The challenges for Video in Mobile Engagement

Apr 17

The challenges for Video in Mobile Engagement

In this podcast series we ask experienced Appreneurs for one success story and one fail. In this episode of the Mobile Engagement Podcast:

Incoming Media is a fascinating technology that uses “Machine Learning” on the user phone/tablet to adapt to the user’s video viewing patterns and therefore pre-load future for future consumption. They are attacking the problems of slow video load times and the impact on user acceptance of video on mobile and also monetization. David McKeague has plenty of great insights and its a peek into the future as I’m sure mobile engagement will have video as a centrepiece.

Engagement Score

We score all out interview anecdotes based on 3 axes of: acquisition, UX and retention – I’ve listed the actionable points below in TL;DR but David breaks out “sustained engagement” & re-engagement as the key goals. Therefore this technology is focussing equally on UX and Customer Retention.

Let us know on our Twitter account how you score it.



  • If you want to add video to your App – test it out and figure how you can have it start to roll in 300-400 milliseconds!
  • Relying on video advertising is tough because of the failure of load and slow load times.
  • Its a physical issue of content size (1080p resolution on phones now) and slow bandwidth
  • David’s big fail was the long time it took to get to “product-market-fit” – he explains that they did not understand the problem domain until they spent extended time with Warner Bros.
  • When pitching your product/solution – make sure you know who are the “committed” stakeholders and who are just “supportive”. For David, the Telco’s were interested but won’t drive success for the company – however, Media companies very much need their solution.

Transcript Detail

DAVID: Hello, it’s David from Street Hawk. I’m here with David McKeague. David is, I think, VP of Products of Incoming Media. It’s a really interesting company looking at what the future of video is on mobile. How are you, David?

McKEAGUE: Very good.

DAVID: Tell us a little bit about what problem Incoming is actually trying to solve.

McKEAGUE: Yeah, thanks. Incoming actually came out of research lab called NICTA and was really working on the problem of how would mobile video-watching play out over the next 10 or 20 years. The issue that lives inside, if we look at a video app, typically you get a spinning wheel for one or two seconds as it’s coming out of the wide area network, many times it’s 270p adopted bit right and the end users paying $10 a gigabyte for watching video, and that is a set of technology problems.

DAVID: You’ve got this incredible kind of resolution there with iPhone 6 and Nexus 6, where you got really high res screens.

McKEAGUE: Exactly.

DAVID: So that demands a lot more video density, I guess.

McKEAGUE: It does. In the fixed network, the answer to this was called the CDN, and companies like Akamai, and they created content, they cast towards the edge so that when you press go on a video with ABC iView, it actually happens all very quickly. But that problem doesn’t exist in a mobile network, so basically the solution is prepositioning of content. So we’re pulling forward content video actually into the mobile phone, and that’s because the phone’s storage is growing much more quickly than the wide area network and then you get this instant start, you know, beautiful HD instant experience.

Some of the tricks that we do—I guess we’ll dig into this—is we give media publishers the ability to pull the content when the network is free. What that means is that if you want to use opportunistic Wi-Fi so that’s not coming out of your data plan, you can load. So the broad technique is what’s called “predictive prefetch.”

You’re quite right. What we found when we started two years ago—so we’re a Series E-funded, we have an office in San Francisco, an office here in Sydney—this resonated with the larger Android phones and the note devices, and soon it’s the iPhone 6 and the 6+. You know, getting that 100-millisecond start, minimum 720P rock solid play-out is important to media properties. From a target market perspective, these are the new apps, the sports apps, episodic, short form, MCNs, music media companies, all trying to build businesses around mobile.

DAVID: Okay. So let’s see if I got this right. ESPN application—I’m just making something up here, but say this is an ESPN app—what they want to do is be able to show something from a baseball match but actually give the user instant gratification so that when the user presses the button about that, the video actually runs straightaway and I don’t get the spinning wheel that you’re talking about.

McKEAGUE: That’s right. Basically we’re using ESPN, BskyB. We talk about CWC, we talk about Maker in the US. These are professional companies, video-orientated companies, trying to make money around video-watching apps. Today, because of that spinning wheel and the unreliable experience, people get frustrated and tend to give up earlier than they should.

