recency, frequency, engagement, and then a couple of others. But this is like an old school, relatively old school as it did come from the digital marketing world. This is like a legit business framework, you know, from like kind of classical marketing and business academia. I studied it at business school, and so it’s great to see that, I always think…
Peggy Anne Salz: A comeback, it’s a comeback, yeah.
Andy Carvell: Great to see it being applied in digital businesses, because I do think, you know, a slight book bear for me is that a lot of these startups these days everyone thinks that they’re inventing everything from scratch and they’re totally breaking new ground.
John Koetsier: They’re not?!
Andy Carvell: And they’re doing something completely new, and often the marketers and the folks working on these products and marketing them don’t really look back to classic business theory and look at some of the great stuff that’s out there that could really help them, because at the end of the day, these are still businesses or they should be.
So yeah, RFM, a great framework. Basically by calculating this and then basically doing this, it’s a segmentation tool essentially, or at least that’s the way it’s typically used. So there are some tools and packages out there which make it simpler, or you could do this on analysis, like in Excel if you want. But by looking at users and again with past data and understanding of your users, who’s monetizing the most, who’s been here the most frequently or been using the app the most frequently, and who’s been using it the most recently.
And by kind of combining these three variables and doing a segmentation based on those, you can start to kind of identify your most valuable users, your at-risk users, your kind of really loyal ones, and do these kind of segmentations. And then, you know, of course you can then devise strategies and tactics to actually interact with these segments possibly in different ways, based on that understanding of where they are in terms of their loyalty journey or their kind of affinity with the product.
Peggy Anne Salz: So I love that. So we have RFM as sort of our basis, our foundation, our old school coming back to be new knowledge. We have AIC, which is that framework but also that lens, it’s a very nuanced way of seeing engagement. I like the way that we can see it in gradient because otherwise engagement is this big clump.
And then we have, of course, we’ll finish out our tour of alphabet here, RRF, the other model that we want to sort of, I want to understand how that interplays because I think the three of these together is the holistic way of seeing retention now and going forward.
Andy Carvell: Sure. So RRF, which I developed based on my experience building out a real time notification service at SoundCloud when I was running the retention team there. What I learned from building that system was that not all notifications are equal, and I was able to kind of distill it down to these three variables: Reach, Relevance and Frequency, and these have a certain tension between them, as in it’s quite hard to kind of get them all to be high. So the ideal campaign or notification, so now we’re talking about interacting with users, right? Interacting with the user segment, and you want to ideally be able to reach a large chunk of the user base or a large chunk of your target segment with any particular campaign or notification.
You want to reach them with ideally the highest relevance of message that you can, which is typically achieved through better personalization of that message, and you can proxy relevance fairly well with click rate if they’re clicking and opening on your message, whether that’s a push or an email, or in-app message, if the interaction with that is high that gives you a good indication that your relevance is high.
And finally, the frequency. How often can you send this particular campaign or this particular interaction to the user and it’s still being valuable, and not just tolerated, but ideally actively appreciated by the user. And so the holy grail is really having something that you can reach all of your users or a very large amount of them with a high relevancy message at a frequency which they’ll appreciate because the relevance is so high. And so there’s this tension between them. And so RRF is really a … it’s the actionable framework, whereas the other two are kind of more like your sort of planning stages, identifying and analysis. And then when you come to actually prioritizing your interactions with these users you can use RRF to kind of prioritize the ones which are likely to give you the most impact.
John Koetsier: I love what you said, that you can’t always get each of them at the same level. I mean, because reach and relevance are often at war with each other, right?
Andy Carvell: Totally.
John Koetsier: I mean, you can reach everybody and you’re probably relevant to 5% or 20%. There’s some things that are relevant to everybody, don’t get me wrong, but many things in a mobile marketing framework are only relevant to a segment. So do you apply a segmentation to that?
Andy Carvell: Yeah. I mean, I think you should always think about, okay, what’s the target segment that’s going to be most relevant for this particular message or this particular interaction, right? And then I would also just throw in there as well as just segmentation, you might also want to think about specific triggers. Triggers are great for increasing the relevance within that segment.
