Mobile Search
June 10, 2008
I try to steer clear of mobile search. The notion is broad and like most terms used to describe information retrieval the phrase mobile search is frequently undefined. The idea, I assume, is that everyone knows what mobile search is.
I asked my neighbor what mobile search was, and he said, “I just use my phone for calls.” Functions like sending a query to Yahoo’s mobile service aren’t used very often by me, not at all by him, and probably not by you, gentle reader, either.
But if you you get text or graphic information on a mobile device, it’s mobile search. Most pundits feel that this definition is close enough for horse shoes. The problem is that it is the equivalent of cutting a cherry pie with a Husqvarna 455 Rancher chain saw, a popular model here in the hills of Kentucky.
This is a photograph of a Beyond Search programmer expressing dissatisfaction with the mobile search function on an Apple iPhone and a Treo 650. “Both are terrible,” says ArnoldIT.com’s chief technical officer.
The USA Today business section ran this front page story on June 10, 2008: “Are Google, Yahoo the Next Dinosaurs?” I couldn’t find the story on USAToday’s Web site. If it does appear online, I think this is the link that will display it for you. If you can’t locate this story online, you may have to hunt for a tree-unfriendly printed version.
The story, written by Leslie Cauley, is that “many [vendors are] on the hunt for a way to cash in on wireless search.” The idea is that no one, not even Google, Microsoft, or Yahoo have cracked the code for mobile search. The “dinosaur” part is a bit of color. The notion is that because neither Google nor Yahoo have cracked the code for mobile search, these two firms could be left in the dust by younger, more hip innovators. Ergo: Google and Yahoo become the brontosauri of online with regard to mobile search. Ms. Cauley mentions an up-and-coming company called Medio, careful to explain that this is just one interesting company among many. You can read more about Medio here. Could Medio be the next Google?
Because mobile devices are more plentiful than other types of computers, whoever cracks the code can make boat loads of cash selling ads to mobile phone device users running search. I’m not going to cite USAToday’s statistics. I have heard that Gannett takes a dim view of old researchers tapping into their high-value statistical data captured in bar charts without data tables.
I urge you buy “America’s newspaper”; make Gannett’s accountants happy.
The challenges of mobile search are formidable. There are established business models ossified in the American telecommunications industry. There are device issues; namely, screens smaller than the 48 inches of flat panel I have in front of me at this moment, lousy keyboards, and users who aren’t too keen on taking time to paw through a laundry lists of results.
The June 10, 2008, USAToday article triggered a lunch-time conversation today. When I returned to my desk, I rummaged through my collection of mobile search data, became inspired, and decided to create this short essay. What I propose to do is identify several issues associated with mobile search and then offer some observations. Relax, Ms. Cauley, I’m not commenting on your story, just using it as a spring board. I even think the headline was a bit of editorial magic to increase the chance that a reader would dive into the article. I think Yahoo is more like a snail darter; Google is not a dinosaur, but it does have movie monster characteristics which is the reason I refer to the company as “Googzilla”. Now mobile search.
Mobile Search Issues
Let’s run through the items I plucked from my mobile search files. A quick fact: more smartphones shipped in 2007 than PCs. In two years, there will be somewhere in the neighborhood of a billion smartphones in the hands of mobile users.
Now the issues:
First, mobile search begs for a definition. In fact, “search” will be different on different mobile devices. Let me illustrate. My colleagues and I fiddled with an Asus eee700. The device had a reasonable screen and a keyboard that was light years better than the buttons on a Treo, BlackBerry, or iPhone. Once a browser rendered a search page–search.live.com, if I recall the default the owner set for the device–mobile search wasn’t much different from “regular” search. With a wide range of mobile devices available, this example makes clear that the general notion of “mobile search” must be defined in terms of the device’s capabilities, not left as a zippy sounding phrase that somehow defines itself. This means that there must be different types of information access for different types of mobile devices. These will range from mobile phones to GPS gizmos that let me search for information about roads closed north of Seymour, Indiana. (As an aside, this type of search is truly terrible as I learned when I got caught in the great Indiana flood of 2008. My GPS couldn’t find a route, so I resorted to a paper map purchased for $20 at a truck stop. The paper map worked; the mobile device was useless.)
