Author Archives: Nadir Garouche

Mobile Search Engine Taptu Wins $9.6M in New Funding

Taptu, a pure mobile search engine that I reviewed a while ago on this post just announced that it has won $9.6 million in extra funding from its existing investors.

This news confirms that there is still room for growth in the mobile search market. Taptu is specialized in mobile content search, as opposed to traditional search engines that still mainly focus on web pages. Taptu’s main competitor is ABPhone, a French mobile search engine, which is also seeing an exponential growth in the number of searches made.

Handheld devices will be the primary connection tool to the Internet for most people in the world in 2020, so there is a need for mobile vertical search, and players such as Taptu are already filling that need.

Via Mobile Marketer.

Google AdWords Allowing Targeting of iPhones, G1s and similar phones

If you’re a mobile content provider, then you’ll enjoy the new feature available in Google AdWords.

You can now display ads exclusively for the iPhone and Google’s G1. For example, if you are providing games for the iPhone, you can run AdWords Campaigns to only target these users, so you’ll get more targeted users when bidding for “games”, as opposed to buying the keyword “games” for all users (PC users, non-iPhone users, etc)

To do so, you need to do the following in your AdWords account:

To target ads for G1 and iPhone, go to your campaign settings tab in your AdWords account. Then for the “Device Platform” option under “Networks and Bidding,” select “iPhones and other mobile devices with full internet browsers.”

To only target iPhone, G1, and other phones with full HTML browsers, unselect “Destop and laptop computers”. This way only these kind of phones will see your ads.

Also note that if you’re currently running “regular” AdWords campaign, they will be displayed on iPhones, G1s, and other mobile phones with full internet browsers. You’ll need to opt-out in your AdWords account if you do not want to have your ads displayed on these devices.

I wish Google allowed to target even more, for example, only iPhones, only G1’s, etc. This could be very handy for content providers.

Also, I hope that this kind of customization will be applied to organic search, for example giving more weight to sites designed for the iPhone when people are searching from an iPhone…

Source : Google Mobile Blog.

Confirmed: Yahoo! OneSearch Comes to T-Mobile USA

I blogged last month about the mobile search deal that was being tied up between T-Mobile USA and Yahoo!. It has officially been announced today.

T-Mobile USA has launched a new mobile portal called web2go, that has a Yahoo! OneSearch searchbox right at the top of the page, according to a screenshot from Venture Beat.

In addition to that, some cell phones that T-Mobile will launch from now will have a shortcut button to Yahoo, except of course Google’s G1.

Users will then be more tempted to visit off-deck sites and mobile site (or regular site) owners should see an increase in their traffic from Yahoo! OneSearch if they’re properly indexed there.

Google will still remain a partner according to T-Mobile.

I wondered if Medio would remain present on T-Mobile’s deck and it seems to be the case, Medio will still provide mobile content search results (games, news, ringtones, etc). Yahoo OneSearch is capable of indexing and ranking mobile content but I guess Medio still has the contract for that type of service.

Mobile Search Patent: How Google Would Blend Mobile Search Results

Back at the end of July 2008, a patent from Google named “Blending Mobile Search Results” has been published. It’s only until recently that I decided to read it.

You know that the new version of Google Mobile Search now by default presents a mix of generic pages and mobile pages. In this patent, Google explains how it would blend mobile search results, that is, increasing the quality score of a mobile page so that it is displayed higher in the search results.

BLENDING MOBILE SEARCH RESULTS
United States Patent Application: 0080183699
Inventors: Hu; Ning; (Pittsburgh, PA) ; Ha; Vida U.; (New York, NY)
Assigned to Google
Filed: January 24, 2008
Published: July 31, 2008

Abstract

Methods, systems, and apparatus, including computer program products, for blending mobile search results. A method includes receiving a search query and multiple search results. The search results each satisfy the search query and have a respective search result quality score. The search results include generic and mobile search results. The generic and mobile search results each identify a generic and mobile resource, respectively. The search result quality scores include mobile and generic search result quality scores for the mobile and generic search results, respectively. The mobile search result quality scores and the generic search result quality scores were generated according to different scoring formulas. Based on one or more terms in the search query, the search query is classified as a mobile query. As a consequence, one or more search result quality scores are modified to improve the sorting of search results that include both mobile and generic search results.

Here is how Google would proceed, I have outlined some of the most important points:

9. The method of claim 8, wherein modifying the mobile search result quality score comprises:increasing the mobile search result quality score if the mobile search result links to a mobile resource that links to downloadable content for a mobile device.

10. The method of claim 8, wherein modifying the mobile search result quality score comprises:identifying a first language of the search query;identifying that the mobile resource identified by the mobile search
result is written in a second language; anddecreasing the mobile search result quality score if the first
language is different from the second language.

11. The method of claim 8, further comprising:ranking the mobile search results and the generic search results in an order, the ranking being based on the search result quality scores; and removing one or more duplicates from the order.

29. The system of claim 28, wherein modifying the mobile search result quality score comprises:increasing the mobile search result quality score if the mobile search result links to a mobile resource that links to downloadable content for a mobile device.

