Is mobile search broken? Let’s just say the experience leaves a lot to be desired.

A major obstacle is usability. Inputting search terms on a mobile device is tedious, but the more words in the query the more accurate the result set is likely to be. You’re sort of stuck between a rock and a hard place, which is why query completion and/or multi-modal technology (voice) have a top-notch spot in many mobile search strategies.

But input is just part of the equation. Another important issue is context/personalization, and the requirement for mobile search services to “learn” with the user to understand that when they type in “bond” they want financial news (stocks and bonds) and not entertainment (James Bond).

Many tools and technologies are coming online to tackle these issues, but kannuu, a newcomer that broke on the scene with its patented technology in 2007, has the capabilities mix to address several key issues at once. In fact, I have had kannuu high on my radar since the start of the year when I was requested to judge the 2008 Meffy Mobile Entertainment awards in the mobile Search and Discovery category for the second consecutive year. (BTW: The awards provide an excellent snapshot of cutting-edge technology and business models, and are presented by the Mobile Entertainment Forum, the global infotainment industry association. In the end, kannuu made the shortlist, along with some other companies you can read about here.)

I was initially impressed by the company’s approach to usability. But a recent briefing with Sean-Michael Daley, kannuu CEO, has made it clear that kannuu has its eye on a much bigger prize: Applying its self-learning algorithms and key IP in indexing to improve mobile search and enable new monetization schemes that go far beyond Google AdWords. (Thanks again to Sarah Duckett, kannuu VP Communications, for arranging the briefing and providing me with first-rate background materials such as the video below.)

Sean-Michael tells me that recent usability tests estimate kannuu’s technology “generally delivers a 3x improvement in the speed at which users can find something.” But it’s not just about speeding up mobile search. The pay-off is in fine-tuning mobile search to enable transactions and commerce on the fly. (Sean-Michael’s remarks remind me of a recent panel on mobile search trends in Japan that MSG covered here. The bottom line: Commerce, not search advertising, will be the way content owners and brands make their money.)

Sean-Michael tells me that mobile commerce pilots are already underway. “That’s where there is a really good return on investment…We can improve the quality of the experience people go through and [enable them to] actually complete transactions. Because of the volume there you only need to increase that by a fraction of a percent and our technology is well and truly paying for itself.”

Listen to the podcast here. [11:30]

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

In my view, kannuu is one of the few companies in a position to both recognize and address the growth opportunities in mobile search and content discovery. That’s a pretty bold statement, so I’ll begin with a review of the key capabilities that allow kannuu to organize content sources so items are searchable and findable in the first place (thus going one better than most).

As the demo above shows, kannuu’s user interface is built from the ground up to present the user with relevant options at each menu level, along with a way for the user to make their next move. In the case of a device with a screen and four-way directional button, the user gets four choices at a time, and can select between them by pressing the button up, down, left, or right. The system also provides an option for the user to request more choices if the desired option is not displayed.

The interface is cool, but I’m even more excited about the algorithms and the indexing capabilities under the hood that work with the interface component to ensure the user gets shortcuts to content that is indeed available. It’s not about telling the user what could be there; it’s about showing the user what really is there.

In a nutshell, kannuu only presents data that exists in the database – that is, data that belongs to the partner mobile search provider or content partner – so users never receive the dreaded “no results found” message. As Sean-Michael explains: “We index your content [the content belonging to kannuu's partner] and enable the lookup on that content only. … We present the user with the most likely bit sized chunks of characters required to complete the desired lookup phrase.”

In addition to indexing content items in the database owned by the customer company, kannuu can also index content stored on the individual user’s device, making it searchable, findable and monetizable.

“Just say there may be someone in your address book whose name starts with a “Z” who you call often…If we’re looking at it [your behavior]  we see you call this person often. We can say the algorithm should be influenced by frequency and, since you call that number on a regular basis, we can promote that number higher up the list of choices. So you’ll actually get to that choice much earlier.”

So how is this possible?

Kannuu technology begins by analyzing source content and organizing it into a hierarchical tree of string sections, a structure we’ll refer to as the kannuu index. The kannuu index is the basis for kannuu lookup. It is also the mechanism that determines what to show to the individual user. At a deeper level, it’s the interplay between the index and kannuu’s own algorithms that makes doubly sure each individual user is shown the results in the order they will appreciate.

Once the kannuu index tree is generated, lookup is all about presenting the user with the most likely sets of characters needed to build their desired phrase. The feedback loop between the technology and the user – as options are selected or rejected by the user – give the kannuu algorithm the clues it needs to analyze the kannuu index and bubble up the right answers faster.

Kannuu’s technology can also be fine-tuned to learn over time which content is the most useful to the individual user and place those results at the top of the list. To this end kannuu can arrange results according to the user’s preferences or shuffle results to match the business rules of a business partner. “So if the advertiser is promoting a particular [item], the algorithm can be influenced so those particular choices are pushed further up the list and therefore appear earlier in the user’s selection process.”

