T-Mobile Says Federated Mobile Search Is THE Next Big Thing; Is It Also Google’s Achilles Heel? » mobilegroove

In-Brief: As I explained in this earlier post, I’m in Phuket speaking at the 3rd meeting of the Mobile Search Expert Group (MSEG). No live blogging due to NDA, but I do have the green light to print this in-depth and exclusive Q&A with Holger Knoepke, EVP, Product Design & Provisioning, T-Mobile International. He talks us through his pick of key prerequisites for a profitable mobile search service and reveals why T-Mobile replaced Google with Yahoo as their “partner of choice.”

As regular readers may recall, an ankle injury forced me to miss out on Mobile World Congress, but the buzz that T-Mobile’s session on mobile search created convinced me to reach out to T-Mobile directly for a deep-dive into the presentation and an exclusive Q&A with the presenter. Thanks again to T-Mobile’s Klaus Czerwinski for arranging both.

So what precisely is the business case for mobile search? Holger Knoepke tells me the answer is a mix of models.

It’s about generating direct revenues (by monetizing sponsored links via the search results pages); communicating the brand (strengthening the brand by providing a superior user experience); and creating indirect revenues. He splits the latter into three approaches: driving content revenues by recommending relevant content and providing easy access to it; driving mobile advertising revenues by making customers return to the web ‘n’ walk portal because it provides the best search experience; and driving mobile access revenues by underpinning the “free Internet” message of the web ‘n’ walk proposition.

This thinking isn’t new at T-mobile, whose web ‘n’ walk service gained it the reputation as a pioneer for the open Internet model. But I would say it is more pronounced, particularly now that the operator’s future strategy emphasizes federated search and aims to avoid the shortcomings of a universal approach.

Indeed, when it comes to mobile search and advertising, more is better, but useful is best. As Holger pointed out in our discussion: “The challenge is how to make the information, which you display as part of the search, extremely relevant.” The first release of web ‘n’ walk in 2005 was about offering users the opportunity to browse and search the entire Internet. “We saw the walled gardens and made a conscious decision to tear them down and we did. We wanted to provide the customer choice to search wherever they want to and this is basically what Google brought to the table.” In web ‘n’ walk 3.0, which T-Mobile launched in 4Q2007, the focus shifted to personalization and allowing the user to personalize their own web experience – down to a single sign-on.

“We needed to change because the mobile Internet changed. In view of this, we said we need to boost ease of use and provide users easy access to favorite services. So if you’re interested in e-Bay, you basically get a single sign on e-Bay access, so you’re automatically logged into e-Bay when you launch web ‘n’ walk and any other services you’re interested in….That’s what we thought was really key and you’re going to see more of that [from T-Mobile] going forward.

That thinking also laid the groundwork for T-Mobile’s partnership with Yahoo. “Mobile search is now part of the mobile Internet experience, different from the Internet experience at your PC. The mobile search is tailor-made for the mobile user who gets an even better output than on his PC.. The best partner is a company that sits with our brand and provides the best search experience. That’s why we went for Yahoo.

More about the Yahoo deal and Holger’s views on the success factors for mobile search services in this Q&A excerpt:

Q: At Barcelona, you provided a checklist of the five characteristics of successful mobile search: convenience, federated search, contextual relevance, personal relevance, and partnering. Why did you put federated so high up in the list? Some operators wouldn’t have listed it at all.

A: We think federated search is something really big – it’s not a small matter. It really takes search to the next level. Instead of only getting a list of links, and having to click on a link to see the result, federated search gives you the answer. This plays back to the convenience factor. Federated search also means that [results] won’t be limited to information that is published on the Web. Over time we will also make [as an inclusion in the results] available different content sources. It could be any database, a news source like a newspaper or The Economist, or whatever. The idea is to open this up to allow users to search on those databases that will deliver a better and richer experience back to the customer.

Q: You talk about opening this up to content sources and allowing their results to figure in your results. Without asking you to name the content companies on your radar, can you tell me your vision for this and the types of content sources you are considering?