So, let’s take an example, like I maybe have my target segment, which is people who have never made a purchase in an e-commerce app. So that’s probably quite a big segment typically. And then of course, you could just bless them with a special offer and hope that that’s relevant to all of them. Maybe you could even increase that relevance by further personalizing that based on the history of who’s viewed certain categories of products, whatever, so that you could further segment that and personalize it further.
But then you could also think about increasing that relevance by building that rather than just on like a time send, building it on a trigger. So trigger the message when they’ve just looked at a product and then send them that special offer based on that specific product, or at least on that category, maybe 15 minutes later. So you’re reacting to something that they’ve actually done, but of course that will reduce your reach somewhat because they’re not always like trigger that trigger.
So this is what I mean, there’s a real tension between these three variables.
Peggy Anne Salz: I mean, you bring up the tension. It would be very interesting to hear a little bit about how you’re also addressing that. Because part of this is what are really the levers that a marketer has that they can be activating and pulling, pushing, playing, experimenting. I want to understand just how broad, how free, how possible is all of that because there are certain confines. If we look at the models themselves and how the framework works it seems as if there’s a lot of fluidity and flexibility, but I’d like to hear about that.
Andy Carvell: Yeah. I mean, I think there’s always room for marketers to be creative.
Peggy Anne Salz: Well that’s good news.
Andy Carvell: And, you know, these models are not meant to actually confine you as such, they’re just really kind of hopefully representing something that you should think about in terms of some of the natural constraints that you’re going to come across and you’re going to see in the data. But in terms of how you approach that and how you try to overcome them, I think that’s where marketers still need to be creative. So, I mean, I’m not sure if I’m answering your direct question, so pull me back if I’m off.
Peggy Anne Salz: I’ll pull you back just a little bit, Andy, but I have to say I do love to just let you go too. I really do, I enjoy this. I’ve known Andy, full disclosure, for what like maybe five years, and every time he speaks it’s just another insight. So I’ll challenge you on this and I’ll bring some practical stuff back in, Andy.
Andy Carvell: Sure.
Peggy Anne Salz: I mean, it’s really about at the end of the day, influencing behavior, you know, it’s about influencing how I’m going to engage, how long I’m going to be in the AIC, if I’m going to go up to converge, if I’m going to feed back into the RFM model. But the big question is what can marketers do to influence behavior? You talk a little bit about that tension, so how do they get to the best possible outcome?
Andy Carvell: Yeah, great question. So, in the mobile growth stack, I’ve kind of built out the various levers that I see as being most relevant for increasing engagement and retention.
That starts with product, right? If you don’t have an engaging product to start with, you can kind of forget about everything else. You know, if you don’t have that at some level of product market fit where you’ve got at least some audience out there who really appreciates the product proposition, understands it, and gets value from the app. If you’re not able to see that from the numbers, that it’s probably too early to be working anywhere except in the product itself.
But once you do have some indication that you have an engaging product with some product market fit, the levers available would be things like working on your activation flows, your onboarding, doing things like life cycle marketing and activity notifications through a channel such as push, email, in-app messaging, possibly even SMS. SMS is actually still a valuable channel in some categories and to some audiences. You might want to look at doing more community engagement and/or community support in order to make sure that the user doesn’t feel like they’re on their own with this experience, but actually they’re part of something bigger, or that they’re just well supported by the company that’s providing this experience to them.
Yeah, and maybe also gamification I’d say is something which not universally applicable, but you can incorporate gamification in lots of kind of quite interesting and subtle ways to non-gaming products, things like run streaks and sort of celebrating continual usage can be a great way to increase engagement within app over the long term.
Peggy Anne Salz: Interesting.
John Koetsier: I love some of what you said there, I’m going to highlight SMS for half a second because it seems like just before kind of apps became massive, SMS had a moment and everybody was bringing out an SMS marketing tool and it was really cool, and it was the next big thing, and then we all forgot about it… we all, many of us completely forgot about it.
Apps became everything and of course there’s much more data, there’s relationship, there’s all these other things that apps can bring, but the immediacy of SMS is really impressive. I mean, most SMS is a very high percentage get read, get seen, and a lot of them get acted on as well. So what I wanted to ask was of the tools in the toolbox that marketers have, whether it’s email, whether it’s SMS, push, in-app messaging, all that stuff, what are they using too much of?
What are they using too little of?
