Roman Karachinsky Interview

Dec 22 2011

Roman Karachinsky, CEO of News360, talks about delivering relevant content. Roman explains how they use information from your “social graph” to create an “interest graph”. Later, he talks about supporting all major mobile platforms and shares observations on differences between Android and iPhone users.

Links referenced in the show:

News360

The music in the show, Have Mercy — Big Walter Horton, was provided by Mevio’s Music Alley.

Transcription

Roman 00:00:22

My name is Roman Karachinsky and I’m the CEO and cofounder of News360. We’re a company that makes mobile apps for news discovery. We aggregate a whole bunch of sources around the Web, understand what’s important, and then try to understand what’s interesting to all of our users by, you know, looking at their Web usage, Facebook, Twitter, how they save stuff (articles), how they use content. Once we understand what they like then we try to generate feeds (to find things for them) as soon as they come out rather than, you know, having people search for things and discover them on their own.
Ryan 00:00:54

That’s great. You’re definitely solving a problem that I would like to have solved for me. I think the idea of curating my own OPML of RSS feeds that I like is not really a hobby that I would like to pursue.
Roman 00:01:07

Yeah. I mean, like, when we started the company that was because we had this problem as well. It was a very personal thing for us because back in, like, a few years ago I was spending maybe an hour or an hour-and-a-half a day just making sure my Google Reader said “no unread articles” at the end of the day. And I realized that it was kind of a chore that, you know, that was not the most efficient way to spend time. I had to -- you know, I looked through tens of different blogs that I generally perceived as interesting but, you know, I had to discard, like, 90% of the headlines that I looked at and I only read 10%. And our thinking was, you know, that this thought process that goes into understanding, “Okay, do I want to read this or not,” is something that can be automated. And, furthermore, because, you know, our assessment and just, kind of, the way people consume content might be focused on sources is very limited. There’s not a lot of room for discovery there. I mean, there might be a great source out there; like a really awesome blog about space exploration that you’ve never heard about even though you’re really interested in space exploration. And, you know, unless it finds you somehow, either through search or through, like, social media through your friends (which can be, even in some cases, like, unlikely if you don’t have a lot of friends who are also into space exploration) and you’re never going to find it. And one of our goals was to make sure that, first of all, we remove this chore of having to filter stuff yourself but also expose you to other things that should be interesting to you.
Ryan 00:02:31

Yeah. It’s tricky trying to find new sources for information. Well, I guess it depends mostly on what it is that you’re interested in. If you’re just interested in something like generic pop culture, like, that kind of thing, you know, there’s sites for that. Or if you just want tech news, you know, there’s sites for that; you can do that. But what if you want to reach outside of, you know, like, a blog that has a niche. You know, how do you discover stuff? That’s something that I’ve been dealing with recently, actually, because I found that I was mostly just always reading about, you know, like, tech news and gadgets and things like that because, well, basically, that’s what I’ve ultimately curated my RSS feeds to be. But it isn’t the only thing that I’m interested in; I just don’t know how to find other stuff, really.
Roman 00:03:15

Yeah. And that problem actually goes kind of both ways that you don’t know how to find good content that is interesting to you outside of those mainstream or the ones that you’ve already found. And, you know, the authors of that content also have a problem in terms of how they find an audience. If they’re a niche or a local kind of publication or blog, it becomes really hard for them to market; especially on mobile because if you look at the top 20, you know, free apps on the iTunes App Store it’s basically, like, huge brands. It’s, you know, Times, USA Today, Wall Street Journal. And if you’re, you know, a blog that maybe you have, you know, awesome content, maybe you’re really knowledgeable, maybe you’re a great writer. But unless you have a big brand associated with you or unless you’re a genius in terms of social marketing or SEO then, you know, you’re going to have a hard time actually making sure that somebody’s reading your content.
Ryan 00:04:09

Yeah. I mean, the great thing about the Internet is there’s no barrier to entry; anybody can write about anything. And the bad news is who’s the curator? And, you know, who’s the gatekeeper of this information so that I can easily find something that’s valuable, right?
Roman 00:04:24

