Analytics for Growth – Tobias Vogels from Flurry
Analytics for Growth – Tobias Vogels from Flurry
In this podcast series we ask experienced Appreneurs for one success story and one fail.
Tobias covers the new features in Flurry – its had some huge improvements in the last few months and its awesome that its fighting back. Tobias also talks about some of his wins & fails at startups prior to joining Flurry.
StreetHawk support automation for push notifications, geofences and emails using Flurry data – so we love that things are being updated there!
Here is a quick TL;DR:
- Acquisition is tough if your virality is limited to just couples – I guess this was a major problem for Path as well!
- In consumer Apps having a super-unique value-prop is critical. For Darkroom its pretty clear that its for people who want absolute privacy in their messaging.
- Some good insights on ASO – having the right mix of 5-start reviews and keword mixes that target the “super-unique value-prop”.
Mobile Growth Score
We score all out interview anecdotes based on 3 axes of: acquisition, UX and retention.
Let us know on our Twitter account how you score it.
David: Hello, this is David Jones from StreetHawk and I’m with Tobias Vogels. And Tobias is at Yahoo! And he represents Flurry. And I think you’re a developer advocate, aren’t you?
Tobias: Yes, that’s correct. Hey, how’s it going?
Tobias: So yeah, I’m a developer evangelist at Flurry and Flurry got acquired by Yahoo! about, I think, two years ago. And so as a developer evangelist, basically part of our growth team right here and working on building relationships with our developers, and our developer community, and trying to get Flurry into the hands of more awesome users.
David: So how come you guys need to grow your user base? You’ve got a massive amount of apps that are covered by Flurry?
Tobias: Right, right. I mean, we have over, I think, 800,000 apps on it.
David: It’s amazing.
Tobias: So that is a lot. I think we track over 10 billion sections per day. So that’s why I’m also like really focusing on building relationships with those apps that are already on there, and trying to make it better for them, trying to make it easier for them to analyze the data that they really need and figure out things that they really want.
David: Yup. You know, at one stage I was chatting to Sean Byrnes who was the Founder of Flurry, and we’re talking about how was it that the SDK was so incredibly successful? And you know, he said, “Obviously, it was kind of like our first-to-market type thing.” But also they did some ease of onboarding from a developer’s perspective in terms of SDK. Is there any other kind of insights you have about it as to why it’s worked their way?
Tobias: I mean, obviously, we are free, right? [Laughs]
David: [Laughs] Yup.
Tobias: I think that definitely helps. But I mean, it’s like, it’s a really big platform and we offer like so many different metrics on it. I think that’s like one of the reasons why people like it and why it fits like different needs. I mean, we have start-ups on there. We have really, really big apps on there like some of the biggest apps on the App Store or Google Play. And I think that’s why it’s successful because we have like a very broad range of different metrics. And we just did like really big redesign of our platform and there’s a new modern UI with some more intuitive workflows. And we’re also working on the onboarding to make it like even easier to get people to use it and to get to that data.
David: Okay, great. Yeah, let’s come back and talk about that a bit later. But I wanted to make sure we touched on your bio or your resume. You’ve done a couple of startups and been involved in various projects. Tell us a little bit about what you’ve done.
Tobias: Yes. So in Germany, I worked at my—I actually founded my own company in Germany.
Tobias: -which was focusing on content creation. So I built this kind of market place model, like a two-sided market place for content creation, where basically writers could connect to businesses, similar workflow as it is on Airbnb or like others, so those two-sided market places. And that’s still around even though I’m not active on that anymore. I was working for Bookatable in Europe which is the open-table of Europe-kind of, so like the largest online restaurant reservation service in Europe.
David: Right. Was that web and app?
David: Was that both web and apps?
Tobias: Yeah, that was web and app. Exactly. I was more focused on the web though. But I was responsible for user acquisition over there, so we also worked on some app campaigns. But at that time, our web traffic was way higher than apps. And it’s where our conversions basically came from.
David: Gotcha, gotcha. And you worked in Silicon Valley startups as well too, right?
Tobias: Right, yeah. Actually, like I was working on two different messaging startups over here. The first one was called Twyxt. And they develop messaging apps focused for couples, so it was like the kind of model social network out there.
David: [Laughs] It’s two people.
Tobias: [Laughs] Exactly, yeah. Just two people.
David: Although in San Francisco, I’m not so sure. [Laughs]
Tobias: It was just focused for two people. [Laughs]
David: [Laughs] You didn’t handle all these other freaky relationships—oh I shouldn’t say “freaky,” not that there’s anything against that. [Laughs]
Tobias: Yeah, I mean, we got some requests of them and some of them were serious, which is—I mean, sorry if I interrupt—but that was just focused on two people, right?
