Super Charge Your Mobile App Onboarding & Retention

Engage 2017

Kathy M. Chao, Director Global Marketing

Video Transcript:

BabyCenter, who’s going to talk to you about onboarding and retention through the eyes of BabyCenter. Welcome, Kathy. Thank you. You’re welcome. Thanks. Hi, everyone. My name is Kathy Chao, and I am here representing the BabyCenter team. So I am based out in San Francisco, which is our headquarters. And just by way of introduction, I actually went to school down the street in Medford Somerville, so it’s kind of great to be back in Boston. It’s been a while. So what I’m going to walk you through today is about 25 minutes’ worth of content. And I’d love to save about the last five, 10 minutes for questions. So if you can hold on to your questions, that would be awesome. So here’s what we’re going to discuss. A quick intro to BabyCenter. Some of you might be saying, who is BabyCenter? I might have heard about you guys, I’m not sure. I also want to talk about team structure. I think [? Lou ?] mentioned it earlier today. I think it’s pretty fascinating. I’ll tell you about how we work as a team, some of the considerations, and also I’m going to be focusing very specifically about our app. So not just mobile engagement sort of universally, but very specifically our app. We’ll talk about how we achieved success so quickly on the Localytics platform. And then I’m going to do a deep dive on app registration onboarding, as well as retention. I’m hoping that you guys can walk away with some very clear action steps that you can apply to your own business and your own challenges. So who is BabyCenter? We’re a global brand. We are reaching currently one in five moms around the world. So it’s pretty incredible. We’re the number one pregnancy and parenting resource. So we are in 14 markets, 28 different languages. So essentially, we’re a global publisher. So with that comes a lot of challenges in terms of taking something that’s so universal like pregnancy, or even parenting, and really trying to make it very specific for specific cultures and specific nuances in different geographies. Our company’s vision– and I’m bringing this up here because I think it’s going to tie in later when I start talking about our messaging programs– is our North Star, it’s we believe in a world of healthy pregnancies, thriving children, and confident parents. And that’s super motivating for myself. It’s also motivating for all the employees who work there. I mean, this is really something you know, we feel like we are contributing to the world and trying to do good here. We just want to give information so people can make the best choices for themselves and their families. So there’s four– let’s see– there are four strategic imperatives that we think about at BabyCenter. The first one is relevance. So this is super important when it comes to targeting and some of the specific push programs I’m going to talk about. So what is relevant for somebody who is four weeks pregnant– they just find out they’re pregnant– versus somebody who is 24 weeks pregnant is very different, right? The user journey is extremely nuanced. And we need to make sure that we are reaching her with the right message at that exact right time. If not the right week, the right day of week. Second is the experience. So BabyCenter was founded about 20 years ago. And we started with this idea of mass personalization. When you register today, we only ask for one piece of information, your email address, and also– if you’re pregnant– your expected due date, or your child’s birth date. And based on that one key piece of information, we’re able to spawn a very extensive, sophisticated series of messages outbound to our users. Personalization. So I think about this in a few different flavors. There’s the explicit. So for instance, I ask you at registration what’s your due date. It could be implicit, just based on, you know, what we’re seeing in terms of behavior on the site, or in our app. It could also be– you know, what we’re moving into now is more inferred data. So what I mean by that is I can look at a user’s profile and see how many children they have in their profile, and I can infer whether or not it’s their first pregnancy, or maybe they have three children listed. OK, maybe you have three kids. And now we’re moving more into the predictive, you know, where we’re really using data science to help us get more personalized, because you can’t just give me week by week information anymore. There’s a lot of nuances within that pregnancy. And finally, solutions. This is sort of how we wrap it up for our partners and our clients. So we always have to keep that in the back of our mind, right? It’s the monetization. So some relevant background now. Now I’m going to get into the Localytics piece of this. So we have four products integrated into our app. We’ve got app analytics, we have push, we have in-app, and we also have Connector, which is email responses and Localytics. We’re not quite live with that, but we will be in Q1 of 2018. So how we function right now as an organization is that we are sort of a hybrid, I think, of a couple of the models that were presented before. We have a cross-functional team that have different managers, essentially, but the same shared goal. So what this allows us to do is really talk with the same language about, OK, what is the KPI, what are we trying to do for each campaign, each strategy. So