Mike Rowehl: This is Mobility

Oliver seems to be really into Bluepulse. After I saw the posting at MobileCrunch I signed up for a Bluepulse developer kit. I’m one of those propeller-headed folks, being of the general mind that knowing how things work out under the covers normally helps understand how the system as a whole functions. But I’m left blinking at a mostly empty screen when I take a look under the covers of Bluepulse. Then I saw the post at MobHappy and decided I needed to figure out what I was missing.

I was expecting something along the lines of the Opera widget toolkit, there were a couple of comments to that effect, and the talk about the platform making applications work on a wide variety of devices. Inside the Bluepulse developer kit I find a description of a stripped down set of XHTML, a set of static stylesheets that the “platform” includes, some extra HTTP headers, and very minimal handling of multimedia types (mostly by just downloading them to the phone). “This can’t be all of it” I thought to myself. So I dug around for a while longer in the docs and examples looking for what I could be missing from the tech end. I’m just not seeing anything. No Javascript or other scripting languages, no real extension of the display language on the client side and no support for the interesting aspects of the existing spec, PNG images only! Overall a long step backwards and at least a short step to the side.

I can’t bring myself to think that this method of delivering applications to mobile devices is the future of the mobile web. So what am I missing here? It’s not like we haven’t seen attempts to mobilize application using a subset of standard web markup and some client side shim to display it. The argument against it last time around was that no one wanted to develop a crippled version of their service specifically for mobile devices. There was also the problem of install base, which I don’t see Bluepulse being able to address. Though maybe they have an in somewhere that’ll get them on all the decks. Whoring themselves out to carriers would certainly do it, but based on the answers from their VP of Answering Questions that doesn’t seem like the route they’re interested in.

So maybe there’s an evolution in the business model that I’m not getting. The little applets seem to be pay-per-day usage charges using a credit system. The kinds of content in there - restaurant and bar info, TV listings, movie locations and times. Nothing that isn’t getting delivered via the likes of Yahoo or Google mobile at this point. The only real differentiator there is detection of phone types so you can do ringtone and wallpaper downloads. So of course promoters and advertisers should be very happy to have a channel like this. But they aren’t going to drive the channel in terms of adoption of an application that folks have to install on their handset. And the use as a promotional tool is out of line with the credit system which seems to be one of the major points that Bluepulse is pushing.

So what else is there? Is this a tool for MVNOs to use in a way that I’m not picturing? They have the reach to get this installed a the kinds of numbers of handsets that would make it a compelling development platform, and the right kind of demographic segmentation to make it worth the publishers while. If that’s it then what’s the prototypical app for this and what’s the cost model? Any propeller-heads out there who picked up something I missed? Or is this not a tech play at all?

Carlo has a great follow on to the recent discussion about mobile advertising:

For instance, there’s a “channel” on the service from Adult Swim that has a bunch of short video clips, and Sprint charges $4 per month to access it. $4 isn’t necessarily a high price, particularly if you’re an Adult Swim fan, but $4 here, $3 there, $8 over here adds up pretty quickly. What about offsetting these prices, or eliminating them by showing a brief targeted ad before the clip? Sprint wouldn’t have to sell much advertising to make back that $3 per month.

I totally agree with that, and would like to add a bit. The problem from the carrier end is normally one of capacity. Their system just can’t handle the volume necessary to model themselves on broadcast advertising. Depending on who you talk to you’ll hear somewhere between 20% and 3% thrown around as the high water mark for data activity on EDGE and 3G networks. I tend to think the number is somewhere around 7% to 10%. What I mean is that if just 7% of the people who are subscribed to highspeed data services were to use them at the same time for something like video download the system would grind to a halt. And if you think that 7% of people would never be active at the same time just think about 9/11 or the Michael Jackson verdict. Or the final moments of any given season of American Idol. If people could whip out their cellphones and see video for events like that they definitely would.

There’s been a lot of talk about this whole mobile ubiquity thing, the assumption is that these “always on” mobile data services are scalable solutions. But that’s just not the case. Networks reach capacity and cease to function, and particularly under the model the FCC chooses to enforce for cellular services here in the US bandwidth is a scarce resource. Unlike the television and radio models, there is an incremental cost for a new user on the cellular networks. Adding a subscriber consumes an increment of resource, whereas adding a subscriber under television and radio models did not reduce the number of watchers/listeners the system could add in the future.

