Under the Radar

126: Retention


00:00:00   Welcome to Under the Radar, a show about independent iOS app development. I'm Marco Arment.

00:00:05   And I'm David Smith. Under the Radar is never longer than 30 minutes, so let's get started.

00:00:10   So today I thought it would be fun to unpack one of the metrics that we can use to track our apps.

00:00:16   So it's no surprise I love metrics, I love numbers, and the particular one that I want to,

00:00:21   I think it would be fun to talk about today is retention, which we'll talk a lot about what that actually is

00:00:28   and kind of unpack it, but it's one of the analytics that is, if you have an app on the App Store,

00:00:33   you have access to, because it's just like one of the built-in ones in iTunes App Analytics.

00:00:38   And I was recently digging through some analytics related to my Sleepless Plus big update and launch,

00:00:45   and I came to a realization that I think retention is probably the most important metric

00:00:55   for the overall health and vitality and longevity of an app that we have available to us.

00:01:01   That in general, it's a great measure of how your app is performing, both from a usability perspective,

00:01:08   a stability perspective, how useful it is. There are so many things that kind of are attached to retention

00:01:15   that it seems like a great metric to kind of dive into and make sure that we understand,

00:01:20   because at its core, retention is about how often people are coming back to your application,

00:01:28   how often people are coming back to it and launching it on subsequent days after they first install it.

00:01:33   And so I think it's so easy to initially focus on things like download numbers,

00:01:39   which are sort of useful and sort of important insofar as if that number is too small,

00:01:45   it obviously will have an impact on how viable an app is.

00:01:51   But if your retention numbers are really small, so you could have huge download numbers,

00:01:57   you could have 10,000 people a day download your app, but if only 2% of them ever launched it ever again,

00:02:03   you're not in a great place. You're not building a viable business if you have lots of people come in

00:02:08   and then immediately leave disappointed. That's only viable if you're building a pyramid scheme or something.

00:02:13   But in general, you want to have people come and stick around.

00:02:16   You want people to download your app and to continue using it.

00:02:19   If you have an advertising-based business model or an in-app purchase-based business model,

00:02:24   you need them to continue to be coming back to the app in order for those to work.

00:02:29   Maybe if you have an upfront business model, it matters a little less,

00:02:32   but still happy customers coming back to the app on a regular basis seems like it's always a good thing.

00:02:38   It seems like a worthwhile thing to unpack to make sure that everyone understands what it is

00:02:44   and then some things we can do about it to improve it or make our retention better seems like a good thing.

00:02:52   - It's part of what business people talk about that we all kind of zone out or make fun of at first.

00:02:58   When we first hear it, we talk about the sales funnel. It's one of those kinds of things where they talk about,

00:03:03   "Oh, well, you have this many people who are out there and this many of them are going to learn about your app

00:03:09   and then once they learn about it, this many people are going to actually go seek it out

00:03:13   and then this many people are going to download it and this many people are going to actually ever buy it

00:03:17   or use it or whatever else and this gets smaller and smaller and smaller as you go."

00:03:22   If you're trying to increase the parts of your app that are valuable to you,

00:03:28   like if you just sell the app up front for a flat price, then what you want people to do is buy the app.

00:03:37   That is your end goal financially. You might want them to use it for other reasons,

00:03:42   but for the most part, your job is done when they buy the app.

00:03:45   You kind of stop caring about the funnel or the retention after that point.

00:03:51   It's a lot less of your problem after that.

00:03:53   Whereas if you have a business model or an app model or some kind of goal that requires ongoing usage

00:03:59   or that you don't maybe make money if people stop using the app.

00:04:04   For instance, if you're based in advertising as the source of your money, then you kind of need more than that.

00:04:11   Your funnel needs to continue past the point of downloading the app.

00:04:15   You also need to start caring about how many people stick around after X amount of time.

00:04:20   That can also help judge things like how much you should plan to spend for search ads

00:04:26   or for other types of advertisement for your app.

00:04:28   You have to know what's the average lifetime value of a customer.

00:04:34   What's the average lifetime value of an installation?