we have an app product manager, we have a dev team, iOS and Android. We’ll have QA, marketing, which is my team, and then we’ll also have the analytics team very closely working together on all these projects. And I’m just going to focus a little bit on the marketing side. So the things that we are owning right now, we’re aligning on strategy. So yes, we’re a global publisher, and so there’s a lot of voices. Editorial team wants to be involved, tech team wants to be involved. At the end of the day, marking is really the centralized hub, if you will, for determining what the roadmap is going to be in terms of consumer-facing messaging. We own the creative development. So, you know, while the editorial team might be experts in actually writing that content, we’re experts in packaging it, and how we’re going to promote it, and how we’re actually going to get somebody to take action on it. Next we’ve got technical troubleshooting. And this has been actually something I’ll talk a little bit more about, as well. This has been huge for us. We’ve been on the Localytics platform for about one year now. And if we did not have the skill set on the team to really dig deep and understand how audience targeting worked, we probably would not be getting as much out of the platform, because the tool’s made as a WYSIWYG, right? This dropdown, this dropdown. But if you don’t truly understand how your audience is being created, you could message people at the wrong time with the wrong thing. And for us, that would be a disaster, because we’re so focused on that relevant message at the right time. And then also, we are driving the testing and learning and the execution. So that’s basically run by a couple people on the team. We have somebody who is 50% of their time. She also managed our email channel, which, if anybody out there is familiar with BabyCenter, you know we have a very robust email platform. So we’ve taken a lot of our learnings about what you have to do to create a really strong email program and high engagement, and taking that over to the push and app side. A couple of considerations. We actually started with the point solution Urban Airship– I don’t know if anybody in the room used them or looked into them– in 2016. And we were with them for a whole year, and we never sent one campaign. So that was definitely in the OK, we tried that, and that didn’t work. It didn’t work for a couple of reasons. One, they didn’t have marketing involved in the decision-making process. It was purely a tech-led decision. So what happens is one side of the house said, yeah, that sounds great. They hit all the marks in terms of got the right APIs, all the things that we want in terms of SDK. When it actually came to who was going to be looking at creating the campaigns, evaluating the campaigns, they fell short. They couldn’t actually get us to that exact due date that I talked about. We need to know day one of week four. They couldn’t help us calculate that. Other considerations. I think I talked about this a couple of times already, but success for us is when the campaigns are highly personalized. So our goal is– you know, we have 13,000 pieces of content. It’s about discovery. We want to make sure that people can discover that content, and actually take action, find it useful. And the last thing here is we have over 500 campaigns right now in the Localytics tool. And this includes the liquid template for our global footprint. So if we’re thinking about an edit, it’s not something we can really make on the fly. We have to be considerate of how it’s going to affect other campaigns. And you know, you don’t want somebody spending the whole day updating campaigns. That’s just not an effective or efficient use of time. So these are all things we kind of keep in mind. Sorry. Let me see. Apps funnel. So this is a very simplified look at some of the key metrics for our app, which is a pregnancy and parenting app. So at the top of the funnel– I think anybody here who does acquisition marketing is going to know this– it’s all about discovery. You know, how are people discovering your app? Organically, paid, social, word of mouth? Whatever it is, that’s the very top of your funnel. Install conversion. We still care about that. I think if you talk to a lot of practitioners out there, they’ll say, don’t worry about conversion, it’s all about downstream. Well, second step in your funnel, you still need to have people converting, right, at that install. So the things I’m going to focus on in my use case and my deep drive is pregnancy registration. So for us, we only have a free product. It’s reg-gated, so we have to make sure that people can get into the product. They might install your app, but if they don’t go the next step and actually install it and start using it, I spent that money for nothing, right, to get somebody there. And then for us– this is very nuanced for the BabyCenter business– currently we’re looking at retention a couple of ways. We’re talking about it as short-term retention, so day two to 14. So what percentage of people come back to the app two to 14 days after they’ve installed it. And this is, for me, all about velocity, right? It’s the faster we can get somebody in to the app, the more likely they’re going to get hooked and really engaged. And it’s probably similar for a lot of your products. And then we also look at first trimester to second trimester. So what we’ve seen is that