Yeah. And we’re hoping to basically, you know, we don’t want to take on the role of, like, an editor but, generally, we want to be able to editorialize for everybody in their own way. So, I mean, if we see that you are interested in something by looking at how you read content in the past then, you know, we can very easily find things for you. And, you know, and we can generally determine the quality of the content pretty objectively as well. If we understand, like, who are you looking for, like, technical, really in-depth content about technology or are you looking for a layman’s overview of technology or science. We can see those indicators and we really try to find content that should be right up your alley. Yeah. And we can, you know, we can hypothesize and give you different sources of things and see which ones you like or don’t like, but the key thing there is really identifying those interests. It’s really understanding what your general areas of interest are.
Ryan 00:05:23

Yeah. And let’s dig into that a little bit more. So the first thing that happens, you know, I create an account, I log in with Twitter, it lets me, what, connect with Facebook, Google Reader, what else? You take signals from all over the place to try and learn about me right up front.
Roman 00:05:40

Yeah. So right now we support basically five outside platforms. We try Facebook, Twitter, Google+, Google Reader, and Evernote. And but, I mean, that’s kind of just the start. The idea is we developed a technology that can take any sort of connection that you have with a piece of content whether you, you know, shared it or liked it or read it or commented on it or saved it or whatever. You know, it can take all of these connections and the more, the better (the more accurate it will be). And then for every connection we can identify what the connecting of interest was. What was the thing that prompted you to act on this piece of content? And, I mean, obviously, you know, we have to aggregate a lot of this information; you know, discard the outliers that, I mean, maybe you read something about Justin Bieber because you accidentally clicked on it; that doesn’t mean you’re actually interested in Justin Bieber. You know, we have to take that into account. You know, this is an interesting time because for most people there is already enough information in their Facebook account and their Twitter accounts about their use of content; there’s more than enough to get a pretty accurate view of what they’re generally interested in. And I think that’s something that hasn’t been true before. And so, yeah. So our goal is basically to look at any place that has an API you are comfortable giving us access to and then use all the information in that place; just look at it and, I mean, kind of, the goal is to distill it to a very understandable, very transparent set of parameters which we call an interest graph. And we actually want to show you the interest graph. But, you know, we don’t to make a mathematical model; it’s too complex for you or us to understand. We want it to be very human-readable. And we show it to you, we ask you, like, “Okay, is this right? Did we get, you know, did we get anything wrong?” We allow you to edit it, to add things to it. Maybe, you know, you never read anything about bonsai trees because you were never able to find anything, but you were really interested. So we want you to be able to add things or curate your own interest graph. And then we use that interest graph as, basically, it’s almost like a reverse search engine; where we know what you’re looking for all the time and then as soon as we see something published on the Web then we know, like, if this is really important to you, we feel that it touches a lot of your interest points, it’s like urgent, it’s like from a high-quality source, we can push it to you either via push notification or e-mail or, I don’t know, a text message or something. And if it’s just something that you should be interested in and, you know, it’s something that is going to be, like, entertaining or useful or enlightening to you then we just pile it in a pile. And then when you have the time, instead of, you know, spending an hour filtering through headlines, you can just consume all of the things that we’ve kind of identified as interesting for you.
Ryan 00:08:23

It’s interesting that you do let the users kind of self-curate their interest list because while I think that oftentimes that sounds like a good idea, it seems like that could be a really good way to muddy up the water because, I mean, you know, the phenomenon of people not being extremely honest about (not honest isn’t the right word, but not realistic about) what they’re interested in, right? Like, if you ask people, like, what kind of music do you like, people are more likely to say probably something that they think will impress you than maybe necessarily what is the truth, right?
Roman 00:08:57

