Mobile video: Is there a there there?

[Reposted due to a correction. Sorry if you get this twice on your feed.]

I recently I spent a couple of days at the Global Mobility Roundtable, an annual conference that brings together mobile-related academics and a selection of people from the mobile industry. This year's conference was in Los Angeles, so it also drew a number of attendees and speakers from the major entertainment firms. It turned into a kind of a mobile meets entertainment event, and the results were interesting. Mostly, they underlined how far we still need to go in bridging the gaps between the tech industry, mobile, and entertainment.

There's a lot of information to cover, so I'm breaking this post into two parts: mobile video in this part, and in part two the status of mobile data in general and the relationship between Hollywood and the operators.

Is there a pony in the stable? If so, it's a very small pony.*

There was a lot of disagreement about whether mobile video will take off, which may be just as well because the economics of it are seriously dodgy. It's not certain that users really want it, no one knows whether the revenue will come from sponsors or from user fees, and even if video does take off, it's not at all clear that the mobile operators can deliver it without bankrupting themselves.

Other than that, the prospects look great.

One panelist compared the situation in mobile video to a company running a health club: they want to sell a lot of memberships, but they don't want anyone to actually use the facility.

The information below is drawn from a series of different sessions I attended. I've mashed them together so I could organize the information by topic. All quotes are as accurate as I could make them. They are definitely correct as to message, but I probably missed a few words here and there.

Who wants mobile video? A segment of the market.

There are plenty of people in the industry who are enthusiastic about mobile video. One presenter quoted Rob Hyatt, executive director of mobile content at Cingular, as saying, "Watching video on cell phones could eventually easily surpass [demand for games, ringtones, and wallpapers], to reach 100% of the population." That's pretty remarkable, since even SMS doesn't reach 100% of the mobile population yet. (You can find the original quote from BusinessWeek here).

Telephia, a mobile industry research firm, reported that revenue from mobile video is growing rapidly, from $35m in Q3 2006 to $146m in Q1 2007. In that same period, the number of mobile subscribers in the US using video services grew from 5.7 million to 8.4 million (for comparison, there are 77 million MMS users and 148 million SMS users). The Telephia numbers imply that revenue per video user has grown from $2 per month to $5.80. Unfortunately, they didn't give any details on which particular services are growing.

The base is still very small, so it's dangerous to extrapolate from those numbers. But they're definitely hopeful. A number of other speakers were much less optimistic, though.

At the conference, USC presented the results of the sixth annual Worldwide Mobile Data Services study. It showed that about 30% of 18-24 year olds and 20% of 25-34 year olds in the US felt that video downloads to mobiles were an important feature, about the same percentage as wanted games on their mobiles. That's nice, but not the universal usage that Cingular talked about.

Sanjay Pothen, CEO of Pliq (a mobile video production company), claimed that 44% of mobile users are interested in mobile video -- but only 4% are willing to pay for it. That's the typical pattern for mobile data features -- most people don't want them if they have to pay anything for them.

Frank Chindamo, CEO of Fun Little Movies, which produces short video for Sprint, asked the audience how many people in the audience had Sprint phones. About five people raised their hands. "If you all subscribe, that will double our revenue for next month," he joked. [For the record, Frank asked me to make clear that he was only joking; he says he's actually quite happy with the Sprint relationship.]

Is the glass half full or half empty? As I've said before, I think there's abundant evidence that the market for all mobile data products is highly segmented, and we need to learn to make money from products that appeal to ten or fifteen percent of the users. I heard nothing at the conference to change that view.

But overall demand for mobile video is just the beginning of the story...

What sort of video will people watch on mobiles?

This one is still very much undecided. The usual assumption is that because short video is popular on the Web, it'll also be popular on mobiles. For example, Funny Little Movies is creating original short animated films for mobiles. (The place is run by a USC film professor who has his students create a lot of the content.)