we have a couple of natural drops. From the time somebody registers on our site and they tell us they’re pregnant– and what’s really amazing is, at BabyCenter, people often tell us first. They don’t tell their partners, they don’t call their doctors. They probably just took a pregnancy test and they’re saying, oh my god, I think I’m pregnant, question mark, and they go to Google. They find BabyCenter. And immediately, we’ve got that emotional engagement with them. And so what we really want to do is make sure that they carry over from first trimester into second trimester. We see that they have a much longer engagement with us down the road. So a few thoughts on keys to our success, specifically I’ve mentioned this, but I’m going to kind of keep bringing this up because I’ve found it’s been so important for us is that we really have dug deep to understand the deep audience targeting. And that’s been critical. If we didn’t do that, I think that we probably would be saying to ourselves, oh, great, we have a high volume of campaigns, but we wouldn’t really know if we were accurately impacting people the way we really want to be. So I encourage you guys, if you don’t have somebody who’s technical– but not necessarily from an app development technical, but technical to, like, kind of dig in and understand what those dropdowns mean and who exactly you’re talking about, it’s a really– it’s an important first step that should not be overlooked. Second, you know, we’ve been partnering with Localytics from day one. Weekly meetings with them. My team works with Bryan Dunn’s team in product to raise issues, to challenge them, to push them. And that’s really helped, too, so that we can get some visibility into some things that, you know, they’re not just for our use cases, but multiple– you know, you can imagine how they can be used for different clients. Rapid test and learn. So as soon as anything new comes online from Localytics, we’ll try it right away, be that animated GIFs, liquid templates. We felt like we were kind of behind the eight ball in terms of getting push and in-app started. So we wanted to test as much as we could, not for necessarily incremental learning, but to make big impacts. Like, we want to understand emojis. How much can our audience tolerate? And we found they can tolerate a lot. They probably want more of it, in fact. So it’s stuff like that. It’s thinking about the A/B message testing, like all the classic things you would take from any other platform, make sure you bring it over to Localytics, as well. Fourth here is ruthless prioritization. So we are– I mean, I’m probably preaching to the choir here, but this is something you have to do constantly, right? You start off a month– or maybe you have a quarterly roadmap. For us, it’s quarter. And things change really quickly. So you might lose an app developer. OK, well, that just means the three things we had cued up are gone. OK. So you have to really constantly be looking at where do we need to make an impact and where do we need to– oftentimes, things will happen sequentially. So thinking about what I need to learn early on that’s going to help me downstream is really important. And prioritization isn’t just for marketers. We are trying to talk the same language with our prod dev team, too, so we have the same priorities. The fifth one here is the trust and partnership with the product team. Because our structure is somewhat matrixed, we don’t all have the same bosses or managers. So what’s really great is we have constant weekly touch bases, not just about push, but about everything that we’re doing in terms of customer communication, so that when something does come up, you know, you can just walk over to somebody’s desk and say, hey, looks like there’s an issue. Or maybe you guys had an app release, and something happened and, you know, something was in QA– this has happened multiple times for us– and something broke in a deep link. And, you know, we didn’t QC it correctly or something. So there’s a lot of benefit of building that trust so that, when something does happen, everybody can jump on it, you know, together. And the final thing here is internal visibility and communication up and out. And what I mean by this is, like any project, you’re probably going to have a core group of people who are working on something. But what we have found is really helpful is this is not common language. When we get up in front of our company, we might do– we do weekly all hands. So we’re talking about push messaging, we’re talking about– there’s still some people in the audience who think push messaging is SMS strategy. And so it’s like we’ve got to bring people along, make sure people understand we’re using the same language. You know, not everybody understands, yeah, people have to opt in to push. Why can’t you just email– why can’t you just push, you know, your message to 100% of your audience? Well, because you have to get people to opt into your message, especially if they’re on iOS, right? So there’s a lot of education that we’re kind of constantly beating the drum on, and just helping kind of everybody get up to speed at the same time. So here is a very quick look at our push landscape right now. And I’ve bucketed it into these four different kinds of communication streams. They are actually much more complex. But there’s two audiences here I’m showing you, the pregnancy stage, if you will, and then