David: Right, gotcha.
Tobias: It had like some other tools in there besides the messenger, like share a calendar, to-do list, kind of like a digital keepsake box for your messages and photos. But it was really focused on like the messaging part for couples.
David: Right, okay. And what was your role at that company?
Tobias: I was also doing marketing at that company. So I definitely come from a marketing perspective, marketing side, yeah.
David: What does that actually mean? Does that mean acquisition though or does it mean retention or what?
Tobias: I think it means both. So—because, I mean, especially at a small startup, this early stage startup, retention is often times like so much more important than user acquisition because you need to learn from those users, right, from the beginning and the reason is to give them—the main purpose is to keep them, right? And especially when you’re on a tight budget that’s most what the startups are when they’re very early stage,
David: Every user counts. [Laughs]
Tobias: Exactly, yeah. You can spend like $5 or let’s say $2 on the user and then only 1 out of 4 users sticks around after two days.
Tobias: So that’s just not—that just doesn’t work.
David: Right. So where’s that startup today?
Tobias: They’re still around, but we focus a little bit more in a different product afterwards and that’s what I did before Flurry, basically DarkRoom. And DarkRoom is like a secure messenger that doesn’t store anything on servers or devices.
Tobias: So it’s a very interesting concept and it’s based on live messaging and all messages only get stored in the cache of your phone, so-
Tobias: So it’s really nothing on a server or your device ultimately. So if you have like something really confidential to say and you don’t want to share through iMessage or Facebook or all the other things…
David: Well, leave it in something like Yik Yak is presumably—well, it has to be. Yik Yak is storing stuff on the server because there’s threaded conversations, right?
Tobias: Exactly. Yeah, yeah. I mean, even though it’s encrypted, at the end, if it’s on the server, I mean, it is accessible, right?
Tobias: So that’s why we came up with this idea and that’s what I was doing last year.
David: Right, right. Okay. And so there was another startup that went bust, wasn’t it Secret? Secret, I think they were called, that was in sort of like the anonymous space.
Tobias: Yeah, totally. I mean, I think this space is getting more crowded right now definitely. I mean, server security is getting a very big boost, and yeah, they are like more, I mean, more and more messengers are also going for a more secure role. And I mean, WhatsApp just added encrypted communication to their platform.
David: Mm-hmm. Yeah, yeah. And so was there another startup in there as well too or just you said something about two messengers, you mean basically that same company doing two different—
Tobias: Yes, that we started out a new company out of the first one.
David: Right, gotcha. Gotcha. So you know, in the format of this podcast, we like to talk about like one win and one fail, and specifically around app stuff. So if you got something you can talk about on the win side?
Tobias: Yeah, for sure. So let me tell you about the company that I worked before, DarkRoom. So as I said we were in the position. We were really already staged right, and I think our main goal was basically to learn as much from actual users as possible, you know, like learn from—rather from users’ stand from just inviting people over focus groups, which we also did, right? But I think it’s more valuable to actually see what actual users are doing on your app.
Tobias: So we were looking into some cheap but effective user acquisition channels to get like a good base of users on it to keep using it so we can run our experiments and see what people are dropping off and what provide this sort of stickiness that results on something where we use Flurry Analytics for Growth. To really find like the user behavior patterns that provide the stickiness. And so it was interesting that I was starting to experiment a lot with ASO at that time.
Tobias: And I think that was like a really successful thing for us as we were like small and we were able to get about 300 organic downloads per day through ASO.
David: Right, wow.
Tobias: Which was, I think, really valuable at that time because we were at this early stage, I mean, our main purpose for user acquisition was to learn from those users, right? And every day, we got like 300 new users through ASO which is, I think, really, really great because this is like this channel that you have on mobile and compared to app, I mean there’s like not a free channel anymore that you can leverage at this stage, right?
Tobias: -because that’s the old SDK is just bad for people that are just starting out because it’s impossible to get a decent ranking there when you’re starting.
David: Yup. So you’re basically saying that, “Oh, if somebody actually found us by ASO, you were using keywords that are actually relevant to your app.” And then, in a sense, that user who downloaded and installed was effectively pre-qualified and so therefore should be a pretty good user as well too?
Tobias: Right, right. And it’s really about finding the right keywords and then optimizing the conversion from people that are looking up your App store side and then converting to downloading this.
David: Yeah. So you weren’t keyword bombing, like you weren’t saying Kim Kardashian or something like that just to get—yeah.
