MOBILE SEARCH ENGINE SURVEY
Lauri Kolmonen Helsinki University of Technology, Telecommunications Software and Multimedia Laboratory lauri.kolmonen@hut.fi
Abstract
The immense amount of information available on the Internet makes it virtually impossible to find relevant information just by following links from page to another. Because of this, the search engines have become commonplace entry points to Internet in desktop use, shaping the way we browse the Internet by the result they provide. Advertisement based revenue models, that is, displaying advertisements related to search results, have proven to be a successful business. The rapid growth in number of mobile Internet users creates new business opportunities in the communication industry. The success of search engines in desktop computing lures many players also to seek profits from the mobile search engine business. However, there are challenges that hinder deploying a successful mobile search engine business, such as the immaturity of the current mobile Internet. In this paper, we discuss mobile search and related business models and opportunities from multiple different user and business perspectives. We also take a brief look in technologies which form the basis of search engine services today and the different challenges that the services face in their way to maturity. The paper also discusses locationawareness and how it can be exploited in mobile search engine business.
more than three billion people in total have a mobile subscription. (ITU 2007) Increasingly many of these subscribers also use the device for browsing the Internet. According to a study, in year 2007 globally 28 percent of mobile subscribes have used their device for accessing the Internet. (Church, Smyth, Cotter and Bradley 2007) The high number of potential users creates opportunities for mobile Internet service business. There are some enablers, such as flat-rate tariff pricing, which help to promote mobile Internet use. However, there are also limitations, for example in usability of mobile devices, network reliability and bandwidth, which set barriers for the adaptation of the mobile Internet. These barriers also hinder entire mobile Internet based service business, including mobile search engines. A sufficiently large user base opens up possibilities to provide services based on advertisement based revenue models. Success stories, such as Google, show that these advertisement based revenue model can also be highly profitable. The success of search engines in desktop computing attracts companies also to seek profits from the mobile search engine business. Many companies have already entered mobile search engine market with services especially designed for mobile devices. Most important players include large companies that also have a strong foothold on desktop search, such as Google, Yahoo!, Microsoft and AOL.
Keywords
Mobile search engines, advertising, location awareness, operator business
2. Overview of Mobile Search Engines
Mobile search engine is in a piece of software designed for a mobile device to provide a service, or a portal, through which the user may submit a query (usually by entering a set of keywords) and get list of results matching the search criteria. Mobile search engines do not differ a lot from regular desktop counterparts. In fact, the key components behind the engine, crawler, parser and indexed usually remain the same, only the user interface has to be designed to match the limitations of mobile devices. (Paulson 2005) Please refer for example to articles Web Search Engine, Part 1&2 (Hawking 2006) for more specific details about the internals of a typical search engine. Some of the mobile search engines prioritize and filter the search results in order to highlight more relevant result for mobile devices. For example, the mobile version of a search engine could be configured only to search pages designed for mobile devices (e.g. by
1. Introduction
Because of the immense amount of data available in the Internet, offering search services has become a big business in the Internet. The actual amount of data available in the Internet is very difficult to estimate, but some numbers are available. For example, The March 2008 Web Server Survey by Netcraft (Netcraft 2008) estimates that there are currently 164 million registered hostnames on the Internet with the growth of millions of hosts by each month. Finding relevant information from a network of this magnitude is practically impossible without appropriate tools. The statistics by International Telecommunication Union suggest high growth in the number of mobile subscribers worldwide. The mobile penetration rate is estimated to be more than 50 percent globally, that is,
looking at the content type of the pages). (Paulson 2005)
2.1. Usability Issues
The limitations of current mobile devices, such as small screen resolution and cumbersome character input, create challenges for designing mobile search engines. However, there are studies describing solutions for improving search engine usability on mobile devices, for example by auto-completion, related queries (Church, Keane and Smyth 2005) and text summarization. (Sonera Media Lab 2002) Other solutions include for example speech recognition: In 2007 Microsoft acquired, Tellme networks in order to take advantage of voice-based user interface to improve the user experience. (The Associated Press. 2007) The Yahoo! is also currently exploiting speech recognition in their mobile search. Limited network bandwidth and latency have an impact to usability of mobile Internet as well. However, the data rates have been steadily increasing in the past years. Other problems include limited support of protocols and file formats on mobile devices. A study (Church, Smyth, Cotter and Bradley 2007) shows that the users who use mobile search engines tend to use more sophisticated devices compared to browsing users − probably due to a better usability.
one session compared to desktop searching, most likely due to much more troublesome input methods. The amount of required input effort to enter a query with a cell phone keypad is more than double compared to full QWERTY keyboard. However, the average query lengths seem to be of a very same length as desktop queries.