Advertising is normally based on a large number of customers. The more readers you have the more interest advertisers have, the more effort you can put into getting readers and improving your content, which generates more advertiser interest, etc. That model is based on being able to keep adding more and more customers. In the first round of Internet advertising it was a major realization for some folks that there was a non-zero value in serving a pageview to a reader. In the early days banners were paying so much that it was pretty much trivial to add network bandwidth and server capacity as more readers arrived. But as that market constricted only the sites providing high value were able to keep their advertising inventory prices well above their provisioning costs.

What we have in the mobile environment is effectively a constrained potential customer base as well as a very high provisioning cost. It’s going to be very difficult to make a 15 second commercial clip pay off on cellular networks. Because “scheduling” that ad doesn’t mean tossing it into a slot where you know about 3000 people are tuned in, it means transferring that 200 to 300K of message (and I’m just making up that number based on 3gp video sizes) to 3000 individual handsets. That’s 600 megabytes of transfer over the network where bandwidth is a scarcity, just to get out one 15 second clip to a relatively limited audience.

I do like the idea of advertising supported mobile services, as I’ve mentioned before. However I think the existing television advertising model is a poor match for what the cellular video advertising model would have to be. It is out of necessity a premium model. But expecting the customer to bear all the cost of that premium service is wrong. It’s a new channel with radically different and powerful attributes. I’ve yet to see anything that really attempts to exploit the new medium to full benefit.

Interesting essay over at Network Publics titled Beyond Locative Media:

In contrast, Sterling provides us with a darker, more idiosyncratic vision. Humans don’t control Sterling’s world of Spimes. On the contrary, it is an unruly object world in which people are, at best, ”spime wranglers.“ At the dawn of the Internet of Things we have to wonder if we aren’t entering a world in which the object becomes sentient, thereby finally liberating itself from human bondage. If, in the Enlightenment, we learned that nature—in its role as background to human activity—had been replaced by human second nature, then today we are perhaps at the threshold of a machinic third nature. It is the task of whatever remains of art after the locative turn to get involved in the messy business of this new world of objects, even if the Utopian and critical moments that can emerge as a result are only temporary and contingent.

The overall message I took out of it is that the current round of “media” is much less interested in distancing itself from everyday practices. However I can’t help but think about the massive gulf that I see between the issues written about in that article and the kinds of questions I’ve been getting in preparing for the Mobile Monday on location based services. There just still seems to be such a disjoin between the two that they don’t even seem like the same topic.

I’ve posted about Google Talk before, and bitched about them having federation off somewhere else. Now they’ve turned on federation, which I just verified with a chat between my gmail.com account and the server I have at bitsplitter.net. It even seemed to carry presence info well. I like. I like a lot.

Speaking of mobile advertising, Omar from fotochatter just posted to the Mobile Monday SV Yahoo group announcing AdMob. It’s a pay per click advertising service focused on mobile sites, with targeting specific to mobile clients. Here’s a bit from their site:

Buying mobile advertisements

• Reach consumers right on their phone or other mobile device.
• Target ads by region, manufacturer, platform, and device capability.
• Personalize ads using device model number.
• Easily build a quick mobile ‘MobPage’ to reach customers even if you don’t have a mobile site.
• Pay only for the click throughs on your ad.

Phonescoop posted about a few carriers trying out advertising, seemingly as an opt in feature and delivered as video. I’m actually all in favor of advertising supported services on the mobile end. Contextual advertising has brought online advertising in general out of the gutter that it was in. Businesses are starting to spring up which are viably and sustainably based on advertising revenue. I think advertising and/or affiliate based services on the mobile end makes good sense.

What I’m concerned about is that carriers just are not good at working with alternative business models. I wouldn’t be too surprized if we saw carriers trying to charge for the data traffic caused by delivering ads, as the folks at Engadget have joked about. I think advertising would work, but it needs to be done correctly. And like everything else in the mobile environment it’s made somewhat more complex by the limited devices and relatively low speed and high latency networks involved. I also think there are a lot of opportunities for media companies and established brands looking for additional reach to provide content to carriers. The PSP portal and XBox online both offer free downloads of movie trailers. I think that kind of content would be very well suited to handset delivery. As long as we can keep the carriers from trying to charge the media companies to put out the video AND the users to subscribe to the service and for the bandwidth used during the download I think there would be some interest. Overcoming the user expectation that their carrier is once again trying to screw them also factors in there.