baby. We’re defining baby from zero to age one. So the first one at the top is what we call stage-targeted. It’s a little bit of internal language, but that just means down to the week. So my example is, you know, week four is very different than week 24 when you’re pregnant. So we’re going to make sure we communicate a different message. And this is all about your baby’s development, so the fetal development. And it’s really important that this is on the first day of the week that you turn from week 25 let’s say to 26. And these are reoccurring. The second stream are what we call milestones. So this is going to be popular content and tools. So for instance, after a well check for your baby, you know, those are pretty– we already understand when those happen. So we understand the customer journey. So we’re going to make sure that we send you a push that week to say, now go to the Growth Tracker and update your baby’s head circumference, length, weight. You know, things like that. So we’re making sure we use those ongoing prompts to bring people back in. The third thing here is behavior. So this is, I think, where there’s tons of more opportunity for us. Here’s a couple of examples. Welcome message, right? We want to trigger this. Somebody just installed the app. The first 24 hours, we want to make sure that we have a series of messages that people get. The at-risk group. So these are people who are at risk of churning. So we have different models which show us we can increase our future retention if we get somebody to come back three times in the first week versus two. So we’ve tried a lot of different things to try to get people just over– that seems, like, easy, right? Just get them there one more time. But to get somebody over the hump in two weeks to take one more action sometimes is really difficult. So that’s what that category is. And then there’s other things that are ongoing. Of course, rating the app, hugely important for the visibility of your app, as well as word of mouth, sharing the app. And then the final category here is promotional. So these are going to be things that are PR, marketing, maybe branded content, which is how we monetize our content with partners. And these are going to be more one-time, right? We’re targeting a very specific audience around how to create the best, you know, budget nursery. OK, well, we’re going to identify a group of users we think that’s relevant for. If you’re in second trimester, you’re probably not thinking about your nursery. But if you’re in third trimester, you’re getting ready, right? So like, we know the user’s mindset, and when we should be communicating with her. And these are just a couple of Looker dashboards. To show you, on the right side is Android, left side is iOS. We see much higher click-through on Android for different messages. These four groups, these were colored dots actually correspond exactly to the slide I had before. So you can see that purple dot on Android, which are the [? stage ?] week notification, we were really excited that we finally got data. We moved this over from our local CMS, which has been on for years, and we finally got it over to Localytics, which means we can A/B test, we can get analytics, means that we can do things like rich animated GIFs. So those are very engaging. And so this is something that– you know, we’re looking at this stuff all the time, from a campaign level, and obviously as an aggregate so we can understand what the landscape looks like. So now I will take a little bit of a breath and I’ll go into a deep dive on our app onboarding. And when I talk about onboarding, I’m going to talk about it in two different ways. There’s the preregistration onboarding, and then there’s the post-registration onboarding. So I’ll talk about both of those, and different tests that we did. So this was very interesting data that we pulled from Localytics analytics. We knew that– this is from April, and this is a blended Android/iOS– we knew that about 70% of our users– sorry, not users yet, 70% of people who had installed our app were actually converting. So we thought, OK, 70%. But we didn’t understand what that was made up of. Looking at the data, we were able to see that there were actually different cohorts, or groups, within that. So within the success category, you have people who take your happy path, right? I designed my app to have these three steps to registration, and somebody did exactly those three steps. Happy path. Great. But then we discovered there’s this group called butterflies. Clicking all over, you know, 10 clicks when you only need to do three clicks, forwards, backwards. And they ultimately got in the app and registered, but they were looking for something. And we don’t know what it was. And then we start to break down the fail rate. So that first group was fail. Those are the people who bounce, essentially. They download it, they open it up, they leave right away. OK? Maybe they didn’t mean to download it. You don’t really know why they left. The second group was more interesting to us, which is the frustrated. People were taking the time to download the app, open the app, click around, but then they bailed. They never converted to an install, not in that session or future sessions. So what we started to do was say, this is super interesting. What can we do? What I want to do is I want to make the butterfly rate go– I want to see that success rate go up