A: Our vision is about offering what we call a third party enabling platform, which basically makes it possible for any content provider to plug into our platform. There probably will be differences [in how we do this] because it will likely depend on whether certain content providers want to have a commercial agreement. That’s of course different because they maybe would not do it for free but would ask for a return. In the case of [including] user-generated content, a lot of people are probably very happy if they just see their content on display as part of a search.

Q: Federated search theoretically paves the way to blend the search results of search engines as well. Are you thinking in this direction?

A: We don’t want to create a metasearch engine which includes Yahoo, Google and so on. Search here is really about Yahoo, and it will be branded with Yahoo, so that’s what we’re focusing on.

Q: Overall, T-Mobile is in a vocal minority talking about the benefits of federated mobile search. Do you consider yourself to be a pioneer?

A: Absolutely. As I said before, we really did a lot of customer insight to understand what the customer really wants to have as part of a mobile Internet. We found out people want to have ease of use and easy access to favorite services, and, in mobile search, you can substitute services for answers. Instead of just looking for links, people really want to have the answer. Instead of having a list of links, customers want seamless zero-click access to their favorite services. This is part of the reason why we want to have strong partners and we want to really personalize [the service] and make the search results relevant.

In mobile Internet services, search as a service is key and federated search is a perfect fit. But search is just a part. You want easy access to services such as email. That’s why, on the home screen of web ‘n’ walk, you can have zero sign-on access to Yahoo, to Hotmail, whatever you can think of.

Q: Not so long ago you could frighten operators with nightmare scenarios that users would change their starting page, and now you encourage this. What has changed?

A: We estimate 30 to 40 percent of our web ‘n’ walk users have actually customized their home page now, so the key is to get the best partners on the portal. And that’s easily done because, for the partner, it’s an interesting opportunity to get in front of that audience. It’s what I would call an additional distribution channel.

If we had just stayed with web ‘n’ walk, then users would have likely lost interest. But now we have enough customer insight to really know what the customer wants – and they want choice. We give it to them and we are very happy with the numbers we’re getting.

Q: Can you share your insights into search usage or behavior?

A: [Note: Holger did better than that; he supplied a list. I have no numbers, so I can't tell how popular each term is relative to the others.]

The winners are:

Q: Relevance – both personal and contextual – figures on your list of key success factors for mobile search. How does this fit together? Are you thinking that a federated search approach means more content sources, more inventory, and a better chance of delivering relevant advertising with search results?

A: Absolutely. I think we offer a superior product with federated search. More content and more results mean more eyeballs and, of course, more eyeballs generally mean more advertising and revenues.

I think the challenge here again is to be extremely relevant in the advertising you present as part of this. You can’t display a full list of advertising links, like what you see when you use Google, for example. It has to be reduced – and therefore it [advertising] has to be relevant. Otherwise, users will be annoyed. If I’m looking for a vacation, then showing me an ad for a cheap car doesn’t make sense.

Q: If I want to advertise on T-Mobile, how does that actually work now that you have partnered with Yahoo? You do have a subsidiary for advertising as well…

A: To be honest, we’ve not finally decided how this will evolve over time. We have a sister company where advertisers can buy media space and, of course, Yahoo has also brought its capabilities to the table. We have not yet decided how best to fully leverage that.

Q: When you talk personal and contextual relevance, what do you really mean? Is this relevant to search terms? Location? Behavior? All of the above?

A: Using just the search term for us is clearly not good enough. We know much more about the customer [because we are a mobile operator] and we are convinced that social demographics and location will also play a big role in this. We are currently developing those capabilities. We have the first capabilities in Germany and we are also trialing something in the U.K., and we will roll that one out over time and actually provide that information back to Yahoo. This will improve search results and relevancy.

Q: Let’s talk about your partnership with Yahoo. It’s about search and advertising. Is there a benefit to having the same partner for both?