Pothen of Pliq said the ideal sort of video for mobile is neither short individual clips (like YouTube) or long-form video (like a TV show), but chunked content -- an engaging story told in two-minute segments. He said excerpts from reality shows can work well -- highlights from America Idol, for instance. But original content seems to be his main target: soap operas, telenovelas, and cooking for young women, comedies and dramas for young men. The goal is to get people hooked by an ongoing story so they'll keep coming back to watch every segment.

Derek Brose, SVP of business development for Paramount Digital, was also excited about short video. He said the company is cutting all its movies into clips of different lengths, for various mobile usages. Two second clips -- something like Harrison Ford saying, "trust me" -- are for embedding in an MMS message. Twenty second clips are for use in ringtones. Two minute clips are for streaming your favorite scene from a movie. Paramount's goal is to teach consumers a variety of different things that they can do with mobile video.

But some people were skeptical about the prospects for short video on mobiles. Bill Sanders, VP of mobile programming at Sony Pictures, said that in Japan people are watching broadcast TV shows on their mobiles rather than short video streamed over 3G. He said 3G in Japan is great for certain kinds of applications, such as e-wallet. But he said data is priced so high that streaming video barely exists on 3G at all.

"The only thing you find in 3G is porn, because it's the only form of video where people will pay $10 for three minutes of content." --Bill Sanders, Sony

USC's mobile survey also strongly implied that the biggest demand is for broadcast TV. More than 40% of users said they thought that was the most interesting type of video for a mobile, compared to about 20% for short video.

David Tilson of Case Western University supported that view. He said that in a UK test of DVB-H (a broadcast video standard for mobiles), users watched three hours a week of television on their mobiles, with viewing concentrated in the lunch break and commute hours. That's very intriguing, because it implies that mobile video might add new television viewers at times when people don't usually watch TV. Unfortunately, the users were not charged anything in the test, so it's very hard to tell how much usage mobile TV would get if operators started charging for it.

I have no clue what the answer is on this question. People may say they prefer broadcast television just because that's what they're used to. Their actual purchase behavior might be very different. I think price will make a huge difference in adoption, which brings us to the next subject...

Who will pay for mobile video?

You've got two choices -- users pay, or advertisers pay. There are good arguments on both sides.

Sanjay Pothen of Pliq made an interesting case for having the advertisers pay. Since his company is involved in that business, his argument was not a surprise, but it was still interesting.

Pothen claims that neither paid nor ad-supported video are taking off today in the mobile world. As I noted above, he said few users are willing to pay for video, which stops the user-funded scenario right there. But ad-supported video is also problematic on both PCs and mobiles because users are not very tolerant of watching even a short commercial in order to see a two minute video. So what Pliq does is build the sponsor into the video itself, through placement and other promotion within the video.

Pothen said advertisers are willing to pay significant sponsorship fees for these videos. He wouldn't go into details on his financials, but someone I talked to privately said the revenue can be dollars per viewer for a three-minute video. That's impressive, and far more than you could charge a viewer for a few minutes of video.

Unfortunately, Pothen said, the operators want to take 50% of the revenue from these videos. He said that's not acceptable, that the revenue split should be more like 20% of revenue to the operator. "If we work in collaboration and the walled garden is down, we're willing to create original content (for mobiles)....We can drive mass adoption." But he said that won't happen in the current revenue situation.

My take: I don't think it has to be one or the other. Apple's selling a lot of video downloads to iPods, and that won't just dry up. But I think it's going to be very hard to make paid downloads the leading mobile video product, because they'll be competing with free video from places like YouTube, and because ad-supported TV teaches people to expect their television for free. Besides, if advertisers really are willing to pay dollars per viewer, there's no need to make people pay.

The revenue split is an ongoing problem in every mobile data category. There's no immediate solution, at least in the US. I think we're stuck in a chicken and egg situation in which the revenue split discourages the kind of programming investment that might drive a lot of usage, thereby justifying a more generous split.

That may be just as well, though, because video might break the mobile networks if it did take off.