Andy Carvell: So I think push is fairly saturated as a channel at this point. I’m a big fan of push notifications still, I’ve personally seen the value that push can drive when it’s done well, and hopefully we did it reasonably well at SoundCloud. We certainly saw some really good uplift, like five percentage points on retention when we launched that activity notification service, and we were able to increase it further through its duration on that. But I do see a bit of push fatigue in the market, and also Apple and Google see that too, because they’re making it easier to opt out of push and sort of turn to make apps silent if they’re being too noisy with push. So I think push is fairly saturated, although it’s still possible to drive value there for sure.
Email, I mean you could argue that email was saturated a long time ago, but it’s actually still a valuable channel, so I would never advocate ignoring email as a channel. I think you’d be leaving impact on the table, but it’s a tough channel, particularly now with GDPR, double opt-in and all this stuff, it’s harder than ever to get email working for you as a channel, but I’d say it’s worth the investment. But it’s also not the first one I would build.
John Koetsier: Very interestingly just on email, I noticed on my last newsletter I had a 35% open rate and I was like, whoa, I’m doing something right!
Andy Carvell: Yeah, that’s great.
Peggy Anne Salz: I’m seeing that there is a shift in all of these, to be honest, we’re going to have to look back at the metrics because there’s a shift now in all of these different formats. Email is cool because, hey, we’re sitting at home, a lot, and this is what we’re doing. It’s like, oh, maybe I will look at that after all.
And congrats on that, by the way, John. Awesome!
Andy Carvell: I think it’s also a lot about understanding the dynamics of these various channels and understanding the strengths and weaknesses, right? Like there’s some things which are going to work much better sent as an email than trying to do it as a push or an in-app message.
I didn’t mention in-app messages yet, by the way. That’s the channel which I think is still in its infancy in terms of the amount of impact it can drive. And certainly what we’re seeing at Phiture is companies that really leverage in-app messaging to the fullest are driving tons of impact with it, like way more than the other channels that I’ve just mentioned. However, if you want to kind of go into detail and have some long form content or maybe a kind of a compilation of lot of things that you want to say, or a lot of new content you want to communicate to your user, something like email is a much better channel for that because you know they can read it in their own time.
You’ve got much more real estate to kind of go deeper on a topic than you can in an in-app message where you have just limited real estate on the screen, or push where you have even less, you just have a few characters really to get your point across. So understanding the channel dynamics, I think is really key to a successful CRM marketing mix.
Peggy Anne Salz: So you do, speaking of longer form content Andy, that’s what you do at Phiture. You put out these blogs, they’re not one part, they’re three part, five-part, and you just finished a three-part blog looking at the formats, you know, in-app, email, notifications. You talk about where they fit. I’d love to know where are we, I wouldn’t say failing, but where are we missing some opportunities?
Because I think that you have a few favorites where you’re saying if marketers just doubled down and got it a bit right, there would be great returns.
Andy Carvell: Yeah. So I already mentioned in-app messaging. I’m going to keep banging that drum just because I’ve seen firsthand how impactful it is. I saw that first of all, when I started experimenting with that channel at SoundCloud and we see it with our customers at Phiture every day, you know, the ones that really embrace that channel and do it well are really able to drive a ton of engagement, whatever that means for their app.
You know, it works for so many use cases, but of course only with their active users. So you can’t send an in-app message to someone who’s not in the app. So in terms of like other wins that I would say, I’d say coming back to the RRF, I think about that frequency, right?
What I see a lot of, and maybe this is a bigger topic, but I do see a lot of fear in general around CRM and growth experimentation. But I’d say particularly with push notifications, fear of sending too many, and I just said myself that people are a bit fatigued of them. So it’s not that that fear is completely unfounded, but there is a lot of potential impact on the table there.
And I think that if you frequency cap your notifications at a really low level and say, okay, we’re only going to send like two notifications per week because we don’t want to overwhelm the user and we’re afraid of overdoing it. If you’re doing that and you’re not basing that on data then you don’t know. Maybe you could actually send four or five and actually the users would be very happy about that. Or maybe you can actually start to segment your users into folks who actually respond well to push notifications or indeed any channel, and send them more, put them into a higher track where you’re increasing the frequency because maybe they’re actually fine with getting ten notifications, maybe in a single day even, as long as the relevance is high.