DAVID: Right, exactly that situation where I just hit the back button if I’m an Android user and I’m just basically out of there because nothing worked for me in that situation. So if I’m the developer of the ESPN application, what am I actually doing there with you guys, dropping an STK into the application?

McKEAGUE: Yes. So it’s an STK, and in this STK is machine learning, so Incoming is still relatively early, so seven staff, five PhDs in machine learning out of seven. [Laughter]

DAVID: Is that all!?

McKEAGUE: “Is that all?” that’s right! So what we’re leaders in is slow power machine learning. So inside the STK, it’s loading what your content interests are, what the networks are doing, and enabling the media company to preposition content.

DAVID: I see, okay. So this prefetch that you talk about, you know, if I’m watching a movie, then that’s going to be on high res that’s going to be a couple of gigs, so I guess you’re saying that you really targeted that smaller things like 30-second vids or something like that?

McKEAGUE: Yeah, so short form, which is typically in our 4, 5, 7-minute videos this works fantastically for. And really the way people “snack” actually on mobile is that short form is the dominant experience. So if we think of news and sports, episodics, so TV can be chopped up into short form, but Netflix long form, this is not an STK for long form.

DAVID: Right. So I’m the ESPN developer, I’ve dropped the STK in, and your STK understands where the short form videos are hosted. You don’t try and do “CDNs yourselves; you just basically decide which particular piece to preload into the phone.

McKEAGUE: Right. And with the media company, they have content metadata systems and recommendation engines, I’m sorry, machine lining plugged into a content metadata systems for prefetching.

DAVID: Content metadata system, what does that mean?

McKEAGUE: Content metadata system, CMS, is actually it, but inside the CMS, it’s the tags that describe the video. So all video properties or media properties will have CMSs. So something that we found is since really November last year, you know, with Apple coming up with the iPhone 6 and the 6+, this has really changed the landscape. This idea, because we enable the media publisher to preposition videos in the background, so that you don’t see it actually as it’s loading, and we can let them load videos into the notification system which we’ll talk about actually into the main app or it can be display ads.

DAVID: So when you say notification system, you mean push?


DAVID: Right, okay. So when I pull down the menu from the top of the phone, I actually see a still of the video and I can press the play button from there or something?

McKEAGUE: Right. For a media company, this is called program promotion. This becomes a natural sort of thought for news property and a sports property is there might be a new sports goal, like a new soccer match, something brand new happening, instead of a text notification, you create a two or three-minute clip and then you push it as a notifications system. Then as you watch that video, it drops your deep link back into the app.

DAVID: Right.

McKEAGUE: So that’s really around the re-engagement. So what we’ve found with media companies, their media apps have a poor mobile engagement and it really fits in to a couple of areas: one is reengagement, the other area is sustained engagement, keeping people in the app for longer.

DAVID: Session times you mean?

McKEAGUE: Session time. And then the third thing, which is very much a media company issue, is the display ads do not play up properly. So if they’re trying to make money out of their media app, they need to keep you in there as long as possible, but they do need the advertising to work. So typically in the US, banner ads are starting to fall away. Sure, there’s native ads, but in video advertising, display advertising, with streaming ads, many times the ad mechanics fail, so that the ad which is trying to make an ad decision and that sort of goes for one or two seconds and then it actually can’t fulfill the ad properly in the time, so it becomes an ad fail.

DAVID: Right, okay, which is like the worst case scenario for somebody like that, that’s where they make their money.

McKEAGUE: That’s right. The media properties are trying to make money from their media apps. They know that millennials are there, so the way this has really kicked into real action in the last 12 months, if you’re a broadcast TV organization, you’ve been defending TV, you’ve been looking at the internet, you know that you should be doing something with millennials actually on mobile, but you haven’t been able to make the business work. You can’t keep people in the experience long enough, the video isn’t playing up properly, the session times are too short, and also the extra part is the revenue side is not working.

DAVID: A lot of stuff wrong there. [Laughter] So much is broken.