00:04:36   If you know that after a certain number of weeks or months that X percent of people will have just given up on the app

00:04:44   and X percent will still be using it, then that can also help you figure out what should I be doing.

00:04:49   What should I be willing to pay for a new customer that comes into my app?

00:04:54   There's all sorts of reasons why you tend to need to know this stuff for your business.

00:04:59   You can get away ignoring it for a long time.

00:05:02   I'm guilty as anybody. You, David, you are very good at metrics and tracking these things.

00:05:09   I really am not. I'm slowly getting less terrible at it.

00:05:14   But for most of my business, most of the time, I don't do nearly enough tracking of metrics or trying to optimize for certain metrics.

00:05:22   It's kind of tempting when you see retention numbers.

00:05:28   Then one way to make more money is, of course, to just get new customers in.

00:05:33   But another way is to try to get more engagement, more usage, more retention out of the customers that you already got in the first place

00:05:40   or that you're getting in every day.

00:05:42   There's a lot of potential value here where you don't have to be a sleazy business person to make use of these tools in a useful way.

00:05:55   I think, too, the thing that is so essential about retention is that it allows you to continue to...

00:06:06   It's incredibly difficult to get a customer in the first place.

00:06:10   We're going to get into it in a little minute. We're going to talk about how depressing retention numbers can be.

00:06:16   You want really real depression, go look at your App Store impressions versus conversion rate.

00:06:22   It is astonishing how hard it is to get a customer to download your app in the first place.

00:06:29   But once you've gotten them, they have at least made a choice.

00:06:34   Here's a person somewhere in the world who has looked at your App Store page and hit the "Get" button, and they have downloaded your app.

00:06:42   You're at this tremendous advantage with that person.

00:06:47   Trying to then sell them on your app and make them want to continue using it is such an easier problem to solve than the first one.

00:06:58   Obviously, ideally, you want to be optimizing both your acquisition of new customers as well as your retention of your existing ones.

00:07:05   But I feel, as a developer, much more in control of retention than I do of acquisition.

00:07:10   Because with retention, in many ways, it is, "Am I making a quality app that is compelling and useful and interesting and creates a desire for somebody to want to remember it and to come back to it?"

00:07:23   I can see if changes I make in the app are causing that change.

00:07:29   In the example of C++, what was really gratifying for me is I spent a lot of time and effort working on this update.

00:07:37   I put it out, and my retention numbers roughly doubled from what they were before I did the update, which is huge.

00:07:44   It's this great reinforcement, though, that all of the engineering work, all of the stuff that so often we end up on the show talking about how it's important for us developers to not just get stuck in developer land.

00:07:58   There's other business and marketing things that are important for our business, and that is absolutely true.

00:08:03   But this is one area where raw engineering and user design and those types of things come back to play in the business side of things.

00:08:14   By making my app more compelling and understandable and easier to use, I was able to increase retention, which is a really nice virtuous cycle to end up in.

00:08:25   The best place to dive into before we get into some of the other aspects of retention is probably just to define what it is and to clarify how you get these numbers and a little bit of how to understand them.

00:08:38   At its simplest, retention is usually measured in it as a percentage.

00:08:45   It is the percentage of users who have launched your app since they installed it, and it's typically measured in one-day, two-day, three-day, four-day, five-day measures.

00:09:01   One-day retention is the number of users who launched your app the day after they installed it.

00:09:08   A seven-day retention is the percentage of users who launched your app seven days after they installed it, and usually you tend to go up to maybe about 30 days.

00:09:17   And then after that, I think the measure tends to be not as interesting because at that point, at least in my experience, it becomes very stable.

00:09:25   Your app tends to have this stable point where after--and it's probably, for me, I found that after about a week,

00:09:33   if I've hooked a user who they found the app useful and interesting for more than about seven to ten days,

00:09:41   they're probably going to continue using it in the long term.

00:09:45   And so that's at its core what retention is.

00:09:48   You find it by going into iTunes Connect, go to App Analytics, and then the rightmost tab is called Retention.

00:09:55   And there they have a bunch of different graphs.

00:09:57   But the main graph--or I guess it's a table--is they have this kind of triangle table which shows your last 30 days in the App Store,

00:10:09   and then for each of those days, the relevant retention numbers for each of those days, as well as the overall retention.