Tobias: No. That’s also not possible if you don’t have enough downloads, right?
Tobias: I mean, you need like a good base of reviews and you obviously need like five-star reviews to make it to the first five results probably, and those are the only ones that matter. And so you need to find your niche keywords which is also really—it’s getting more and more difficult especially after the algorithm changes that Apple made, I think, from last November? But yeah, if you optimize those keywords and came up with my own algorithm there based on reviews, ratings and then some data that we triggered or got from Sensor Tower about traffic and numbers of apps competing for different keywords-
David: Yeah, right.
Tobias: Yeah, so basically, I built this formula on how to predict downloads and predict rankings for certain keywords, and now work really, really well.
David: Right, that’s cool. So you mentioned they were doing an algorithm and stuff like that. So you mentioned before you were doing marketing. You are now a developer—sorry, you didn’t use advocate, what did you say? Developer—evangelist?
David: And you’re writing algorithms and you’ve been in marketing. So what are you? Are you a growth hacker or developer? [Laughs]
Tobias: [Laughs] I mean, the algorithm of more like a formula, right? It’s not—I didn’t write any code for it. I wrote like a formula and then calculated everything in a spreadsheet.
David: Yeah, right.
Tobias: More for marketing side. [Laughs]
David: Yeah, right. [Laughs] Okay, so you said that this sector is kind of like getting a little bit more cluttered now. So do you think that particular approach would still work now or is it that the keywords that you’re going after at that stage were kind of like green fields and be a bit more tough now?
Tobias: It is a little bit tougher and you definitely need to—I mean, it’s not like you found the right keywords and then you’re done with it, right? So like every update that we push to the App Store, we were updating keywords, we’re updating screenshots and video, and title et cetera for ASO, and you definitely need to look out for new niche key words and new—also for new hypes, right? If there is a good traffic for certain—if there’s like a trend for certain areas, then try to hop on it obviously. But I think it’s still a really valuable channel and they help some other people also to get some good results on it, and I think it still works.
David: Okay. So just give us an example in regards to DarkRoom, what were the kind of keywords that you’re finding were high-impact in your situation, just so you can kind of make this real?
Tobias: So they were all like very security-related, right? I mean, they were—some were like in the sense of a formality, like or content as disappearing, and some others were like more in the security space, in the private, so we try to get those people that were actually looking for something like this because if you just go for any niche and then try to convert those people, then they’re not that likely to retain on your app, right?
David: Yeah. And you know, in that situation, keywords—it’s probably quite different to the long trail keywords on Google, where, you know, length is okay, because you got kind of like autocomplete in the App Store and stuff like that. I’m guessing that most keywords have got to be sort of really, you know, high-hit on the frontloading of things.
Tobias: That’s true. The good thing is that you can combine keywords if you just separate them by comma. Apple automatically combines the words, so that’s a good thing.
Tobias: If you go for long trail, and yeah, what I found out, I mean, obviously the title is like—it matters like way more than the keyword field. So if you combine the title or keyword with the keyword fields and find like a good way to do that, that can be very impactful, so.
David: So, okay, you guys had found a way of actually getting some decent downloads in order to get real user feedback.
David: How did you guys actually make the decision to build the product in the first place, like what were the data points that actually led you to actually build it? Was it because it was so easy to build off the social network thing that…
Tobias: No, this actually came like from real user feedback. So some of our users, they were really asking about our security and security standards at Twyxt. And because they were like talking about sensitive issues-
Tobias: -they wanted like really secure solutions, so that got us thinking and we started thinking about user case that other people have. And that’s where we came up with this.
David: Far out. Okay. Cool. Alright, so yeah, so literally you’re pivoting out based on the feedback from users. So you got a better product market for you effectively out of that?
Tobias: Yeah. [Laughs]
David: Yeah. Great. Okay, so tell us about a fail then. What’s something in the journey where you’ve thought something was going to be awesome and it just fell on its face? [Laughs]
Tobias: [Laughs] Yeah. Maybe in—at the company before, since we’re already talking about pivoting, it makes total sense.
David: Yeah. [Laughs]
Tobias: [Laughs] And so, at Twyxt, we had this idea about—I mean, Twyxt again was this one-on-one messenger for couples and we had this idea about conversational messaging approach and thought about that this would basically give us a competitive advantage in that space of couples app that was kind of popping up in that time. So this conversational messaging basically means that it was a messenger that had multiple channels, but just to talk to the same person, right? So you can talk about different topics. So let’s say, you have one channel to talk about “What are we eating tonight?” And then the next channel would be about, I don’t know, “Where do you want to go on the weekend?”