2.3. Related Technologies and Standards
There are some related standard technologies, which for the most part are related to presentation format of the search results and the data transfer medium. In this chapter we are going to discuss only the relevant current standards in brief. We are not going to cover e.g. WAP because of its irrelevance with modern solutions. As most of the mobile search engines are common web applications, they are built using related world wide web standards, such as HTML (HyperText Markup Language), XHTML (eXtensible HTML) or some of their mobile versions, which are more suitable for mobile user interfaces, such as XHTML Mobile Profile, Palm HTML or Compact HTML. (Sonera Media Lab 2002) Some of the current mobile devices support the same technologies as desktop computes. That is, they are equipped with browser capable of rendering XHTML (The Extensible HyperText Markup Language) with JavaScript scripting language and even handling proprietary de-facto standard software, such as Adobe Flash. Devices with higher resolution screens and touch screen input, such as Nokia N810 and Apple iPhone, respectively, bring also the usability of mobile devices closer to desktop PCs. Because of the support for common technologies and better usability, many of these mobile devices can now access the same web pages and services as a regular desktop computer. This reduces development costs of mobile search engines as there is not more need to develop a separate interface for these handheld devices. With the evolution of future mobile devices, it is likely that the services they support will move closer to them supported by standard desktop computes. Some of the services are built on SMS (Short message service). SMS can either be used as data transfer medium behind a separate mobile search application installed to device or as an actual interface to search service. SMS is an attractive technology for because it is open standard with wide support over multiple carries. (Paulson 2005) However, using SMS as a user interface to search has some limitations, such as complexity of use (the user has to know the number and message format required by the search service) and the limited message length. (Schusteritsch et. al. 2005)
2.2. Mobile Search Behavior
The mobile search, as well as whole mobile Internet, seems to be in early stage of maturity. A study (Church, Smyth, Cotter and Bradley 2007) reveals that only 6 percent of mobile users actively use mobile search (in year 2007). The number is relatively low, compared to browsing-only users, which cover 94 percent of the sessions. On the other hand, the study also shows that mobile Internet sessions involving search tend to be longer and include more data than browsing-only sessions. Studies (Kamvar and Baluja 2006 and 2007) reveal the most popular search categories for mobile search engines: adult, entertainment, internet & telecom (such as site specific searches, e.g. “hotmail”) and local services, respectively. The relatively high number of adult queries seems support the fact that the mobile search follows the trend of set by the desktop search queries. Other studies (e.g. Church, Smyth, Cotter and Bradley 2007) as well have found the close resemblance of the evolution of mobile search compared to early desktop search and suggest the early stage of maturity of these services. These observations support the hypothesis that in order for the mobile search engines to become successful business, the mobile devices and the mobile Internet usability as a whole have to develop further. Studies (Kamvar and Baluja 2006 and 2007) also show the more homogenous and significantly fewer queries in
3. Mobile Search Engine Business
The industry has proven the success of search engine business in desktop computing, but the low number of
usage reveals that the real success is yet to take off in mobile devices. However, we believe that the whole mobile Internet is still in its infancy and the mobile search engine business is likely to develop with the growth of mobile Internet. Studies (e.g. Church, Smyth, Cotter and Bradley 2007) also suggest that the mobile search engine business can be considered to be in a development stage of the life-cycle model, with only the early adopters using the services. Despite this, mobility provides new opportunities for businesses, such as larger customer base unbound to a specific location.
3.2. End-user Service Types
Most of the search services are targeted to very wide market segment. The same search engines serve all types of customers, as their only purpose is to find information from a large set of data. However, for example SMS based services may seek differentiation by focusing to a specific segment, in order to serve the needs of niche groups. The most common mobile search service type is a separate application or a browser-based web site that can be used to seek pages from the Internet. Because of the customers’ unwillingness to pay for mobile search, the service is typically free of charge, ignoring the data transfer costs. Typically these services are not targeted to a specific market segment. In addition to application and browser-based services, there are SMS based mobile search services as well. A user can request information by sending a SMS message to a service number and getting requested data as a response message. (Schusteritsch et. al. 2005) These services can include for example a phonebook services or querying public tax information. Typically these services are offered as a premium service described in section 3.1. Another option is to offer these services as a sponsored service, that is, with some contextual advertising. However, there are some regulatory limitations, as well as broad negative consumer attitude towards SMS based advertising. (Immonen et al. 2006) There are also other types of search services in addition to text based services. For example, Google has integrated mobile search, location awareness and maps software as a single service. After the location-aware Symbian application has been installed on a mobile device, the software allows user to search e.g. local businesses by providing set of keywords. The search results are listed on a map which shows the location of the matching businesses.