00:10:18   And so what you'll tend to see is something where your day one retention is going to be your best, typically.

00:10:24   Very few apps would need to have some kind of cyclical nature to it if this wasn't the case,

00:10:29   but typically your best retention is going to be the next day, the day after that you first installed it,

00:10:36   is probably the most likely day that they'll remember and come back and install it again.

00:10:40   And then you'll see this fall off from there.

00:10:43   These numbers, in my experience, are sort of brutal because when you get this--

00:10:51   there's this part of me that wants to think that someone downloads the app, they're excited about it, they want to use it,

00:10:57   and then you see that after a week, 10% of people are still using it, or 5% of people, or 20% of people.

00:11:04   These numbers are relatively small.

00:11:06   For every 100 people who download my app, a month later, only four of them or five of them are still using it.

00:11:13   And that can be kind of brutal.

00:11:15   Obviously, these numbers could be better. It'd be awesome if your retentions were 80%, 90%, that would be great,

00:11:22   but more likely than not, don't be discouraged if these numbers are lower than you would hope for them to be.

00:11:29   Because really, most of what you're doing is--

00:11:33   the actual number is a good measure of how interesting and compelling the app is.

00:11:39   What's probably more important is if you make changes that you hope will affect the retention,

00:11:45   that you can then compare them over time and see the number go up or down as a result.

00:11:51   So don't get too discouraged if they seem kind of brutal to start with.

00:11:56   Yeah, because mine are terrible. I don't really know what's normal here.

00:12:01   I'd actually be curious to know what is the industry average for some of these numbers.

00:12:05   But the highest number on my entire chart is about 43%. That's on day one.

00:12:11   And then it seems like the 30-day average that I'm running for Overcast is between 10 and 20%, depending on the day.

00:12:21   But I'm hovering between 10 and 20. And that's not great. That's a horrible thing to see.

00:12:29   But I look at how I use my phone just personally, and it makes total sense to me.

00:12:35   I download tons of apps that people recommend. And then for whatever reason, for some reason or another, some of them stick.

00:12:41   Most of them don't. And that's fine.

00:12:45   If you look at these numbers and you see, "Wow, I have 15% of people actually using this thing after a month,"

00:12:51   that sounds terrible. But I think that reflects just the reality of there being a lot of apps out there.

00:12:57   And a lot of people willing to explore new things, but not necessarily, not everything's going to stick.

00:13:03   And 15% sounds terrible as a ratio of whether your app sticks or not.

00:13:09   But that actually, when I think about the entire market as a whole, when I think about what percentage of apps I download I actually stick with,

00:13:17   that actually seems kind of in line with that.

00:13:19   Again, I do wish I had more points of comparison for what's the industry average.

00:13:25   I think this also will depend a lot on whether your app is free or paid.

00:13:29   One of the advantages of free apps that I've enjoyed quite a bit with Overcast that I never had with Instapaper was that

00:13:37   the raw download numbers that you get from the app store are way higher.

00:13:42   Oftentimes like 10 times higher than what you get from a paid app of how many people are willing to install this thing for free versus how many do it when there's a pay up front price.

00:13:52   But I think the downside of that is like, yeah, you lowered the barrier for them to come in the door.

00:13:59   But then I bet, not having any information on this, I bet that the retention rate over time of free apps is noticeably lower than that of paid apps.

00:14:08   Because the paid apps, it's like you kind of had to get them over that hurdle of believing in your app up front.

00:14:13   And so more people who download it will be more committed to the app that they paid for, I bet, versus the one that they just took a chance on because it was free.

00:14:22   Maybe. I have a suspicion that the short term retention may be slightly better.

00:14:28   But I think the same forces that cause retention to drop off so dramatically will immediately come to play eventually.

00:14:36   That initial sense of like, oh, man, I paid for this and I want to use it, I think will fall off once you get to the point of like, it's just, is it really compelling and useful enough for them to come back to it?

00:14:48   Because I think after a couple of days, you're going to kind of forget and it's kind of a sunk cost at that point.