Yeah. I think that it’s really important to have a good starting point; to have something that we can fill out. You know, the main point of curation is not to ask people what are you interested in because, you know, it’s a very hard question to answer truthfully and completely and exhaustively. And especially for, like, just if it’s an area as broad as just content; like, what are you interested in knowing about? You know, there’s a lot of stuff and, you know, I’m probably not going to remember it all. I’m probably going to, you know, mention something that I’m not really interested in because I would like to be interested in it or something like that. But yes. So our goal when we allow people to edit it is not to have them fill it out; it’s to have them correct mistakes and to have them add things that are, you know, maybe a recent interest or something we might have missed. And the goal is not to have everybody even do it. You know, we want people to be able to look at their interest graphs so they know what we’re doing. We just don’t want the personalization kind of engine to be a black box. Because, you know, I think that leads to a bunch of other problems. Like, if you look at how Google does it, I mean, they have a ridiculous amount of information about me, you know, from my Web searches, from my Gmail account, from my Google Reader. I mean, there’s, like, basically everything I do, Google probably has some idea of what I’m doing. And I have no idea how they’re using that information. It’s like, they probably have, you know, some mathematical model that’s really complex; it has, you know, thousands of dimensions that determines, you know, whether the search result’s going to be good for me or not. And it’s really optimized towards, like, click-through rates and towards advertising. And probably, like, even if a Google engineer looked at this model he wouldn’t be able to understand who I am or it’s going to be very hard for him to make sense of it because it’s, you know, it’s optimized to very statistical, you know, making statistical conclusions about whether this ad is going to be effective on me. And our goal is very different. We don’t want it to be, you know, in an obscure kind of statistical model. We don’t want it to be optimized towards click-throughs or anything like that. We want it to be an actual representation of your interests and that’s why we really feel that it needs to be transparent. It might be, like, large, it might be very detailed, but it needs to be something that you can look at.
Ryan 00:11:12

You know that springs to mind kind of a common defense about the opacity of how Google search works is it’s, “Well, that’s their secret sauce. That’s where all the money comes from so that has to be the secret. I mean, they can open all kinds of things, but they certainly can’t take their core business and have that be the transparent part, right?” And that’s basically your bread and butter too and you’ve taken a very different direction than maybe what’s intuitive. So would you like to speak to that?
Roman 00:11:41

Yeah. I mean, sure. I think you’re right that Google is probably very protective of that data because, you know, that is what gives them a competitive advantage. I mean, I don’t think that our interest graph is really a competitive advantage right now because this market has no good solutions yet. So we don’t really have to compete with anybody. Because even our own platform is not at a point where we feel that it’s, you know, it’s a really good automated personalization solution for content. That’s something I think that’s going to be, like, a year or two out for anybody.
Ryan 00:12:14

So basically what you’re saying is even when it is core to your business, go ahead and learn in the open. You’ll probably learn faster.
Roman 00:12:21

Yeah. And I think, you know, it’s less about the interest graph itself and it’s more about how you use it is what is going to define the quality of the product. Because the interest graph, at least in the way that we think about it, it’s not really, you know, because it’s not complex it’s not really that, you know, competitively valuable.
Ryan 00:12:44

Yeah.
Roman 00:12:45

You know, we know about you that you’re interested in these hundred topics. I mean, that’s -- somebody else could probably learn about this as well if you give them access to your Facebook profile and your Twitter account. I mean, it could be -- I mean, they might not be as accurate about it, they might not be as quick or, like, you know, involved about it, but it’s not something that we feel is really worth kind of keeping a secret or protecting. I mean, there’s probably a lot more indicators that we can collect that are less obvious and might be -- but I don’t think they should be part of your interest graph. I don’t think they should be kind of obscured. Our goal, really, is to describe you in very simple terms in very distilled and, you know, human-readable terms; like, how you yourself would describe you.
Ryan 00:13:32

I like that. I like the way that you, I mean, if you scope it down it can be really easy to just kind of run down that rabbit hole, right? Like, once you start capturing and trying to map out, like, a social graph-y kind of thing, I mean, how deep do the tentacles run?
Roman 00:13:48

Yeah. And I mean, I think it’s very easy as well to get lost in this kind of complexity because -- and I don’t think the complexity itself is necessarily something that’s making things more accurate; it’s just making things more complex.
Ryan 00:13:59