But maybe there’s another segment of users who are really not responding to these and not opening them, so maybe you put them into a lower engaged track where you put them into a quieter track for your CRM. And so by kind of stratifying the frequency in a bit more of a nuanced way, rather than just saying one global frequency cap for our entire 10 million users audience, it’s just not going to fit everyone.
By doing that and doing that exercise, I think you can unlock a lot of extra impact.
John Koetsier: I love that. I love that because I know that even for an individual, it’s on a per-app basis in some cases, where I can get a notification from in-app, I’m like ticked off. I mean, why are you sending that to me? That’s not relevant, I don’t want that, I’m deleting you. Right? But in other apps maybe that are more core to what I’m doing, what I’m working on, or something that I’m having fun with, I don’t mind getting that messaging.
Andy Carvell: Sure. Yeah. And it all comes down again to the relevance, right. If it’s relevant enough, then you don’t resent it.
Peggy Anne Salz: And you talk about core, and that’s also part of this because it all goes back to me understanding what are the core actions? What is that interplay between what people want in my app and how my product is delivering that, right?
And you’ve given us some great thoughts. You have some ideas about how we can move these levers around, some of the formats we can be experimenting with or taking more of an advantage of, you know, focusing more on. But at the end of the day, it’s that interplay between product and marketing. And I love it because I’m hearing a lot of people tell me product is the new marketing, and that gives us a new sandbox to play in. Of course it’s maybe not all that easy because product is product and marketing is marketing, and it’s a lot about bringing the two of them together as teams, not just the two concepts together. It’s easy to say product is the new marketing, but what are you seeing with your clients? I mean, it’s about also making it possible for them to work together.
A little bit of politics as well to be fair, Andy.
Andy Carvell: Yeah, I do see politics in various organizations and…
John Koetsier: Never, never.
Peggy Anne Salz: Never. It’s all kumbaya over here.
Andy Carvell: Yeah. I mean a lot of this also comes down to fear in my book. You know, fear often between different departments, it’s fear of the unknown sometimes. Particularly like growth experimentation, it’s a relatively new thing. When we were doing this every day and thinking about this it’s kind of old news, you know, test often test, like data-driven, run fast and break things. This is all kind of just part of our language, but still there’s a lot of more old school thinking going on in some product teams and some design teams, even in some marketing teams that I’ve come across.
And so I think these teams have to really work hard together.
And I’d say that being part of the newer discipline of data-driven growth, the onus is really on that team to take the lead here in evangelizing and explaining the processes and the methodologies around the work that they’re doing and to kind of get the other stakeholders of the company on board with this idea of growth experimentation to start with. And then you can really start to talk about specifics, like maybe running some in-app messages, which product sometimes object to because they feel like it’s kind of running over their product and their nice designs.
And you know, there’s various kinds of tactics and tricks, not tricks exactly, but things that I would suggest, like I’ve seen product teams get won over on in-app messages particularly once they see that actually they can prototype some new things which they were going to build out over three or four months as features. You can actually get like a smoke test done with kind of a prototype version.
But it does have to be based around this common understanding of metrics and data, which is not as universal as you might expect.
John Koetsier: That’s super interesting stuff, and you may see some research come out from Peggy and I about that in the next few months or so. I want to turn our attention to the growth stack. That’s super terrifying to some, image of all these boxes and layers and tiers of tools that you can use to grow in mobile.
It’s something I’ve been working with in various different fields and various areas, for probably over a decade. I’ve often said that people have a growth blob rather than a stack, or they have a lot of tools they don’t know how they all work together. I don’t know if anybody has them all in every box and are they slimming down because of recession or scaling up because they have to be more intense.
Talk to us a little bit about what’s essential and what you look at as maybe a starter kit, what you absolutely need to have, what you can add, and if you’ve ever seen anybody with a tool in every box.
Andy Carvell: Yeah, sure. Yeah, there’s so many tools out there, you know, a lot of off-the-shelf stuff and then there’s also a lot of teams are building their own tools and technology to do bits and pieces. So it can also be a mix of inhouse and commercial stuff.