So, if you have a mobile site that provides downloadable content for mobile devices, such as mobile games, ringtones, or all kinds of mobile applications, Google would improve your quality score. I wonder how Google would proceed exactly though, because on paid sites, you need to go to the billing page before being able to download a .jar file or any other type of mobile content. Or maybe if it will see if it finds words such as “download” throughout your pages…

Google also explains how it will rank a mobile page higher if it had a generic (desktop computer) version that linked to that mobile page, by actually removing the generic page:

32. The system of claim 31, wherein removing the one or more duplicates comprises:identifying a first
mobile search result that identifies a first uniform resource locator;identifying a first generic search
result that identifies a second uniform resource locator, the second uniform resource locator being the
same as the first uniform resource locator;removing the first generic search result from the order; and if the first generic search result had a higher rank than the first mobile search result, moving the first mobile search result to the position in the order that the first generic search result occupied.

I invite you to have a look at the images that come with the patent.

In the image annotations, there are also some interesting tidbits. For example, Google explains that it uses different algorithms for mobile search results.

[0034]The mobile search engine calculates a search result quality score for each of its search results (step 240). To calculate the scores, the mobile search engine uses a different scoring algorithm, or formula, than the one used by the generic search engine. Using the search result quality scores, the mobile search engine ranks the mobile search results (step 245). In other implementations, the results mixer ranks the mobile search results.

Sometimes, if the number of mobile search results for a query is too small, Google won’t display mobile pages at all, since having too few mobile sites pages will decrease the relevancy of the returned results.

[0036]As illustrated in FIG. 3, in some implementations, the results mixer first determines whether to present any mobile search results. It may do this, for example, by determining whether the number of mobile search results produced by the search query is greater than a threshold number (step 302). If the number of mobile search results is too small, then the results mixer concludes that none of the results are sufficiently relevant to the user’s search query. When this approach is used, the results mixer will only present the mobile search results when the number of mobile search results is large enough (step 304).

[0037]The threshold number may be either an absolute number or a percentage. For example, the percentage could represent the percent of mobile search results received by the results mixer out of the combined number of mobile and generic search results received.

And now comes the most interesting part of the patent. Google claims that it can increase the quality score of the mobile search results if it’s clearly a “mobile” query, that is, users willing to find content or information that corresponds to a need while on the go.

If a query contains words that Google believes are “mobile”, for example: ringtones, games, wallpapers, or even chat, news, etc, it will make sure mobile pages are ranked first:

[0038]The results mixer determines whether to modify the mobile search result quality scores (step 250).Depending on its determination, the results mixer will or will not modify the mobile search result quality scores (step 255). Modifying the scores may be necessary to make the mobile search result quality scores commensurable to the generic search result quality scores, or otherwise to improve the usefulness of the presentation of blended search results.

[0039]The determination can be made using one or more signals, including whether the query is a “mobile” query, whether the search results exceed threshold scores, the number of mobile search results, the properties of the mobile resources found in the search, or the properties of the generic resources found in the search. In some implementations, when the results mixer uses more than one signal, it will use the signals and modify the mobile search result quality scores in a non-linear manner.

[0040]In some implementations, based on one or more terms in the search query, the results mixer classifies the search query as “mobile,” i.e., whether it is likely to be intended as a search for mobile resources,e.g., pages (step 306). If so, the results mixer will increase the search result quality scores of the mobile search results (step 308). For example, if a search query includes the term “ringtones”, the results mixer can determine that such a query is likely to be intended as a search for mobile resources, because the word “ringtones” is closely associated with mobile devices. In such a case, the results mixer will increase the mobile search result quality scores. It may do this, for example, by multiplying all the mobile search result quality scores by a number greater than one. In addition or as an alternative, in some implementations, if the search query is determined likely not intended as a search for mobile resources, the results mixer will decrease the mobile search result quality scores.

Google also affirm that it may blend search results with results from other mobile search engines. For example, including search results from a mobile content search engine.

[0071]Also, generic search results may be blended with search results from other search engines instead of or in addition to being blended with mobile search results. Examples include search results from a local listings search engine for searching local listings, an image search engine for searching images, a carrier private content search engine for searching mobile web pages categorized as accessible only to subscribers of certain mobile service plans, a mobile news search engine for searching mobile web pages categorized as news, or a mobile marketplace search engine for searching mobile content to purchase.

Mobile Advertising for Beginners

Peggy Anne Salz, publisher of MSearchGroove has been writing a whitepaper about Mobile Advertising, which is now available for download.

This paper aims to help companies understand how to run mobile advertising campaigns, Peggy provides case studies of 3 campaigns she’s made: Google Mobile Adwords, Decktrade, and Mo’Jiva.

In this paper, you’ll learn how to launch a campaign, optimize it, and how to track your results. Peggy used two mobile analytics solutions: Bango Analytics and AdMob Analytics. She compares the two solutions and outlines the features of each.

The guide has been written for people new to mobile advertising but can also be interesting for people already running mobile ad campaigns.

To download the whitepaper, head over to Bango.

Yahoo! to Become T-Mobile USA’s default Mobile Search Engine

Google may have teamed up with T-Mobile to launch the G1, but I heard from someone close to the matter that Yahoo! will soon be the default mobile search engine on T-Mobile USA’s mobile web portal. All phones launching from November 8th will have a shortcut button to Yahoo! OneSearch, except obviously Google G1 and the new Sidekick.

Yahoo! OneSearch will allow T-Mobile USA customers to search for on-portal content such as games, ringtones, news, etc but also off-portal content.

Medio is currently T-Mobile USA’s mobile search provider but I don’t know if it will still remain present on its deck, I should have more details on that soon.

Yahoo! was already T-Mobile partner for mobile search in Europe, with T-Mobile USA it will have the potential to directly reach 30 million subscribers.