There’s a great post over at MobileCrunch about a mobile interface for YubNub. YubNub is a pretty interesting application, I fooled around with the standard version for a while when I heard that Jon was doing some work for Ning. Use from a mobile is certainly one appealing case for us hackerly types. And the stripped down mobile version is certainly much more appealing than what skweezer does. Interestingly enough, the Google Mobilizer just errors out on trying to hit YubNub. Shrug.

I would highly recommend following MobileCrunch if you have an interest in the segment. TechCrunch, which I think was the first blog in the TechCrunch network, grew into one of the best sources for information about early stage companies in a matter of weeks. I’m sure the same will happen with MobileCrunch for mobile companies.

Although of course the lines get more and more blurry all the time, there are two main classes of devices in terms of mobile solutions. I was thinking about this the other day as I was moving around the slew of devices I have sitting (mostly unused) on my desk at home. There are some devices that are meant to be general mobile computing systems: the Palm devices like the Treo650, Nokia smartphones like my 6680, the GP2X, the Nokia 770. And then there are the devices that are meant to be endpoints for some service or controlled channel: the Apple iPod, BREW phones, the Sony PSP, and the NintendoDS.

I tend to call the devices that function as service endpoints “pipe extenders”, because they are generally meant to serve as a way for someone with an existing distribution channel to extend their channel beyond the sedentary devices they already control. The iPod is the way to make music purchase online compelling by allowing the online store to stretch beyond your desktop/laptop, the PSP and DS are ways for Sony and Nintendo to fill up your leisure time when you’re not in front of the tube. In general devices from this class tend to be rather strict about the formats they accept for data and media, they tend to be unabashedly coupled to their upstream connection, and if you can install new apps on the devices they tend to be signed or controlled in some way.

General mobile computing devices are meant to serve more as open platforms. In general they tend to support a variety of formats for the media they operate on, they allow for installing applications by the user, and a motivated individual can create their own applications for these devices without resorting to hackery. Interestingly enough, this is not a new class of devices. The PDA market in general has fallen into this category for years, which is perhaps why that market has never really hit critical mass. Currently these devices tend toward the geeky, tailoring more to expert users than the novice. This doesn’t have to be the case I don’t think. But without someone controlling the channel to the device there’s just little reason for it to be otherwise.

I’m not saying this is the model to dictate everything, but it is a useful model when looking at the cellular industry. Cellphone carriers for a long time had pipe extender devices on their networks. A cellphone was a device which allowed the connection of voice communication when the user was away from their desk or home phone. However as the devices became more complex and gained data functions they started to look more and more like mobile computing systems. If the cellular carriers hadn’t been asleep during this phase they probably would have realized they were missing the opportunity to control the pipe for data services as well as voice services, but it’s my belief that they both didn’t value the new channel and didn’t care about it.

Recently they have started to care about data services. The traditional cash cow of cellular voice services is being threatened, and carriers would like to control the data channel as well. In many cases however, they just don’t have the background to do so. The data channel is mostly a media channel, and the cellular carriers just don’t have the content necessary to fill that pipe. So in general they’ve taken to strangling off that pipe so that no one can use it. Wondering why it is that smartphone usage in the US is still so low even though the ARPU numbers say that smartphone users are great customers and carriers should be pimping smartphones to raise their revenue per user? Wonder why when you walk onto the floor of a carrier store in the US there are normally very few smartphone models and they’re often tucked into some corner? Because I think they know that smartphones break the channel they’re hoping they can eventually control.

There’s a wildcard in there with the MVNOs however. In many cases they do have an existing channel, and they would like to extend it to mobile users but they don’t have a pipe to do it over. If the MVNOs can generate enough off the data services that their share with the carriers is larger than the revenue the carrier would get from supporting a general computing device the carrier should win out. However there are a few big “if”s in there. When the carriers controlled the voice channel they did so because it was the only communication available, and federal regulation said that another alternative network couldn’t be set up. However with data communication it’s much easier for a third party service provider to disrupt the carrier/customer/MVNO relationship with an alternative offering. The carriers really need to keep a hold of the platform and restrict the activity on their networks in order to make the environment appealing to the MVNOs.