Right.
Roman 00:14:00

This is, you know, we’re consciously trying to avoid it.
Ryan 00:14:03

I do oftentimes like to think about, like, when people, like, especially with the self-curated thing, like, you guys, to take your example of Bieber. You know, like, well this guy, clearly this guy does like Justin Bieber, but he keeps insisting that he does not. And now you have an extra bucket to put him in: you know, people who insist that they do not even though clearly they do like Justin Bieber, right? Like, you get that one level deeper and then you start trying to make all kinds of assumptions off of that. I don’t know; it’s just, like, a fun thing I like to think about. Like, especially like Google, algorithms, and things like that; every time they let you customize things. I just see, like, that’s one more single; it’s part of the game.
Roman 00:14:43

Yeah. And, I mean, like, it’s funny to talk to about it. I don’t think there’s actually that many people who are really interested in Justin Bieber but conceal it.
Ryan 00:14:50

Right. Well, I mean, that was the example from your example. But, I mean, yeah, the idea that people, they like to insist they’re one way even though it would be obvious to their friends, not just people who are data mining them, right?
Roman 00:15:02

Right.
Ryan 00:15:03

That they’re -- right. And then using that as a signal and how useful that is. But I think, yeah, you see the same thing with user experience, you know, debates against, like, split testing and things like that. Once things get past a certain degree of complexity then, you know, the relevance is -- you start losing a lot of value and is there really -- is it worth your time because there’s too many things that you’re not thinking about?
Roman 00:15:27

Right. And I think, kind of, there’s another issue there is that, you know, we want people to understand how the filtering works so they understand what they might be missing or they, kind of, understand that their perspective might be limited in a way. And I, you know, that’s kind of the filter bubble problem that people are talking about constantly is that if you’re kind of hiding the personalization mechanism, if you’re making some, you know, decisions behind the scenes that might affect your presentation of reality, then it’s very easy to, you know, to kind of fall into your own little hole where you keep kind of sustaining the idea, “Okay, I’m really interested in Justin Bieber and nobody else,” or I don’t know. Or, “I’m really, you know, a hardcore Republican and I never want to see anything from CNN.” And it’s a self-sustaining kind of problem because this, you know, Google, now that they see that you’re clicking on Fox News links a lot more they’re going to be giving you more and more Fox News and you’re going to be, like, keep clicking on them. And, you know, even though your perspective might change; like, you know, three years later maybe you’re a completely different man. You know, because you don’t see that at work your, like, world view and how you get your search results and how you get your news might still be affected by that, you know, conclusion that they made a long time ago. And if, you know, if that’s something that you don’t see, if you don’t see the process that’s making these decisions, you know, that’s kind of an interesting problem in terms of, you know, that you might have screwed up your own kind of perspective on the world just because now you’re looking down this tunnel instead of having a broader view.
Ryan 00:17:08

Yeah. Well, you know what’s interesting is that kind of ties back to, you know, a very similar kind of talk about how social networks are going, right? So Facebook has been seen as not nearly as good by a lot of people of a source to find news that you’re interested in because people friended people kind of fast and loose and their social graph is kind of out of whack and it doesn’t have any kind of bearing on, like, the kind of person that they are or want to be, right? Because this is lots of signals from your past and not necessarily who you are. Whereas people tend to get more relevant news from, you know, something like Twitter because maybe not even because Twitter’s better but because it’s newer. And then again with Google+. When it came out people said, “Wow, you know, all the early adopters are really smart people so that means this is a better source for news than even Twitter was before.”
Roman 00:18:03

Yeah. Definitely. And I think there’s kind of an even interesting concept that’s coming out now is that, you know, that everybody’s trying to find ways to allow people to curate their own social, kind of, graph to make sure that the quality stays high even as they continue to expand and continue to add people, which is something, you know, Google (I don’t know if they pioneered it but), it’s something that they made a core feature of the network with circles.
Ryan 00:18:27

Right.
Roman 00:18:28

Where you can, yeah. You can just, like, lump all of the tech people and all of the early adopters that you followed into a tech circle, you can put all of your friends and family into a circle, you can put all of your, I don’t know, like, coworkers into one, you can put -- and that kind of helps a lot in terms of as your social graph expands, if you started doing that from the very beginning then that keeps it very organized; that keeps the quality high. Because on Facebook I think that was the main problem where when you reach past a certain friend limit you just have a hodgepodge of, you know, baby pictures from your aunt and, you know, tech news from your coworker and, you know, dancing events from your, like, girlfriend and it’s just -- you know, it becomes unmanageable at some point. And I think, kind of, this idea of we know your interest graph then, you know, it can help us understand what you want to see from different people in your social graph as well. And I think it’s, you know, an interesting, like, problem with circles is that I have all these, like, tech people in my circle that’s called “tech,” but when they share, like, pictures of their dog and what they ate for lunch I still see that in my circle called “tech.”
Ryan 00:19:38