Can mobile video be delivered?

This was the most disturbing topic of all. Even if we can find the right users, the right product, and the right pricing scheme, most of today's 3G networks are not well suited to delivering video.

Tilson of Case Western quoted some very sobering statistics on the economics of mobile video. He said one megabyte of data delivered as SMS messages yields £268 of revenue to an operator in the UK. That same megabyte delivered as video yields 20 pence of revenue, roughly 1/1000 the revenue. Of course, a single user of video is much more likely to consume a meg of data than is an SMS user, so the billing per user might still be fairly good. But video quickly exceeds the capacity of a typical 3G data network. He said no more than six viewers per cell can watch video at one time, and if 40% of users on a typical 3G system watched six minutes of video a day, they would saturate the entire network.

Hardly the basis for achieving Cingular's dream of 100% viewership.

Some of the operators at the conference confirmed this perspective. Francois Thenoz, Director of Strategic Marketing at Orange, said it takes seven minutes to download a 60-90 second video clip on a standard 3G network. 3G "evolved" takes 90 seconds (so you can just about stream in real time). The CDMA 1X network I use to connect my notebook PC is a lot faster, but GSM is the standard for most of the world, so his point was that in most places the wireless network simply isn't ready for video.

Higher-capacity networks are in development, of course. But Tilson said that in the UK, spectrum for a DVB-H wireless video system won't be available until 2102 at the earliest. That implies that for the next five years, mobile video in the UK is more of a science experiment than a serious commercial project.

In the US, the functional equivalent of DVB-H is MediaFlo, which is already deployed in Verizon's VCast system. MediaFlo transmits video one way, using a separate wireless signal, so it gets around the network saturation problems you get in 3G. Similar systems are already being used in Japan and Korea, and reportedly account for most of the mobile video usage there.

A drawback of the broadcast technologies is that they're not streamed on demand. You watch whatever's been programmed at that time. It's like a cable television system, but with far fewer channels. Tilson said one driver of mobile video usage is the availability of a lot of different programming, so limits on the number of channels might eventually restrict usage.

The other challenge for broadcast systems like MediaFlo is that they compete with people using SlingBox or similar products to retransmit their home cable television signals to their mobile devices. "Why get HBO Mobile when you can already get HBO home slinged to your phone?" asked Sanders of Sony. He pointed out that the Three network in the UK is bundling Sling services with its flat-rate 3G service offering.

"Three is like an airline that just bought a bunch of 777s and now they're flying with a bunch of empty seats," replied Brose of Paramount. He claimed that Three has to be betting that video usage will grow slowly enough that faster data networks will be available before the usage of video saturates the network.

The "encoding nightmare"

Then there's the question of standards. Unlike the PC, there aren't one or two video standards for mobiles. Because of the huge array of different screen sizes and software environments, a company that wants to stream video to mobiles supposedly needs to encode it in up to 150 different formats (seriously, that's the figure I was given by a couple of people). An executive I talked to called this the "encoding nightmare." Some companies are starting to offer server appliances that encode the video in real-time from one or a few base formats. But this adds expense to the business model, and real-time encoding is not as high-quality as pre-encoded video, especially if you're trying to compress the video heavily -- which is exactly what operators need to do in order to conserve bandwidth.

What does it all mean?

I think there's a role for mobile video, but considering the limits on user interest, and the huge technical and business challenges, it's not going to be the great horizontal application that drives the mobile data market. At best, it'll be a nice add-on for entertainment-focused users who want video in addition to their MP3s and games.

*This is a reference to an old joke about a boy who desperately wanted a pony. One day he saw a stable stall full of manure, and began furiously shoveling it out. "What are you doing?" his parents asked. "Well," the boy replied, "with all this manure, I figure there has to be a pony in here somewhere."

Good deal: Palm's new ownership

Several people have asked what I thought of the recent change in ownership at Palm. I don't have any inside information, so all I can do is speculate like everyone else, and try to apply the lessons I've learned from working at other companies.