3.1. Revenue Models and pricing
In this section, we discuss the most common mobile search revenue models for vendors. In addition to vendors, there are other players as well. How these actors in value network can seek revenues from mobile search business will be discussed in more detail later in this paper. A research has shown that the users are not willing to pay more for mobile communication (Immonen et al. 2006) compared to services available for desktop use. This must be taken into account when thinking about the pricing and seeking the potential revenue models for mobile search engine business. Typically the vendors gain their revenues from selling advertisements with query results. Usually these advertisements are context-aware, that is, related to search query keywords. For example, search query “web site” returns advertisements for businesses which offer web site development. Another revenue model is directly providing premium services subject to a fee directly to customers. However, with this model it is required that the vendor has ownership of a hard to access data in order to justify the fee of the service. 3.1.1. Location-awareness When searching for e.g. services, mobile users are likely to be interested in finding businesses near their current location. Location-awareness enables users to seek information about local restaurants or entertainment services without explicitly entering the current location. By exploiting location-awareness technology, such as GPS, mobile search engine vendors are able to offer geographically relevant search results for users, thus increasing also the end-user value. At the same time search vendors can advertise local businesses by adding location to context-aware search results, therefore also increasing the customer value (and revenues). For example, when searching for “pizza”, the user gets advertisements for businesses which serve pizza in near area.
4. Mobility and Search Vendors
In this section, we are going to discuss the business aspects of mobile search engines for search engine vendors. These aspects include vendor strategies, analysis of competition and substitutes and how to use standardization to gain competitive advantage.
4.1. Vendor Strategies
A common search vendor strategy for gaining market share is ally with a mobile device vendor or an operator and to bundle company’s search service with a phone in order to make sure that the service is configured as the default search engine of the device. The same strategy has been implemented with desktop software, for example Google has embedded its search engine as the default home page of a popular web browser Mozilla Firefox and Microsoft with Internet Explorer. There are already numerous ongoing alliances with search
vendors, mobile operators and device manufactures, e.g. Nokia teaming up with Google and T-Mobile with Yahoo!. (Gohring, 2008) (Schonfeld, 2008) Google has also sought to gain competitive advantage by pushing open mobile platform developed by the company, namely Android. The company has joined some 30 companies in order to form an alliance and to replace Symbian operating system as the dominant mobile platform. This creates new possibilities for mobile application development using company’s services, including their search, thus increasing their revenue. (Schonfeld, 2008) Smaller search vendors can try to concentrate to a niche market segments or try to sell or license their technology to other companies seeking to gain competitive advantage by innovative solutions. One such example was already mentioned in section 2.1. Usability Issues, where company Tellme Network was purchased by Microsoft.
population finds the mobile Internet access important, the need in Latin America is almost three fold (63.5 percent). (Burns 2007) By pushing mobile search in these countries, the operators could promote the usefulness of mobile Internet and seek to grow their customer base and revenues. As mentioned in chapter 2.2. (Mobile Search Behavior) the mobile Internet sessions involving search are on average longer and more data rich than browsing-only sessions. Thus mobile search engines could provide means for operators to increase revenues from the use of mobile Internet. By replacing operator-owned mobile portals with a mobile search engine, operators can try to promote searching instead of browsing, thus enabling higher data rates and increased revenues from subscriptions with data transfer based billing. However, if the operator sells flat-rate subscriptions, the mobile search engines may have a negative affect since the relative euro per bit is lower compared to browsing-only sessions. Operators can also try to exploit mobile search engine business by directly entering search business by capitalizing on their established customer base and marketing power. For example in Finland, a company, namely Sonera, has done just this by providing SMS based search and other information services, such as phone book and billing information to customers. The business was sold to 3i in the year 2002 for about 120 million euros. (Business Wire 2002) One opportunity is also to act as location data provider for search engine businesses. (Docter et. al. 2007) However, local regulation and the development of location-aware mobile devices with built-in GPS undermine the business opportunities based on providing location data.
4.2. Competition and Substitutes
Competing with large search engine companies requires a major investment, in order to be able to provide a highly scalable service able serve millions of users daily. This includes a large IT infrastructure of servers for quick response times and to support data mining robots that crawl through the Internet to build search indexes. Thus, the barrier of entry is rather high. On the other hand, Internet provides success stories, where small companies have been able to grow and gain market share with their innovative service. Threat of service substitution can be regarded as rather low as the industry has already proven the high demand for search demand services. The incomprehensible amount of information online requires some kind of a search service. So, for the end-user, the mobile search engine technology is there to meet the need of quickly finding the relevant data online. This need is difficult to substitute. However, there are some suggestions for improving the overall internet architecture by embedding semantic information to web pages. This semantic data could be used to improve the relevance of the search results. (Li, Wang and Huang 2007) In order to maintain their competitiveness, the search vendors must keep in touch with the new technologies which could threaten their current position.