Tobias: Right? So it’s kind of like—actually, you know, like some people, they use Slack with their partner and then…
Tobias: –use the channels to talk about different topics, so it’s kind of like the same concept and the same idea. That’s kind of cool.
David: Got it.
Tobias: Yeah, we thought this could be super valuable and it’s like a unique idea. And so we did a lot of focus groups at that time and people really liked this idea because, I mean, it was new, right? So I guess a lot of people like new stuff. And so we spent a lot of time programming it and build it into our app. But then when we really saw like extra people using it, like people were not adapting it. They didn’t really understand. The concept was too complicated. And at the end, we—like after a long time, we moved in and went back to regular messaging stream. You know, that’s just something that we saw, that something because I mentioned like earlier, it’s sort of important that you have like those first users on there and you actually get the data, all of it, and see what the retention is, where they’re dropping off and really understand instead of just, yeah, relying more on focus group and things.
David: Right. So you said that you were a Flurry user at this company. Were you using Flurry to kind of find that particular data rush, like you know, people weren’t really creating channels or…
Tobias: Yeah, we use Flurry at that time too, yeah. So yeah, we saw that—I mean, first of all, the retention was not that good. And then, people dropped off pretty quickly and they often, they just used one channel and exactly, as I said. I mean, they didn’t get it and they just put everything in there. I mean, obviously, we couldn’t—didn’t read what they write, right?
Tobias: They just use one channel exactly.
David: Yeah, right. Okay. With DarkRoom, did that end up being sort of like a one-too-many or a many-too-many type conversation and that’s been better for engagement and retention in that sense?
Tobias: Yes, absolutely. I mean, that’s also—that’s like a growth problem that we definitely had and I think also the other—sorry, and also the other companies in that space. I mean, we couldn’t grow by referring, right? Because…
David: Yeah, that’s it. And we get so much out of one user, right? So, whereas now, what you’ve got is this situation where one user, like all chat apps, with when one user actually communicates with another user, they’re effectively recruiting them into the network. And so…
Tobias: Yeah, totally. And you’re really depending on their relationship, right? I think if the relationship ends, then you lose the chats.
David: Yeah, so you get your K-factor virality instead.
Tobias: [Laughs] Yeah, yeah.
David: And typically, I mean, what were the user cases that ultimately DarkRoom was solving if the first one was solving “What’s for dinner?” and “What happens on the weekend?” the partners, you know, is…
Tobias: Oh, you mean…
Tobias: You mean, tricks, the-
Tobias: -app or the security of the app?
David: No, no, just in DarkRoom, is it, the user case, is it literally, you know, Arab Spring or is it people sort of coordinating around, you know, around something to do with taking over the foreign governments or… [Laughs]
Tobias: I mean, the thing is like, it’s really hard to figure that out, right, because we don’t know what people communicate throughout the app.
David: Yes, I knew you were going to answer that. [Laughs]
Tobias: But, I mean, that’s actually a really big challenge that we were facing as you can imagine. But we definitely have like a business need for that or a business user case, I mean, different areas, right? Like people who are using it to exchange server passwords, for instance.
David: Ah. Yes, yes.
Tobias: Or, you know, like to share confidential information and sometimes also just people I think they just use it for gossip, right? So I think there are like different user cases.
David: Yeah. That’s interesting. Yeah, so yeah, password, yeah, a lot of off-type details and stuff like that sounds completely useful sort of thing because—
Tobias: Yeah, your credit card information, right? If you want to share your credit card, yeah.
David: I get it. Yeah, that’s great. That’s fantastic.
David: Alright, that’s cool. So you’re saying basically, “Okay, no amount of sort of like virality in the first application, second application, you’ve got just by—it’s just basically built-in.” First application, you actually have a channel-type concept but people just, for some reason, got confused. Do you think that third app issue was really a usability issue or just, you know, just people just don’t work that way when they have intimate relationships?
Tobias: Hmm, yeah. I think it was kind of like a mixture of both, and it is, if you don’t have this critical math at the beginning, it’s really hard to like change user behavior, right? I mean, people were really not used to this. I mean, when they were opening a messenger, they all look the same, right? So they know exactly what to do but when they open Twyxt with this conversational messaging, they didn’t, right? So…
David: So you didn’t. Do you think that was just an onboarding issue though, like could it have worked if they spent more time on the UX?
Tobias: Maybe. I mean, some people, they really like the concept of it and I think that’s also the reason why some couples use Slack, right? [Laughs] But I’m not 100% sure on that if that would have worked, but like the critical mass… yeah, I’m not 100% sure.