Overall, I'm very happy for the folks at Palm, and cautiously hopeful about what this might mean for the company's prospects. I think this outcome is a lot more encouraging than any of the buyout rumors that were floated in the last few months. Palm's new part-owners clearly understand the value of systems design, which is Palm's biggest potential advantage in the market. I think we really need another great systems company to challenge Apple, and I would love to see Palm step up to that task.

Although a purchase by a Motorola or Nokia would have been very entertaining from a soap opera perspective, they don't really understand systems design, and it's very likely that they would have digested Palm without a trace. I'm reminded of a joke we used to tell at Apple in the 1990s when there were rumors that IBM would buy the company:

Q: "What do you get when you combine Apple and IBM?"
A: "IBM."

The other buyout option what was circulating, a full purchase by private capital, would have left the company independent, but with a load of debt that might have been crippling. Hardware companies must have a big reserve of cash to fund inventory and tide them over if they launch an unsuccessful product. I don't pretend to understand all the terms of the Elevation deal (they're wickedly complex), but from my perspective it looks like the financials aren't crippling. I am a little worried about Palm's cash levels, though; a lot of their current cash is going into the stockholder payout.

A couple of other thoughts on the impact of the deal:

Bye-bye 3Com. Palm gets three very well respected people for its board, and removes Eric Benhamou, the last vestige of the 3Com legacy. Somewhere I have a photo of the Palm and PalmSource combined management teams from just before the two companies were separated. The photo includes everyone in the company from Mr. Benhamou down to senior directors. That was about 30+ people. Every single one of them is now gone. So if you didn't like Palm's management back then, you should take another look at the company because it's now 100% different.

Irresponsible speculation about politics. After a change like this, the standard sport in Silicon Valley is to speculate about what it means for the job status of the people involved. In that vein, the thing to ask is, "Who's running Palm in the long run?" The weirdest part of the whole Elevation deal is the arrival of Jon Rubinstein as both Chairman of the Board of Palm and head of product development. As Chairman, Jon is technically the boss of Palm CEO Ed Colligan. As head of product development, Jon technically reports to Ed. So Jon is kind of his own second-level manager.

That feels...unstable.

Palm seems to now have a surplus of product leaders. Jon is in charge of product development, Jeff Hawkins is the designated product visionary, and marketing SVP Brodie Keast is supposed to control the product road map, according to the press release Palm issued when he was hired. It's hard to picture a car with three steering wheels. Who will really be in charge? In the conference call Palm said that Jon would be the execution guy and Jeff the visionary. "The combination of those two guys is one of the most dynamic... combinations on the planet." Maybe. Any organization structure can work if the people involved get along well, and I presume they would not have made this arrangement unless they were all comfortable they could work together. So good for them and best wishes.

But if you want to be a cynic, you'd speculate that Jon probably didn't leave Apple just to be the head of engineering execution at a much smaller company. You wonder if the current situation is just a stage in a longer-term changing of the guards at Palm. I don't have any evidence that's the case, and I am not trying to start any rumors. But when you see a nonstandard reporting structure like this, it usually triggers speculation that another shoe is going to drop later.

Only time will tell.

What's the effect on products? That's the most important question, and it's impossible to answer at this time. Hardware product development usually takes 18-24 months, so the earliest Jon could change the Palm road map would be very late 2008. But that's the middle of the Christmas selling season, and you can't announce products then. So realistically, the Rubinstein product era doesn't start until spring 2009.

In the meantime, there's a lot he can do to make the development of the currently-planned products be more efficient and predictable. Palm has said publicly on numerous occasions that its on-time product delivery needs to improve, and presumably Jon can help with that.

But personally, I think Palm's bigger problem has been its lack of innovative new product designs. Unless Palm has a bunch of surprise products already in development, it will take quite a while to turn around the product road map.


Thanks to Twofones for including last week's post on the Palm Foleo in the latest Carnival of the Mobilists (link).