Right.
Roman 00:19:39

So how do you filter -- like, if I know all these people that are really great technological thinkers; you know, they’re awesome, like, CEOs, CTOs, you know, early adopters, kind of, product developers and so on. And I really want to know their perspective on technology. I really want to, like, see what their thinking about that but, you know, they keep broadcasting everything, right? They broadcast their musical tastes, they broadcast their life, they broadcast -- and maybe that’s something that I’m not necessarily interested in. How do you filter that? And that’s, you know, I think that that has to be automated.
Ryan 00:20:14

It seems very similar to, like, the same kinds of problems you have with the difference between folders and tags, right? Like, if something has to be one thing it can’t be multiple things. Like, how do you -- because you don’t really want that (I don’t, personally want that) granular control over these circles. The simplicity is sort of the strength. But I do understand, right, what you mean. But I think an aggregate you can make these lists where people kind of talk less about nonsense in a certain list.
Roman 00:20:41

Yeah. And, I mean, it could be solved maybe through giving the likes through the sharing mechanisms to making them a little bit kind of more intuitive in terms of what you’re sharing; is it a personal thing? Because the problem is, I mean, you don’t see which circles people’ve added you to. So you can’t choose, “I want to share this to people who’ve added me to a tech circle”…
Ryan 00:20:59

Right
Roman 00:21:00

…even though that might be something that would be kind of intuitive and that’s what you really want to do.
Ryan 00:21:06

It’s interesting to me that -- so, you know, the thing before was no one wants to fiddle around and put people on lists was the saying with Facebook’s group system, right?
Roman 00:21:18

Right
Ryan 00:21:18

But then -- and no one did when the feature kind of came out in a crude way. But then Google+ comes out with these circles and everyone does circles from the beginning. It almost seems like less, potentially, less important is do the tools exist and how good are the tools and more it’s just like a brushfire mentality. Well, if I’m starting from scratch I can go ahead and do it right from the beginning with the knowledge that I didn’t have before, right?
Roman 00:21:44

Yeah. And I think it’s just about, kind of, the barrier; the amount of work that you have to do. Because, you know, if I think about having to kind of group all my Facebook friends into lists that’s, you know, that’s a ridiculous amount of work. And if it’s spread out, if I do it from the very beginning, as soon as I add somebody I kind of specify what my relationship with them is and add them to the right circle, then it’s the same amount of work but it’s spread out through a longer period of time; it’s not as preventative. I think that’s kind of the main issue.
Ryan 00:22:13

So bringing that back to the idea of these kind of curated interest graphs that you’re working on; is that something that you’re trying to solve actively now or maybe something on the back of your mind as later. Like, how do you solve this, you know, a few years down the line when people have kind of been maybe getting more and more signals?
Roman 00:22:31