overall, obviously. I want more butterflies to become happy pathers, and I want to decrease that frustrated rate. That was the goal. So what did we do? We did a couple variations. And the first variation I’m going to show you is an onboarding test. So you’ll see the area I have circled in red here. We decided we should have a tour. Nobody knows what’s behind this gated registration. Why are we holding it back? Let’s tell people what it is. So what we did is it’s a series of four swipes. The first one animates through four different images. And fetal dev is– fetal development images are by far our bread and butter It’s what people really want to see. They want to understand is their baby healthy, how are they developing. It’s a total mystery, right, if you’re a soon-to-be-parent. And so we talked about our four main benefits here. Second variation. We looked at, OK, well, this was our original welcome screen. We had two different– we had three paths, actually, when you first came to our registration. We asked you to tell us if you’re pregnant, if you already had a child, or if you’re an existing member. And the reason we did that originally was for sales monetization purposes, right? We’ve had clients who wanted to only get in front of pregnant users, and we had people who only wanted to get in front of people with babies. So we built our product that way. Well, that relationship is done. So we said, you know what we need to do is we need to collapse them, and we need to make it make sense for users. So we combined the two so there’s only one. The choice is very simple, register or login. So after we ran these two tests together, and then separate, we found huge success. So this is from April to August. You’ll see our overall success rate went up nine points and our fail rate decreased by 19 points. So this was a really awesome early win that we were able to, you know, go to the organization and say, wow, that investment we made in Localytics really paid off, because look at what we’ve done. I mean, this translates all the way down the revenue stream and the value stream. Something else we did– I talked about the post-reg onboarding– is we also realized, once somebody registers, they’re still not engaged, right? There’s still a hurdle. They still have to actually start using your product. Even if they’re super motivated, people get busy, they put down their phone. They might never come back to your app. So what we wanted to test was post-reg onboarding. And we did four different versions here. One was a one-screen kind of one, two, three. We showed one specific feature which is very popular called Bumpy, which is essentially selfies of your growing belly. Third one was a slideshow, and fourth one is a video. And when we looked at this, we measured it against our two to 14-day retention I talked about. Again, very specific to our business. The benchmark was 76%. So already pretty high. People who were– so they went from discovering to installing to registering to launching the app, three quarters were already coming back two to 14-day. But we were like, we think we can do better. And what we did is we saw users who saw one of those ABCD creatives, OK, a modest, maybe half a point, a point. But when we looked at users who saw the creative and actually took an action, clicked on something, it went up a few more points. So this was really interesting to see, too. So now what we have is learning on the front end adding the tour made a difference, combining our reg flows made a difference, and we can see what happened on the post-reg side. So we will be looking at all this data together and figuring out what do we test next. What is going to be the optimal, you know, puzzle for our users. So some specific onboarding and retention recommendations. This is all based on the stuff I just showed you guys. The first one is to evaluate your paid and non-paid efforts across a continuum. So I’m not sure how it’s run in your organizations, but for instance on the marketing team, we’ve got some people who are thinking about how to use money and budget, get those paid registrations in for the app. We have other people about the organic side. How do I optimize my ASO? What am I doing on the metadata side? What am I doing on A/B testing in the app stores? All that kind of stuff. And what we realized is, you know, this year, we happened to face an issue where media budget was going away on one side, and we said, you know what? That’s OK. We don’t have to live by that media budget. There’s other things we can do downstream. If we can just improve that onboarding and retention a couple more points, that’s going to make up for all those new users I would have brought in anyway through the page. So think about it as a continuum, because it really shouldn’t be a fixed sort of mindset about what your roadmap is. You should always be looking at the whole picture, and how different things impact each other. Second is use Localytics in-app messaging. So the tour I showed you, we were able to do that through Localytics. If we had to wait in the queue for dev resources– I remember [? Lou ?] saying something about 22 weeks, which is 11 sprints, that’s a very similar world that we live in. You know, that would have been like, OK, well, why even bother starting, right? So what we were able to do is go back and say, OK, we want to do this. Are we going to build it in the app, or are we actually going to use