Second, Web search in general is not too good. Google offers a pretty good Web search and because it lacks significant competition in the free Web search arena, Google has more than half the market. Data here. Mobile search is a sub set of search. My view is that “regular” search isn’t very good. It follows that mobile search is not very good either. Therefore, it is not surprising that a large number of companies are working to crack the problem. Mobile search, as a subset of search, is going to be a laggard for the foreseeable future. This does not mean that specific types of search systems tailored to specific types of mobile devices won’t become available and gain traction. Solutions will emerge, but no vendor can rest on a single mobile search success. A number of battles must be fought before the war is over.
Third, mobile search depends on context. Now this is a term that has not become too overused to have its meaning worn off by familiarity. Context implies that a system can figure out the user’s situation, his device’s characteristics, and other factors bearing on the user at the moment information is needed. Context can cut through many of the problems identified by the USAToday’s Ms. Cauley, and it also is the key to making certain types of queries available in a computationally-inexpensive way. Let me give you an example. I had a heart “event”–my family doesn’t like the term “attack”. A mobile phone equipped with a geospatial function, a temperature and heart rate monitor, and the ability to connect to a network could perform the following automatic functions: [a] Know that I experienced a heart “event”, [b] link the heart rate and body temperature data to a look up table for things that could afflict me, [c] determine if I needed help, and [d] send a request for an emergency service to come to my aid. Sounds like science fiction, right? It’s just one example of how smart software, smartphones, and a robust mobile operating system can perform a search without typing, speaking, or pointing-and-clicking. This is one function disclosed in a Google patent application. I mention it to show that context can drive very different types of search experiences than those restricted to a user typing on BlackBerry’s key pad. Oh, Google calls this type of search without search, “I’m feeling doubly lucky.” (See US2006/0230350 “Non Standard Locality-Based Text Entry. You can download the document from the USPTO.)
Fourth, mobile search has to be broken down into types. For each type of mobile search there will be variants. Let me give you an example. Fire up a Treo 650. Launch Opera. You see a menu of pre defined bookmarks. Click on a bookmark. Is this mobile search or an ad? Now, click on the enter URL option and key in a mobile search site; for example, Yahoo. What do you see aside from annoying log in screens? You see pick lists. You have to do some serious poking around to get a search box. When you do, the results are tough to figure out. Even the results displayed from a pick list are tough to read and navigate. Mobile search is a mélange of different methods within a single system. I’m not faulting Yahoo. The same problems surface with most mobile services offering any type of information retrieval. Now, the real test. Try to locate an email on your mobile device. We are not just in early days with mobile search. We are at the gamete stage.
I have more “issues”, but this is a short essay in a free Web log. So. Enough already.
Observations
I had an opportunity to see some nifty mobile applications which will be available soon. I also received from a colleague a super secret proposal about a new mobile information service. One mobile search company is small. The other mobile company is a giant, dwarfing Googzilla in revenue, engineers, and employees. Both of these mobile search challenges’ forthcoming services are quite promising in my opinion. Both are variants of information access and are, therefore, embraced by the notion of mobile search. But I don’t think either solves the broader problem of search, device variance, context awareness, and ease of use. I anticipate more innovation in this space. I wrote about Judd Bowman’s new start up. Maybe he will hit another home run to match his Pinpoint mobile search system, now available as part of the Motricity suite.
Let me offer several other observations:
- We need a clear definition of mobile devices and a listing of the types of search each seems capable of supporting. Lacking that, we can’t discuss what’s available and what’s not.
- There will be considerable experimentation. In the end, mobile information access will follow other services down the path of utility monopolists of the past and present. Where are the many US automobile manufacturers of the not-so-distant past? There will be a high mortality rate among mobile search system providers.
- It’s too early to describe Google and Yahoo as vulnerable in mobile search. These outfits are leaders when you compare their offerings to what I have from my outstanding mobile provider.
Agree? Disagree? Use the comments section to share your views.
Stephen Arnold, June 11, 2008
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