So how does this play out long term? Even without the carriers working against smartphones and general mobile computing devices, I think those devices have generally been ill suited to use by the casual user. It’s starting to look more and more like a setup where the high end traditional offerings from entrenched players like Nokia, Motorola, Sony Ericsson, etc. are ripe to get picked off by inexpensive handsets from small manufacturers that mix in just slighty more than bare functionality and support the point-solution needs of both MVNO and carriers without overshooting. Because overshooting in this case means not just wasted costs, it willfully introduces competition for the MVNO who makes the decision to carry the handset. Ooops, that doesn’t sound like a nice environment at all for us users. Maybe we’ll have to count on 802.11 to save us.

802.11 has tons of problems: licensing, quality of service, capacity, call handoff, etc. That’s the funny thing about disruption though. It often starts out as something not good enough for the current environment, but it evolves into something sufficient while the current environment is tying itself in knots trying to serve a less and less significant portion of expert users. That’s why I think Nokia offering the 770 isn’t just technically interesting, but a great strategic move. Makes me wonder if someone over at Nokia ran some scenario planning and decided to take the “what if 802.11 does actually solve the current round of issues?” question seriously. They say now that the 770 is a different market than their handsets, but I think long term we could see that change.

I’ve had Yahoo! Go on my phone for a while now, trying it out for syncing my address book and calendar in particular. But I’ve been having a bunch of trouble with it. Today it’s refusing to sync a calendar entry that I added earlier this morning. The entry is on my calendar online, but it’s not appearing on my phone. So maybe I’ll use the app every once in a while to push my calendar and contacts online, but it’s not working for syncing. Here’s the breakdown, for anyone at Yahoo who might be interested in a list of gripes:

• It’s too big for the memory profile of the phones you’ve chosen. If I’m running just about anything from the Yahoo Go suite I can’t take a picture with my phone. Because Go really wants to keep itself up and running all this time this is very annoying.
• The initial options for setup are hidden. When I reinstall the app there is no setup for username and password. You need to hit one of the icons on the Go screen to get into the setup dialog. Why completely hide this? It’s just confusing for folks who are familiar with the app and think they know what to expect.
• It’s not “just working” for the calendar stuff at least. Why aren’t my entries from online getting pulled down to the phone? I have no idea, and as far as I can tell there’s nothing to give me additional info. The transfer says it worked, if I add a test entry from my phone it shows up online. But not vice versa.
• I want a simple “manual sync now” option. The “manual” option allows me to set options when to sync myself, not explicitly request syncs directly.
• The inter-application communication could really use some work. The tool seems to be structured as a set of apps that try to communicate with each other. And if you try to set options in one app and then go check on another, the status isn’t synced up fast enough. Sometimes I would go into Go and set an option, and then go to the connection app and check on the status only to find my recent changes weren’t reflected. So I’m stuck killing everything off, starting up Go to set options, and launching Connect just to make sure my tests are really running with the options expected.
• Along those same lines, having the options within the Go app be dependent on if Connect is running or not is problematic. First of all it’s just confusing to change options in one apps softkey menus based on if another app is running. Second, what if I want to change my filter options BEFORE I start to connect? I launch connect and it connects when it hasn’t been up for a while. Put that filter options dialog in the Go app itself and allow it to be used without Connect running.

So for now, unfortunately, Go just doesn’t work out for me. I have it disabled, and given how much of a pain doing a manual sync is I probably won’t use it. Can’t wait for the next release though. I would love to use the app, it’s just currently too much pain to do so.

The topic for the January meeting of the Symbian Programming SIG is “Sockets and Active Objects”. Here’s the essential info from John Kern’s blog:

• What: Symbian Programming SIG meeting (Topic: Sockets and Active Objects)
• When: January 31st, 7PM
• Where: Mozilla Foundation, 1981 Landings Drive, Build K, Mountain View, CA - directions
• Speaker: John Kern
• Audience: Anyone interested in programming Symbian devices.
• Cost: Nothing!