Yeah. No, I mean, our thinking is basically, you know, right now there’s three ways people get content and that’s search (where they have an active query), the kind of sources that they subscribe to (and that could be, like, a magazine in the mail or it could be an RSS feed), and third one is social media where they, you know, where people basically play editor for you. And we think that in each of them, you know, the experience could be improved with using the interest graph one way or another. You know, with search we could basically automate search altogether; we could find things for you before you search for them. With sources we could tell you which sources are, like, we can first of all tell you which sources are going to be interesting to you in addition to those that you already subscribe to. And put those sources that you do subscribe to, we can filter things better so that you can kind of -- you don’t have to look at all the headlines; you can just look at the ones that will make sense to you. And for social, you know, we can -- it’s the same thing. It’s about, kind of, you know, each friend of yours is basically a source that’s pumping a lot of data your way. And that data is, you know, photos, videos, content, links. And the question is, you know, can we optimize this a little bit so that when you have 500 friends you, you know, this doesn’t overwhelm you and you get, you know, the most valuable content instead of just, you know, getting a view of all of it. And there’s basically, kind of, people approach this problem in two ways. The first way is to use, kind of, social data and to try and see, like, collaborative patterns, which is something that Facebook is doing with their EdgeRank; where they look at a piece of content, they see how people are sharing it, they see how it’s spreading through the different groups, and then they kind of, you know, make conclusions about for this particular person do we want to place this higher on their newsfeed or do we want to place this lower or maybe not show it at all. And that works, you know, that works well for things that are socially, kind of, for vital things; for things that are socially, kind of, that spread well. But it doesn’t work for a lot of other things. And, you know, the second way to approach this problem is to think about it in an algorithmical kind of -- to try and understand, you know, we know these things about this person, we have this history of their consumption, we have the things that they said about themselves in their profile, we have their, kind of, geographical location which can kind of yield some interesting results as well and use that data to prioritize the incoming stream to make it a little bit more useful; and that’s what we’re trying to do.
Ryan 00:25:05

Interesting. So outside of the signals, you mentioned things like topical; like, things that people are interested in and what they’re searching. But another signal that you have comes from the actual device that they’re on. Now you support, what? Like, seven different platforms? Something like that?
Roman 00:25:21

Yeah. Yeah. Six mobile apps and the Web. So that gives you a good idea.
Ryan 00:25:23

Nailed it. Why so many? Like, there’s only -- there’s two front-runners. Like, is it really worth the effort to go everywhere else?
Roman 00:25:36

Well, I mean, our goal is basically, you know, we want to be (eventually, we want to be) on every platform that people might have. And the goal is not to be an app, but rather to be a place for them to go to when they are looking for content/when they’re looking for news; it doesn’t matter where they are. Are they on their computer, on their phone, on their tablet? And, like, really you’re right that there are two front-runners but -- and, you know, the six platforms that we support are iPhone, iPad, Android phones, Android tablets, Blackberry, and Windows Phone. So four of the six are the front-runners that you’re talking about. So that’s not really as diverse maybe as you thought it was. But yeah. The Windows Phone and Blackberry; like, for us, both of them were not huge engineering kind of projects simply because at this point, you know, things are reasonably quick to develop. You know, if you want to make an app for a Windows Phone, it’s not a very huge proposition simply because, you know, the development tools are very good. You know, it’s just not a large effort to write something. You know, it still takes a lot of time to make it very good and to make it, kind of, impressive and to make sure that people love it but, kind of, the return on that investment is usually enough to cover the effort.
Ryan 00:26:55

That’s been an interesting thing I’ve heard about Windows Phone. And it’s funny how people say, “Well, have you seen the tools? Like, it’s really easy. Why not?”
Roman 00:27:08

Yeah, I know. I think Microsoft did a really good job of making sure that the development tools are really high quality. And that kind of goes a long way in terms of, you know, in terms of us making the decision, okay, to want to do it or not. I mean, if it takes, I don’t know, like, maybe a month for one developer to create the app or it takes, like, three months for a team to create the app, you know, that makes a big difference.
Ryan 00:27:31

It’s interesting because you don’t really ever hear that about either Android or IOS. I think you will find people that maybe prefer, like, “Well, I already knew Java so that’s great.” Or they’ll say, you know, they really like Objective-C for this, this, and this reason. And then they’ll support the other one because, like, “Well, we have to.”
Roman 00:27:52

Yeah. I know. And that’s true. I mean, like, I think now you have to. Yeah. Because both IOS and Android are, you know, their audience is so big that even if it takes, like, you a year and you have to, you know, and it’s really hard to develop and whatever. I’m not saying it is, but even if that was the case you’d still, you know, you’d have to do it just because, you know, you have to have access to that audience to be relevant. If you don’t then.
Ryan 00:28:17

So a lot of people speculate that there are big differences between, like, Android users and Apple users and I assume also about Windows Phone 7 users. But you actually have some data to back that up and you’ve seen trends.
Roman 00:28:34