And yeah, definitely, it’s more often than not that we see a company that’s got a mishmash of various tools that are not really well-integrated and not really driving the value that they should be, versus a really well-architected tech stack. This is why I added like the tech layer to the mobile growth stack framework in the left iteration that I did. In that layer I’ve tried to kind of pick out the things which I considered to be more or less essential that you would either build or buy. Although I’d say more typically you would buy these in because there are fairly advanced and robust solutions out there in the marketplace, so why would you reinvent the wheel?
So for me, that would be analytics, install attribution, unless you’re not doing any paid marketing at all, in which case maybe you can skip attribution. CRM or marketing automation has so many different names, but you know something that allows you to interact with users on these channels that I just mentioned, so push, in-app, email, in a segmented way. And then, yeah, you also need deep linking functionality, whether you build that yourself or whether you use a tool like Branch to make sure that you can get users, drive users from these messages or these interactions, into every different part of the app with a well architected, deep link schema.
That would be my personal recommendation of like the basic boxes that you want to tick. Everything else is kind of gravy.
John Koetsier: Interesting. I had to think when you’re saying ‘unless you’re going to build it yourself,’ I would say if you’re Machine Zone, right, build everything in house, we’ve controlled it all. That may change, they just got acquired by, was it AppLovin? It was somebody unexpected that just acquired them. I was… amazing. Anyways, we’ll get the facts on that, we’ll get that out there, but yeah, I’ll leave it to Peggy for the next one.
Peggy Anne Salz: I was going to just say I have to look at the latest growth stack actually, Andy, I didn’t know you added… I don’t know if that’s good or bad news if you’ve added another layer to that actually, because…
Andy Carvell: It was back in 2017 so I’m pretty sure…
Peggy Anne Salz: Oh okay! Oh I thought another one on top of that because it’s already, you know, you can’t even get it on one screen, so…
Andy Carvell: No, I’m trying to simplify it. That’s why I’ve not put out an update in two years, because I’m still trying to figure out which bits can I take out that actually people shouldn’t be considering. And unfortunately, it’s quite a complex environment that we’re operating in. It’s designed to be complete, but I’m always looking for ways to simplify it.
John Koetsier: It’s really challenging.
Peggy Anne Salz: So dare I ask, can we drop pieces out of it? I mean, you sort of said, okay, analytics, attribution, CRM, I get that, but that’s three of like 30 boxes. And we are in times that are quite challenging. I’m just curious, can we shave it down just a little bit more?
Andy Carvell: Yeah. I want to make a clarification though here. So the framework itself is a strategy framework based on activities that you might want to consider building out in your organization to help you kind of achieve these different marketing goals. This is not designed to be an entire framework full of technical solutions. So for each of those activities, you may or may not buy in some specific tooling, but the only kind of tech that’s really kind of mandated, I guess in that framework, is in the tech layer, which is, there’s just like five boxes there. So, yeah, but just back to the point of can you kind of do without things? In the framework itself, yes, absolutely. The idea is not that you need to do all of those things and tick all of those boxes. Quite the opposite, actually. It’s like a good strategy should be about being carefully selective about the things that you focus on and all of the stuff that you don’t focus on. You know, you’re making tradeoffs all the time. And what the mobile growth stack hopefully does for marketers is to help them to consider everything, the whole, and then kind of zero in on the things which make the most sense for them at that particular period based on their goals. So the idea is not that you have to do all of the things in the mobile growth stack, that would be definitely overwhelming.
Peggy Anne Salz: Okay.
John Koetsier: I love that. I love that, and I’m super thankful that you made that clarification because I’ve had this impression sometimes from marketers. They needed to fill every box or they weren’t doing their job, and they simply didn’t have enough people to manage that quantity of tools and even know what was going on in all those tools.
I also love that you said it’s strategic, not tactical. Figure out what pieces you need and what you don’t need, and use those. So that’s really, really cool.
By the way, I just did a quick check while Peggy was talking, while Andy was talking, and yes, Machine Zone was acquired by AppLovin, and yes, that did surprise me. But anyways, we’ve got to end pretty quickly here. We know you’ve got a hard stop to go to, but we have this one question about retention marketing, and that is, hey, what’s the kind of uplift, what’s the kind of bonus that I get when I focus on retention marketing, when I focus not just on acquisition, but I focus on, hey, I’ve got these people how do I maximize the value of the users that I’ve gotten?
What kind of lift have you seen when people put a significant focus on retention marketing?