00:14:52   And you're just going to sort of decide if, you know, is it really, you're not going to sort of just be using it out of penance for a mistake you feel like you made in the past.

00:15:03   But I think I imagine you're right insofar as the initial retention is probably slightly better.

00:15:08   Then, I mean, honestly, in some ways, zero day retention is probably way better for paid apps than free apps in the sense that I'm sure a lot of people download free apps and never never launched them once.

00:15:20   And so that side of things is, you know, certainly is going to be better on the paid side.

00:15:25   Honestly, I think your numbers are fairly good in terms of I don't I have, I only have my my data sets to compare them to, but I have more apps than you.

00:15:34   And I think having a sort of having a next day retention in somewhere in the 40 to 50% I think is very good.

00:15:42   And then your 30 day if you can, you know, somewhere between 15 and 20% is also very good as far as I can tell that it seems like anything above that would be really, really impressive.

00:15:53   And the reality is, I think, is if your numbers are in that range, that if you can start between 40 and 50 and end somewhere between like 15 and 20, I think there's a good chance you can make a viable sort of business out of that app.

00:16:06   That that's a very you're going to be growing your user base over time, most likely because really, this is kind of one of those funny things where if you want to think about like your daily active user number, which isn't like another one of the metrics.

00:16:20   That number in order for that number to grow, you need both new users coming in. But mostly what you need is over time to build out your like sustaining user base.

00:16:33   Because, you know, you if you want to say your own say you say you're getting 10,000 new users a day.

00:16:40   You know, you're never going to get to say 100,000 daily active users unless your attention is really good because you know, the new people coming in is just not going to be able to sort of add up unless you can have a sustained, you know, growth over time because your attention is good.

00:16:56   And so that for every 10,000 that come in, you have 2000 still using it a month later.

00:17:02   I would say too, like, you know, if your app has any kind of like service component where you can where you track like number of user accounts versus some concept of active users.

00:17:10   Like for a while I didn't track active users. I only kept tracked, you know, like the number of user accounts that were in my database, which it really is not that useful of a metric.

00:17:18   It's really it doesn't show anything about how active people are. So I have this big spreadsheet where I track like where I just take like, you know, weekly analytics and just to kind of get an idea of how my business is doing.

00:17:28   I'm so proud of you. Yeah, weekly. Yeah, it's I have like, you know, five metrics.

00:17:34   It's better than zero metrics. It's way better than zero metrics. I had zero for a long time.

00:17:40   But one of the things that I started tracking about a year ago was monthly active users. And of course, you know, there's arguments about what that means.

00:17:48   You know, but once you start tracking that, like I realized very quickly, like, okay, first of all, I actually have something like a quarter of the of the users that are in my database about a quarter of them are actually active over a month.

00:17:59   And so like, you know, it kind of it kind of like put some perspective like it kind of cuts you down his eyes a little bit like, oh, okay, I don't have, you know, X million users or anything.

00:18:09   It's not it's much smaller than that. But it is useful to see like, okay, this is how many people are actually like actively using the app at all.

00:18:18   And so it does, you know, create incentives to try to improve that number, try to increase the percentage of users that are active, which we'll get to after the ad break, I think.

00:18:30   But it also kind of shows you like, just more accurate data when you're doing things like, like when I was trying to figure out like how much do I pay for search ads and everything.

00:18:39   It doesn't really matter what percentage of people convert to user accounts, what actually matters is what percentage of people convert to active to monthly active users.

00:18:47   And that is a little harder to figure, but a far more accurate representation of like, kind of trying to give you an idea of like what you should pay for installation.

00:18:57   Anyway, we are brought to you this week by Squarespace. Make your next move with Squarespace.

00:19:02   Squarespace lets you easily create a website for your next idea with a unique domain, award winning templates and more.

00:19:09   Whether you want to create an online store, a portfolio, a blog, even a podcast, you can do so much with Squarespace.

00:19:14   It's the all in one platform that lets you do pretty much anything you want to do with a website these days.

00:19:19   And there's nothing to install, no patches, no upgrades. You don't have to worry about any of that stuff. Squarespace has it covered for you.