Localytics. What we found is we didn’t have the right hook. We didn’t have the right event in order to key off for that tour. So we started talking with the app team very early on and saying, look, in three months or two months when we actually build this app, we’re going to want to make this live, and we’re going to need this hook in there. So having that alignment with the product early on meant we could get something like this in market very quickly, and not have to wait all those dev cycles. Third is always set clear DACIs, or RACIs, whatever you guys, you know, use in your organization. This has a lot to do with sort of who sets– who presses go. You know, I think that was one of the examples earlier on used. As well as just clearing the clutter for other things, like the tour. Product team might say, oh, I own that, that’s product extension. Marketing team might say, well, I own it. It’s how we market. The creative team might say, actually, I want the final say. So what we’ve made sure for just even that small example is to say, OK, who’s really driving this? Who’s going to prove it? Who’s being consulted? And who are we going to tell about it afterwards? That is some– it takes some rigor. And I mean, we find ourselves doing it all the time for small projects, big projects, because it cuts back a lot of the back-and-forth and confusion. Fourth thing I have here is, especially with your app, don’t set it and forget it just because it’s hard to get dev resources. So that reg flow I showed you was an example. We’ve had that reg flow where we ask people to choose if I’m pregnant or if I had a baby for a couple years. And if we had just let that keep riding out, we probably– you know, probably wouldn’t have really impacted the user experience, but we probably would have gotten fewer people through the funnel. And so it’s good to kind of always refresh those things and say, well, why did we build it that way, or why are we doing it that way. And just always bring those kind of bigger fundamental things to the forefront. Let’s see. My last thing here is challenge your internal goals. Monthly active users, does that really drive your business? For BabyCenter, we actually don’t even look at that, because that’s not how we tie revenue back to it, it’s not how we monetize, it’s not how we drive engagement or retention. So you have to make sure that you’re not just using industry standard best practices, but you’re using standards that make sense for your business. And this is my last slide. So here are some action steps I think everybody could take. The first one is get the right team together. At BabyCenter, we have one team formation right now. We’re actually going through an org change where we’re not going to have an app team anymore dedicated to just the app. So what does that mean? Operationally, I think it will take us some time to figure it out. But in order to keep doing push messaging and in-app and all this other great stuff we want to do, it’ll be really important for us to get the right team together. Two, continuously align with your prod, dev, and analytics organizations. Analytics is like our best friends. I feel like we couldn’t operate without our analytics partners, because they really help us, you know, come at things from an objective perspective. And also, you know, we have one analytics team, not multiple, so we can all agree, you know, one set of data at the end of the day. The third thought for you guys is to frame your questions proportionally to the impact they’re going to make. And then be patient on gathering the right data. So what I mean here is, you know, I know, as a marketer, I have lots of questions and ideas and thoughts about how we can make things better. But sometimes, you start looking at the audience size or the relative impact, and you have to say, you know what? I can give on that. I’m not going to actually move the needle. So those aren’t the things that you really try to push. So especially early on, if anybody is newer to the platform, or even just kind of as a mindset, I always find, you know, you want to get those big wins early, so go for things that you’re going to see the most impact on, because then people get excited. Then people start listening and they say, OK, what other ideas do you have, right? What other things do you think we can move. And then you get a lot of buy-in. And the last thing– I already talked about show early wins. Pretty self-explanatory. You know, get some points on the board. It really helps the future conversations. And the last thing is socialize your learnings. Don’t keep it within just the app team. Make sure that you think about this and you talk about how you can apply it to– in our case, to the mobile side of the house or the desktop side of the house or our offline side of the house. Is there something that we can do better in terms of how we’re connecting with some partner or client opportunity. This has been also really helpful so that the team starts to be looked at as more of a thought leadership instead of just the team that is creating campaigns and running content programs all the time. So that’s my slides. And I’m happy to take any questions now or after. I’ll take one from Bryan. He got a shout out in the presentation, so. Yeah, I feel like I had to, you know? No. So Kathy, you guys are tremendously data-driven. You have clear goals. You understand the data. You understand how the funnels work.