McKEAGUE: It is. And when they’re very honest, they say, “Well, we lose money on our mobile apps because we can’t keep people in and the ads actually aren’t playing out, so enough of an ad-driven business.” So in these mobile media apps—and if I just reiterate, news, sports, episodic, MCNs, music—there’s probably only 1200 apps that really matter. So if we went to someone like an ESPN, you know, they’ll have one major app with probably four or five surrounding apps around it. So it’s not millions of apps.

DAVID: But why would that not – I mean, it’s always the thing with startups that something might look like a really small thing initially – but why wouldn’t this be something that is relevant to 50,000 apps in the future? If somebody was sitting out there now doing some sort of social network, some sort of “Vine-ish” kind of thing or something like that, wouldn’t that be applicable for this technology?

McKEAGUE: Absolutely. I mean, you’re right. As a company, we focus really just on a smaller market segment, so that we’re not trying to boil the ocean. But now the technology will go broadly applicable. This idea of…and Chris Dixon wrote a great blog post about November last year about how in the web it was very much a pull model, and then for mobile to keep people in the experience you need to create a push and a pull model. But to get the push right, it needs to be relevant and personalised.

DAVID: Yeah, I think he was talking about the replacement of fundamentally Google was kind of was like the poster child for pull. You go to their site and you search on something and that returns some results to you, whereas the push model flipped over with companies, like Facebook actually pushing data at you, which was fundamentally perceived to be coming from your friends, and so therefore it’s okay to push things at you.

McKEAGUE: Exactly. And so push that is not personalised and relevant is called spam. So what we see is people want to promote. These videos that I’m talking about with push video notifications, they’re promoting new, if there’s something new in your app, so a new sports call, the new piece of content, something new in the game, then instead of sending a text notification, they send a video notification. It connects into their video workflow because they have newsrooms already. And so making video work, I think, is sort of a part of the next big step for mobile, making it work as a medium.

DAVID: So out of the companies you’ve been working with at the moment—because we tend to look on these podcasts at one win and one fail—has there been something where you’ve actually seen a real inflection point or a real success example of where it’s really worked?

McKEAGUE: So the core underlying problem has remained solid over many years. The core problem, many people would say, “Well, won’t the carriers fix the one or two-second start on video?”

DAVID: Yup, good point.

McKEAGUE: Actually the problem is rock solid and it’s fantastic, because there are laws of physics. In mobile—if I can be sort of a bit more brutal about it—if a video takes one second to start, then people are giving up because this is a short-form snacking environment.

DAVID: Yeah, very attention-deficit behaviour for people who on the bus with plenty of other distractions and things like that.

McKEAGUE: Yeah, 300 to 400 milliseconds is a stuck button, so you’re thinking “there’s something going wrong with this thing.” So you need to be consistently 100-millisecond start and the networks can’t really get a low half-a-second or three-quarters of a second.

DAVID: Yeah, I was just actually kind of thinking that Akamai needs to be sitting out at a cell tower. [Laughter]

McKEAGUE: And then you could imagine, so this is the – we know them well. You could imagine a caching point at the top of each the size of the internet and the size of each; it wouldn’t still fix the problem. The only way to do it is to push it into the app or into the phone. So the side effect of what we do is we let carriers, we don’t need to work with any carrier at all, but carriers can then start using – carriers typically have 50% spare capacity, so all of the cell towers are actually all free and vacant overnight. So why not push out content overnight so that when you wake up, you know, you might actually have an extra half-an-hour video that’s being preloaded into the apps? So I think, for us, what’s been fantastic is that there are some laws of physics underneath them but the core problem that we’re trying to solve is solid. We’ve talked to many carriers and many media companies and they’re all trying to find solutions.

I think on the fail side, it’s taken a good amount of time to understand what product market fit looks like, how to get this STK, thinking through the go-to market in enough detail for us. So we’re selected last year as one of Warner Brothers’ media camps. So as a team, we spent three months inside Burbank.

DAVID: Oh, very interesting.

McKEAGUE: Yeah. You know, learning about the media industry and learning about mobile properties from their perspective and what does it mean, what does the value proposition, what does a pitch look like? So we’re a highly technical team but without enough knowledge. Now in that experience in Warner Brothers was transformational for us, in helping us understand what the go-to market means.

DAVID: So you’re saying that prior to that, you were kind of like a technology looking for a problem, whereas working with them allowed you to really understand what the actual problems were at a very tangible and current sense.