David: Yeah. I hear of couples using “TrailIt” [0:2Tobias:41.8] or just as well too. [Laughs]
Tobias: [Laughs] Yeah, it’s funny. We heard like, I mean, people who are like using Google Calendar to like share their events, that where we also like pushed our shared calendar in the app.
David: Ah, right. Yeah, gotcha. Alright. So anyway, you mentioned that there’ve been some cool new additions with Flurry, and things that sort of like all of a sudden, a couple of months ago, all of a sudden sort of updates in, you know, on the daily stats. We’re getting faster and faster, so…
Tobias: Oh yeah.
David: Lots of good stuff has happened there.
Tobias: Yeah. We have done a lot of things in the past couple of weeks and months. And we were like partially rolling out some of the new features. And yeah, we really were really excited about the new things and our clients as well, I think. And so we updated our UI and, yeah, like changed some workflows. We added real-time metrics which a lot of people wanted, so that’s really cool, so now you can see how many sessions you had per hour and then you use this data to like roll out new features, right?
Tobias: So okay, at 1PM there’s not much going on, like to my main channels, we’ll get—1AM most likely not 1PM.
Tobias: Yeah, things like that and we added like new dashboards, data filtering, so you can filter all your data right when you look at it, like when you look at your current retention metrics. For instance, you can already filter that by gender or by language or things like that. So that’s really cool. And yeah, we added like the possibility to compare multiple apps at the same time and group them if needed. So we have a lot of games on there.
David: Yes. So for those who aren’t kind of familiar with the gaming sector studios will have multiple apps running at any one time, and so they can do side-by-side comparison effectively of what’s happening this month from two different games?
Tobias: Yeah, exactly. Yeah, and you can group your apps, right? So let’s say you have like, I don’t know, like five apps like more, like in the board game-style, and then you have two games that are like more in the sports segment, then you can like group those apps and then see how they perform.
David: Yeah, that’s great. You know, the kind of—the question I’ve had about Flurry has been things like gender and stuff like that, like that’s not explicitly entered by the user in the application, so how do you guys pull that off? Is it, you know—and how accurate is it? Is it, I mean, sure, okay, David Jones, I think if you check out David Jones on Facebook, he’s female, right?
David: But, you know, those-
Tobias: It seems pretty accurate. I mean, honestly, I think—so how we do this is we have like this panel of 50 million people and they give us gender information and age information, and all this good stuff.
Tobias: And based on their behavior and how they interact in apps, we see how other people interact and then try to see if there are common patterns.
Tobias: And if there are, then we say, “Okay, we estimated this person is male and between 16 and 25 years.”
David: Ah, okay. Great. That’s very educational. Cool.
Tobias: So yeah, I mean, there are still some unknown people in there. So we can’t like—we can’t say everybody is male or female. But gives you a really, really good idea and it’s actually pretty accurate.
David: Yeah, that’s great. Yeah, so what’s anything you know about what’s coming down the pipeline? Anything you want to talk about?
Tobias: Yeah. Actually, I think the most exciting feature that we just launched is Explore and we had a beta version of that for a couple of months.
Tobias: And now, we’re really getting better at it. And it is like this interactive analysis feature where you can run instant data queries based on events that you trigger.
Tobias: So it’s segmented into three parts, first is like explore segments, where you can kind of segment your users to find out like which users are really, really engaging with your app and which are sticking around and performing certain actions. And then we have funnels which are really instant now so you can do a lot of micro-funneling, and yeah, dig deeper into the different funnels. And the last one is, of course, analysis, where you can basically find out, for instance, if a person signs up what they—do they usually like go ahead and become an active user or like become a customer and purchase something, right? So if you have done purchases, like usually metric that a lot of apps track, at what time? Are they turning into a customer?
David: Got it. Very good. I can’t be too excited because we do some of those things as well too.
David: But it’s kind of cool that we’ve been able to actually integrate with Flurry to take data out of Flurry and then turn that into automation as well too so…
Tobias: Oh, absolutely. That’s really cool.
David: Yeah. Well, mate, thank you very much for your time. I appreciate it.
Tobias: Yeah, thank you. That was awesome.
David: Yeah. Thank you. And we shall catch up maybe in a couple of months, just see what’s happening, again, get an update because obviously, things are really moving with the platform now so we could learn more about that.
Tobias: Yeah, definitely.
David: Okay. Thanks, mate.
Tobias: Alright. Ciao, ciao.
David: Ciao for now. Bye.