00:19:25   They have award winning 24/7 customer support if you need any help. And you can quickly and easily grab a unique domain name for your idea.

00:19:32   All of these templates that they let you choose from are designed by professional designers. They are beautiful.

00:19:37   And they really help you show off your great ideas with very little effort.

00:19:41   Squarespace plans start at just $12 a month. You can start a trial with no credit card required by going to squarespace.com/radar.

00:19:49   When you decide to sign up, use offer code radar to get 10% off your first purchase of a website or domain and to show your support for under the radar.

00:19:56   Once again, that is squarespace.com/radar and use code radar for 10% off your first purchase.

00:20:03   We thank Squarespace for their support. Squarespace, make your next move, make your next website.

00:20:08   So the rest place, it seems to sort of wrap up our discussion, seems to be to talk about A on the negative side of trying to think through reasons why people might want to leave.

00:20:17   Reasons why you may not be retaining people. And then on the flip side, things that you can do to potentially help people want to stay and improve your retention as a result.

00:20:29   And the first thing I think about is so often we have no idea why people leave our apps. It's a bit of a mystery in terms of, you can imagine this is why most big companies, if you want to cancel your account,

00:20:41   they give you a three question survey where they're trying to ask you, "Why you go? Why are you sad?" That's something that they do because it's so hard to collect this data.

00:20:51   But it's certainly something that I think you can speculate on. It's something that you can look at your customer support, emails, or however you collect your customer support.

00:21:00   You can look at that and that probably gives you an indication. Tragically, probably another really good place to find this information if you're looking for it is One Star Reviews in the App Store.

00:21:09   There's a good chance that the reasons people are upset is a subset of the reasons why people might want to go.

00:21:16   And you can put that all together to get a sense of why people might be wanting to leave. And some obvious reasons off the top of my head are things like, well, the app wasn't useful enough.

00:21:27   It didn't actually provide the utility they were hoping it would do. It could have been hard to understand. It may have been an app where they wanted to use it for whatever purpose and utility that it provides and ostensibly shows, but they didn't understand it.

00:21:46   It was confusing. There was something that hit some wall somewhere where suddenly it wasn't able to give them that usefulness.

00:21:55   There's also probably a saying, some types of apps, this isn't really a useful metric. Like if you have a seasonal app or a one-time use app, I know there are certain applications that I download and use for a specific need that I have, and then I don't have that need going forward.

00:22:13   There's some file processing thing that I need to do or a conversion between two data types. There are certain types of apps that aren't expected to be used on a regular basis.

00:22:25   Certainly the best kinds of apps for building a sustainable business are the kinds that I think are used on a more regular basis, but that's just worth saying that if you have one of those types of apps, understand that this kind of analysis isn't really useful.

00:22:37   But in general, I think it's a useful thing to come at it from why people might not want to use it. It's important, I think, to try and come at your application every now and then with a fresh mind.

00:22:49   I always try before I do a big update. I reset a phone, I have a completely fresh installation of it, and I start to try and use it and see if there's pain points.

00:23:00   Are there things in that initial run that would cause someone to go away? And ask yourself, what are the reasons why someone might want to not keep using it?

00:23:09   Because if the app is just not useful for that person or doesn't really do what they want to do, there's nothing you can do to really grab them in.

00:23:16   But if there is something that you can think of, like is there a reason why someone might, you know, whatever, in my sleep tracker, if you don't give the app health permissions, the app is completely useless for the most part.

00:23:28   So if that's the case, then I need to be clear and upfront and encouraging around that and to try and encourage to make sure that that's not a reason people are leaving.

00:23:37   But I need to think through in the first case of what are the reasons that people might want to leave.

00:23:42   Oh yeah, like I faced this problem when Overcast first launched. It required an account, you know, for like the website syncing and everything.

00:23:48   And so I required people to give me an email address and a password. And requiring an account is such a barrier for a lot of people.

00:23:58   And you know, and I know it myself too, again with my personal habits. Like I know it, like if an app opens up and it requires me to create an account or to log in with Facebook or something, I'm like, no, I don't want to do that.