BUSINESS MODEL: “We will make money two ways. If we’re actually on the device itself, [then] we sit on the actual device itself and therefore the manufacturer pays us a royalty per device that we’re on. The other way is that we sit on the server side. So if you’re on say a mobile phone and you’re on a Web connection and you’re looking at, say, Amazon, it would be Amazon that would be paying for our server license to use our technology.”

KILLER OPP(ORTUNITY): “The opportunity to be able to make more sales is there if people can easily and conveniently do a lookup on a mobile device. So, to me, it’s about making it easy. If you make it easy for the user, they’ll do it more often, they’ll complete more transactions, therefore we’re making more revenue. It’s a pretty straightforward story on that front.”

TRACTION: Kannuu is getting it. A major deal for the company is an alliance with Coby Electronics that will bundle kannuu’s proprietary technology with MP3 players beginning this quarter. The tie-up allows users immediate access to on-device content. In the mobile device space, chances are we’ll see more of kannuu on Nokia devices. In March, kannuu won the Technology Innovation Award for Best Mobile Multimedia experience at Nokia’s annual Mobile Rules! competition and has since been in discussions about ways to improve on-device search. “We’re moving down a path to getting [kannuu] into a pilot and that will be on their [Nokia] high-end phones.” (It’s easy to imagine taking it a step further to make the content part of the equation (think Ovi) indexable and searchable.) Sean-Michael tells me kannuu lookup functions can also be embedded into games consoles, media players, set-top boxes, kiosks and GPS-enabled devices. “Kannuu can work on any device that has a screen, a directional control of touchscreen, and the ability to index data from a database.”

The company shines on set-top boxes. “We’re having some really interesting conversations right now with a couple of the major service providers in that area [set-top boxes]. There is no doubt when you’re on a remote control, being able to choose up, down, left, right, center is the natural with your thumb, you don’t need to look at the remote to do that, you just need to look at the screen. I think also you’ve got much more screen real estate than you have on a mobile device, so we can do more things with our technology to make the experience even more fulfilling for the user.”

BTW: Kannuu can also be incorporated into widgets including Nokia WidSets, Yahoo Go widgets, and Opera widgets.

APPLE: Kannuu’s search SDK for iPhone application developers is now available for download. (The free beta iPhone SDK can be downloaded here.) Sean-Michael tells me the technology, which is also now available on the Apple Apps Store, has been well-received by Apple executives. “The popularity of the platform, along with the high demand for applications, gives us an amazing opportunity to connect consumers with content.”

SCALE: “Yahoo and Google can’t take their current search metaphor that they use on the desktop and just resize that to the mobile environment, they’ve got to really re-think it. That’s part of what we are, we’ve re-thought what search is and what look-up is when you’re on a device that’s got limited entry and of course the big databases, that’s where we really come into our own because [of] the way our technology works.”

Kannuu employs an exponential selection algorithm, rather than a linear selection algorithm. This means kannuu can work with source lists of any size, with only fractional increases in the number of clicks required to find content as the list exponentially grows. As Sean-Michael puts it: “As the list of possible results grows, kannuu becomes more efficient.”

“So, as you go from say 1,000 items to 1,000,000 items, on average you’re only adding an extra three clicks into your search and I think everyone realizes that the trend is more data, not less.” A word about speed. Sean-Micheal tells me it “takes seconds” to process an index the size of Yahoo – and kannuu should know because they work with Yahoo often.

CONTENT PROVIDERS: It’s about allowing content providers to provide a better user experience now rather than wait for handset makers to meet the usability problems heads-on. “We allow the content providers to take control of the full user experience rather than rely on the hardware to be the way the text is entered and the search is experienced. In a way it standardizes it – allowing content providers to deliver a consistent user experience regardless of the hardware.” BTW: The kannuu user interface is fully customizable, enabling a consistent brand experience for content providers.

CONTEXT & LOCATION: The kannuu algorithm can take in and process data around the user. “It may be previous searches; it may be location; and [whatever the input] they all have an effect on the [search] results that are presented….This could also be influenced by other searches taking place on the network or by mobile advertising.”

WHAT’S NEXT: Pilots – and lots of them. On mobile devices and set-top boxes, as well as “some interesting experiments in new markets.”

MY TAKE: The demo is impressive and so is kannuu’s three-prong approach to tackling the issues that make mobile search a chore. Wrapping an indexing method and a partial word completion algorithm together with a super-cool user interface, kannuu takes the heavy-lifting out of mobile search and content discovery, requiring fewer clicks to reach relevant results. If mobile search does indeed move across platforms and devices – as we would expect when convergence kicks in – then kannuu is a company that can deliver it.

Tags: Kannuu, podcast

This entry was posted on Monday, August 11th, 2008 at 12:43 pm and is filed under Content Discovery, Local Search, Mobile Advertising, Mobile Search, Multimodal Search, Personalization, Podcasts, Usability. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.