Yeah. I mean, and kind of, basically, the data that we look at is we try to collect, kind of, just statistics about feature usage, about what general topics and categories people kind of read stuff from on the different devices. And yeah. I mean, there are obviously, you know, there’s a lot of difference in terms of form factor. So when you look at, you know, tablets versus phones there’s huge kind of differences in usage because people use their phones throughout the day but, like, with the really short kind of sessions the user of tablets; mostly in the evening or in the morning and, like, have really long sessions. But in terms of, like, Android versus iPhone, yeah. We saw a lot of differences in terms of content usage and, you know. And both -- I think, kind of, an interesting difference was we released our iPhone app I think a couple, like, at least four or five months before the Android app. And, you know, the reception that we got was really good. I mean, we have, like, a very high rating. People review that all the time; like, write really, really nice reviews. But they never, I think throughout the five months that we had that iPhone app out, I think we got, like, maybe ten e-mails asking us for features or criticizing or something. And as soon as we released the Android app, like, in a week we got maybe, I don’t know, 50-100 e-mails. Either people, like, really critically kind of, you know, telling us. Even though, like, on the Android market it was the same exact thing. People really liked the app, all the reviews were great, the rating was very good. But we got so much more feedback of people telling us, “Okay, you did this wrong.” Like, you know, “Fix it now.” And they were really forceful about it. It was kind of funny. But yeah. I think, like, Android users, maybe not in general, but there’s a lot more technical people using Android and they’re a lot more vocal about how they want to use the apps and, kind of, they’re a lot more critical. And we saw the same kind of trends in the content consumption as well. You know, I think our tech category has, like, 50% more, kind of, usage per person averaged out across the whole population than on the iPhone, which is I think kind of interesting. And, you know, instead the iPhone kind of is higher on the politics and entertainment side versus Android.
Ryan 00:30:51

That’s really cool. I think I hear a lot of people talk about how the different users of the different platforms are different. But I think mostly it’s people basically self-identifying and then lumping everyone that chose like them like them, right? So you’ve got your Android guys are like, “Well, you know, it’s for smarter people.” And then you’ve got your iPhone people saying, “No, no, no. IOS is for people who have style.” You know, like, they both pick whatever they think is their own strength, I guess, and then just say and that’s what people of this platform are. But I really like that you’ve, you know, objectively -- you’re actually collecting these stats and you can actually, you know, agree with some of them, disagree with others. But it’s not about assumptions. It’s not about these wild speculations that I feel that we hear all too much of.
Roman 00:31:41

Yeah. No. Yeah. Definitely. Kind of, from what we’ve seen, Android users are a lot more technical; that’s our vision. And, you know, I’m not saying we’re not Nielsen. You know, we don’t have statistically, you know, kind of, reliable methods of gathering data. I mean, who knows? Maybe it’s because most of our users, at least in the beginning, were early adopters who kind of like apps and they were more technical.
Ryan 00:32:03

Right. Well, and you bring up a really interesting point there too. It’s something that’s important. It really doesn’t matter what Android users on the whole are like so much as it matters what your Android users tend to be like, right? Because you need to think about your customers, not potentially this, kind of, fictitious user group that may or may not be valid.
Roman 00:32:24

Yeah. Definitely. I think there’s definitely -- there is a difference in mentality between Android users and, I think, IOS users that you do have to take into account. Both because, kind of, the platform is different so they’re different, so they’re used to different things, they have different ways of kind of solving problems on their phone, and I think, yeah, you can’t just map things. You know, you can’t say, “Okay, they’re more or less the same.” You have to really -- it’s a way of doing product design. And when you’re trying to understand what kind of feedback you’ll get and what the reaction will be to your product, you do have to do it separately for IOS and Android; at least that was our experience.
Ryan 00:33:08

Well, Roman, thank you very much for your time. I’m going to have to have you back on because I think that you have tons more information that is very interesting to me about this topic.
Roman 00:33:17

Yeah. Oh, thanks. Thank you for having me. And I’d love to. I’d love to come back and talk to you a little bit more.
Ryan 00:33:21

Excellent. Excellent. Thanks again for your time.
Roman 00:33:23

All right. Thanks, Ryan.