A: Yes, [having more than one] basically defocuses everybody. This is still a non-mature market and I think Yahoo clearly knows whatever they invest, there will be a big pay-off and the same is true for us. We need to stay focused and that’s what we achieve in this partnership.

Q: It wasn’t long ago that the thinking was that a big search brand was necessary to get people to search in the first place. That was part of your own reason for tying up with Google, for example. What has happened to Google’s brand strength from an operator perspective and are operators perhaps less concerned about brand dilution overall?

A: First of all, Yahoo is also a strong brand, so that’s why we decided to partner with Yahoo. Second, we really like the product and everything else Yahoo brought to the table. We think federated search – and what we have seen [from Yahoo]- is probably the most advanced product. As we all know, it’s not everything about the brand. If you’ve got a superior product, then there’s a good chance that product will be successful.

Q: I hear that branded search providers also offer a very attractive revenue share agreement. Did that play a role in your decision?

A: Of course. But I can’t comment beyond that.

Q: What about the other services, such as social networking, that can intersect with search – or at least become part of the mix? Yahoo has a social network and search…

A: The fact that we partner with Yahoo on the search side and other services doesn’t prevent us from partnering with Google on any other services we might find interesting. We just like to keep them separate; otherwise you’re locking yourself in.

As you said, social networks play a very big role, and of course we are also talking with Yahoo about what could be done there. If you check out our website, we now offer YouTube, which is more like in the Google camp and which you could also say is social network.

Q: And finally, Android and the Open Handset Alliance, which you support. What impact will Android have? Will it reduce fragmentation and help bring the “two Internets” together?

A: Android has the capabilities to make the Internet accessible in a more convenient way because there are advanced browsing capabilities and some other things which make it easier to use Internet services. It won’t really change in itself anything on the Internet; it’s an operating system, so we don’t see a big influence on the Internet or on the mobile Internet.

But federation doesn’t only have a top-notch spot in T-Mobile strategy; it was THE main focus of MSEG today. I caught up with Marc Bookman — CEO of MCN (Mobile Content Networks), a company that has positioned itself as a federated search platform provider from the start –to get his take on the T-Mobile-Yahoo deal and T-Mobile’s new emphasis on federation.

In his view, federation represents the “next level” for leading operators and portals. [MCN built the first dynamic federated search platform (MobileSearch.net) and has launched over a dozen operators and portals covering 200 million mobile users since.] “Recently we’ve watched with interest as various vendors have claimed to have a ‘federated’ solution and are gratified that T-Mobile and other leading operators are beginning to take an interest in federated approaches.”

Q: You don’t think those claims to be “federated” are accurate?

A: Not quite. In every case we’ve seen thus far, with the exception of MCN deployments, the vendors are providing either [what is known as]enhanced indexing; which basically replaces crawler feed processors, or static federation, which federates to a limited set of Internet search engines such as Google and Yahoo.

The enhanced indexing approach is not real-time because the content is still collected and managed in a central index which adds cost and reduces relevance by not reflecting the variety of metadata appropriate to each content type. The static federation approach doesn’t provide sufficient answer-level relevance nor does it reduce click distances enough for mobile users. In contrast, the dynamic federation [which MCN provides] analyzes the query to determine the best set of local content providers, queries them in real time, and returns the freshest, most local answers as the response.

Q: Federation means more results and more choice. Can it do the same for advertising, creating a breadth that allows the delivery of more relevant and targeted ads than is currently the case?

A: No single ad network can cover enough of the mobile keywords with contextually relevant ads. Therefore it makes sense to federate to multiple ad networks to enable the most relevant advertising results associated with user queries. With the profusion of ad networks, whether from the Web giants or mobile ad startups, ad networks are just another content provider for mobile operators and portals. We’ve tested and proven this in Japan with our partner, D2 Communications., and we expect to see an acceleration of federated advertising activity this year.

This entry was posted on Thursday, March 13th, 2008 at 3:20 pm and is filed under Location-Based Services, Mobile Marketing, Mobile Search, Mobile Social Media. 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.