DAVID: That sounds like a win, not a fail.

McKEAGUE: Well, it really took us too long. [Laughter] I wish I’d done it the year before. Well, you know, I suppose it is a win. I think another thing which we’ve learnt is Australian media companies are not early adapters. [Laughter]

DAVID: Yes, what a surprise.

McKEAGUE: I don’t know where the entrepreneurs are. So we’d find out those customers are in California, somewhere between San Francisco and LA. I’m not quite 100% sure why that is but they’re all very interested. But Australian media companies aren’t aggressively driving mobile.

DAVID: So can you distill for me what that win was? That was a little bit subtle for me. You’re just saying that basically the win is that you realised that because of the physics problem that preloading overnight was giving you a lot of interest in the telcos as well too because they can basically spread out their demand. Is that –?

McKEAGUE: So the core problem, the technical problem, is rock solid, and that is subscribers are frustrated, media companies can’t make money out of subscribers, and carriers have capacity problems, and so that problem is solid. When I say the fail of the product market fit, understanding between those three actors where to start and who is our primary customer? So our primary customer is media companies trying to build media mobile apps.

DAVID: So the telcos benefit and are going to be supportive but in the old chicken – what is it, chicken and pig, you know the English breakfast where the pig is “committed” and the chicken supportive? Telco’s the chicken.

McKEAGUE: Exactly. And so they’re terrific partners but they’re not our customer. I think just as a startup, the reason why we put that down as a fail is I wish I’d been knowledgeable enough to get to that point earlier.

DAVID: Yeah, it’s something that pops up in these interviews quite a bit is to understand who are the stakeholders in the ecosystem and whether they’re with you or “agin” (against) you. In this particular case, they’re kind of with you but not in the sense that they’re actually going to help move you forward. They’re going to sit on the sidelines and just be a cheerleader, if you like. But they’re not going to actually fund it or help push it forwards in any active way.

McKEAGUE: And I think that when you do original pitches where you put in all the stakeholders into that one pitch and you take it, say, to a media company and you’ve got the media value proposition mixed in with the carrier value proposition, the media company becomes confused.

DAVID: Right.

McKEAGUE: And then you take it to the carrier, and they get confused.

DAVID: It’s got to be “What’s in it for me?”

McKEAGUE: Exactly. And so, you need to separate these and be very clear in your own mind. So this is where the issue is that it took us too long, but we actually understand that now in great detail what this means. You know, reengagement for a media app is a big issue. We wrote and test that and it’s got 1.5 million installs. It’s in the Google Play Store Incoming TV. It also shows how we triple sustained engagement.

So the key metric that media companies are looking for, if I can triple the length of time, the session time, that people are in watching the video, these video apps—whether they’re news, sports, TV shows, MCNs, properties which are on YouTube trying to get into mobile, music video—if I can triple the overall length of time that you’re in the app, then that transforms their business. So we talk about tripling their reengagement capacity, tripling their sustained engagement and being able to affect the revenue side. Those three messages are messages that media companies understand.

DAVID: All right. So if somebody wanted to reach out to you guys, what is the best description of those people today? Is it an MCN, is it a media company, is it somebody that’s got a radical new idea for video in a social network, or some other contact?

McKEAGUE: You know, all of those. They need to be video-orientated properties. People are trying to make money out of mobile video.

DAVID: So video has to be at the center of what they’re doing in that situation?

McKEAGUE: That’s right.

DAVID: And they may have already identified that this is an impossible problem to solve because of the preload issue or the latency issue of the front end.

McKEAGUE: Latency issue, that’s right. We can affect reengagement, sustained engagement, or the advertising side. So I’d be very interested in talking to people.

DAVID: All right. Well, thank you very much for that. I think it’s a pretty good snapshot of what’s going to happen in the future. It looks like you guys are pretty well positioned to really help people do some interesting things with vids, and it just seems logical to me that that’s going to be a real sort of central piece of the way people engage with mobile apps in the future.

McKEAGUE: Fantastic.

DAVID: All right, Dave, thank you very much. I’ll talk to you again soon.

McKEAGUE: Thank you.

DAVID: Cheers, mate.

  Get our Referral Program Guide