Andy Carvell: Yeah, that’s a great question, John, and I’m going to give you a vague answer on it because it really does depend, right? But there’s always upside to be had, and it requires quite some effort. There needs to be like a concerted decision by the company really, a company level, hey look, we want to really work on this. Ideally it’s an effort that never really ends, but certainly you need to put some effort in and over a considerable period, like retention’s a slow moving metric.
Unless you are doing something drastic with your acquisition actually, which brings in a completely differen cohort of users with a completely different retention profile, that’s actually the quickest way to influence your retention, by the way.
But assuming that you’re already doing acquisition that’s bringing in good quality users and you want to then improve the experience and the interaction for those users to keep them around for longer. In my experience, it’s possible over a 6 to 12 month period to increase that retention, let’s say month one retention by several percentage points, which maybe doesn’t sound like a lot, but if you can increase retention by five percentage points, that compounds over time in terms of, you know, could mean millions of extra active users over the year, for example. On over two, three years because you’re keeping that retention hopefully over the long term can really make a huge step change to the business and the whole dynamics of the business.
John Koetsier: It plugs a few holes in the bottom of that bucket, which is critical when you’re looking at long term growth. Absolutely.
Andy Carvell: So, yeah, I’d say five percentage points would be a really good win over a 6 to 12 month period. Depends on where you start from as well, you know, maybe it’s definitely possible to get more than that, maybe ten? Ten percentage points would be knocking it completely out of the park, right?
So this is important for people to understand because if they’re looking to increase retention by 20%, forget it. Like you’re not going to do that unless it’s at 0%…
Peggy Anne Salz: Or unless you’ve reached marketing Nirvana, which will bring me to my closing question, Andy. I mean, it’s great to have you here, you share so much, you give us some takeaways, but what’s the biggest learning for you in recent times? First of all, you have these clients. Secondly, you were on a trek to Nepal, so you’ve had this experience, a lot of experiences to share, a lot of learnings to share. What would it be?
Andy Carvell: Okay, I’m going to pick two because I’m always on the fence on picking favorites, but, so first of all with COVID-19 the whole coronavirus situation, it’s been a great learning for me to see that certainly like my company’s team is able to work remotely very effectively, but also I see that generally, most people are managing that. And it’s interesting to see how companies that were really resistant to that in the past have actually found like, wow, this actually works just fine, which makes everyone wonder why we all need to go to the office all the time.
And secondly, on a more retention note, even though it’s not specifically related to retention, I think it comes back to me to this organizational fear and stakeholder management question, as well as helping our clients with their strategy and tactics of improving retention. I’ve realized lately that actually sometimes our job as consultants at Phiture is also to help empower the team that we’re working with to evangelize growth and to unblock stakeholders within their company, to actually unblock that team to do the great work that they’re capable of.
And so I’m on this kind of evangelization mission lately, really looking at how can I help our clients to really sell the idea of growth within their company. That’s why I keep talking about it because I’ve seen it lately a few instances where working with a really great smart team who seems like really actually they’re quite restricted by what they can do because they have to run things past the legal team, the product team, the design team, and it’s like, okay, can we get everyone talking together and establish a kind of a rules of engagement that allows this team to do great work and great growth experiments and drive their metrics up.
John Koetsier: Wonderful, wonderful.
Peggy Anne Salz: That’s incredible, that absolutely is. Andy, it’s been wonderful to have you on Retention Masterclass, sharing all of those ideas and sharing this last personal mission. I think, John, gives us something to think about. You know, we are able to pull this together. It’s really about evangelizing, believing, experimentation. Just inspiring. Thank you so much, Andy.
Andy Carvell: Thanks for having me. It’s been wonderful.
John Koetsier: It’s been a real pleasure to have you. It’s our first time meeting for you and me, so real pleasure to get to know you a little bit. For all our listeners, whatever platform you’re on, please like, subscribe, share, comment, all the above. If you love this podcast, please rate it, review it. That’d be a massive help.
Peggy Anne Salz: Absolutely. And until next time, and we’re going to have many more guests, some are coming, some almost as amazing as you, Andy, I have to say. Until next time, stay well, keep safe. And I’m Peggy Anne Salz with Retention Masterclass.
John Koetsier: Wonderful. Thank you, Peggy. I’m John Koetsier, have a great day!