00:24:08   So in most cases I will just abandon it because I just don't care enough to do that. And you know, now I very quickly added a, like you know, basically proceed without an account option, which actually just generates an anonymous account on the backend for sync purposes.

00:24:24   But you know, there was one button to just say, alright, forget this, I don't want an account. And now like my, one of the metrics I track is new users per download of the app. And that has been pretty steady at around 90%.

00:24:40   Which is what I consider really good. Because I know like with Instapaper, where I never had an accountless option, my rate was way lower than that. It was probably closer to half. But now it's 90% because there was this barrier that a lot of people just don't want to give you an account or don't want to make an account or don't want to set it up or whatever.

00:24:55   And if you give them an easy way to just not do that, skip that, then you will lose fewer of them at that stage.

00:25:02   And any kind of barrier to using or continuing to use your app, I think it's worth investigating and considering like, is there a way I can either make this easier, make it less of a barrier, or get rid of it completely?

00:25:16   Yeah, because I think those are the reasons that like, it's important to be honest with yourself. And like, I know this for my own habits that like, it's so easy, I think, as I'm developing something, because I have such knowledge about it, for like, of course, I'd create an account. But then I think of like, my actual habits, it's like, no, I wouldn't. Like anytime I install, I open an app, and the first thing it asks me to do is create an account. Like, I just, I just ignore the app and delete it. Like, I don't want to go down that road nine times out of 10. And so it's important to be to be honest with yourself.

00:25:45   And in your own habits and understand that, like you or perhaps even more importantly, it's like going through the exercise of, you know, the need the initial experience test with people in your life, you know, friends and family, people who don't know your app, especially people who aren't particularly technically savvy, like you can get great and great feedback about what what what was confusing, what was not straightforward, what would be a reason why the what what questions do they ask that lets you know why people are, you know, maybe not used, right?

00:26:14   You know, maybe not used, because the reality is that is, you know, like, your, your, your users are not you, your users are very unlikely to be technically incredibly, technically savvy iOS engineers, like maybe if you're making a hyper focused iOS engineering app, like, great, good for you. But the majority of your apps, you want to be general purpose, you want to have apps that don't require people to really have deep understandings of what's going on that they can just get started. And that's not to say that you need to, like dumb down your app or somehow be condescending about that.

00:26:43   But it's just understanding where their where users are coming from allows you to make an app that directly suits their needs more directly.

00:26:52   Alright, let's talk about spam.

00:27:17   And then if you don't use the app for a week, you find out why they actually want to send you notifications. Because they'll send a little thing like, hey, turns out you haven't used the app in a while, you know, hey, you want to you want to come check it out? Maybe it's kind of cool.

00:27:29   Or they'll send you an email if they have your email address, they'll email you be like, hey, we haven't seen you recently, why don't you come back? Look at all this cool new stuff.

00:27:48   It looks like somebody who has really bad social skills and can't take a hint.

00:28:13   Great, you've increased your retention number, but you haven't increased your like, user satisfaction. Tim Cook's favorite metric, like that's you're not increasing it by very, it's very unlikely someone for you stopped using your app because they forgot about it.

00:28:28   You know, they stopped using your app because it wasn't compelling. You know, the reason you want to help if you want to help people stay and increase your utility, you need to make the app compelling and useful. And you need to have an onboarding process that encourages them to want to come back and find ways to design recurring interest in your app, you know, like making an area of the app where someone wants might want to come back to on a regular basis.

00:28:52   And that might be built into the type of app it is like one of the things I know is useful for me is with a health and fitness app. There's a compelling interest in coming back to it every day because your steps change every day.

00:29:04   So you want to come back to it on a regular basis. But if that's not the kind of app like in a podcast app, it's like for you, I think of the area like the browse area in overcast is a great place to try and create interest where it's like, huh, I want to listen to something and having like their first thought be Oh, let me go browse that catalog.

00:29:21   And if you can create areas in your app where there's that recurring interest, you have a much better chance of building genuine retention like where people are honestly wanting to come back to your app. And like that's probably the most compelling and useful way to to to approach it.

00:29:36   Thanks for listening, everybody. And we'll talk to you next week.

00:29:39   Bye.