6. Conclusions
Industry success stories have shown the attractiveness of search engine business. As Internet use begins to expand more and more also to mobile devices, it is evident that mobile search business has great potential. Advertisement based revenue models combined with location awareness offer interesting service types for used and a real business opportunity for search engine vendors. However, the early maturity of mobile Internet and the usability limitations of mobile devices set barrier for successful service deployment. We believe that as the devices continue to develop further the mobile search can truly show their business potential. Because of this, many of the search providers have already started to build their mobile user base, in order to ensure the competitiveness as the mobile search engine business unfolds.
5. Influence to operator business models
Mobile search engines can be used as a means to promote the mobile Internet use in addition to basic phone calls and SMS. A study shows a high demand for mobile services in developing countries where people lack the disposable income to purchase a desktop computer. The users are likely to be able only to invest in mobile device, which therefore becomes the main communication device of a household (Paulson 2005). For example, compared to U.S, where 22.6 percent of
References
Associated Press. 2007. Microsoft to acquire Tellme Networks. [Online] http://www.msnbc.msn.com/ id/17611739/ (Referenced April 08, 2008) Burns. 2007. Mobile Content Usage is Higher in Developing Countries. The ClickZ Network. [Online]. http://www.clickz.com/showPage.html?page= 3625143 (Referenced April 08, 2008) Business Wire. 2002. Sale of Sonera Info Communications Completed; The Company Continues to Operate Under the Name of Fonecta Oy. [Online] http://findarticles.com/p/articles/mi_m0EI N/is_2002_March_27/ai_84217486 (Referenced April 08, 2008) Church, K. & Keane, M. T. & Smyth, B. 2005. Towards More Intelligent Mobile Search. In IJCAI '05: Proceedings of the 19th International Joint Conference on Artificial Intelligence. p. 1675-1676. http://www.ijcai.org/papers/post-0135.pdf (Referenced April 08, 2008) Church, K., Smyth, B., Cotter, P., and Bradley, K. 2007. Mobile information access: A study of emerging search behavior on the mobile Internet. ACM Trans. http://doi.acm.org/10.1145/1232722.1232726 Docter, J., Licciardi, C. A., and Marchetti, M. 2007. The Telecom Industry and Context Awareness. In Proceedings of the international Conference on the Management of Mobile Business. ICMB. IEEE Computer Society. Gohring. 2008. T-Mobile Dumps Google in Europe for Yahoo!. PC World Magazine. [Online]. IDG News Service. http://www.pcworld.com/article/id,142439pg,1/article.html (Referenced April 08, 2008) Hawking. 2006. Web Search Engines: Part 1. Computer, vol. 39, no. 8, pp. 88-90. http://doi.ieeecomputersociety.org/10.1109/MC.2006.28 6 Hawking. 2006. Web Search Engines: Part 1. Computer, vol. 39, no. 6, pp. 86-88. http://doi.ieeecomputersociety.org/10.1109/MC.2006.21 3 Immonen, O., Galli, L., Haaker, T., Killström, U., Pitkänen, O., de Reuver, M. 2006. Can advertising based earnings logic become a basis for future mobile business models? Proceedings of the 3rd International CICT Conference, Mobile and Wireless Content, Services and Networks, Technical University of Denmark, Copenhagen. ITU (International Telecommunication Union). 2008. ICT statistics. [Online] http://www.itu.int/ITUD/ict/statistics/ict/index.html (Referenced April 08, 2008) Kamvar, M. and Baluja, S. 2006. A large scale study of wireless search behavior: Google mobile search. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Montréal, Québec, Canada, April 22 - 27, 2006). R. Grinter, T. Rodden, P.
Aoki, E. Cutrell, R. Jeffries, and G. Olson, Eds. CHI '06. ACM, New York, NY, 701-709. http://doi.acm.org/10.1145/1124772.1124877 (Referenced April 08, 2008) Kamvar, M. and Baluja, S. 2007. Deciphering Trends in Mobile Search. Computer 40, 8 (Aug. 2007), 58-62. DOI= http://dx.doi.org/10.1109/MC.2007.270 (Referenced April 08, 2008) Li, Y., Wang, Y., and Huang, X. 2007. A RelationBased Search Engine in Semantic Web. IEEE Trans. on Knowl. and Data Eng. 19, 2 (Feb. 2007), 273-282. http://dx.doi.org/10.1109/TKDE.2007.18 Netcraft. 2008. March 2008 Web Server Survey. [Online] http://news.netcraft.com/archives/web_ server_survey.html (Referenced April 08, 2008) Paulson, L. D. 2005. Search Technology Goes Mobile. Computer 38, 8 (Aug. 2005), 19-22. http://dx.doi.org/10.1109/MC.2005.270 Schusteritsch, R., Rao, S., and Rodden, K. 2005. Mobile search with text messages: designing the user experience for google SMS. In CHI '05 Extended Abstracts on Human Factors in Computing Systems. CHI '05. ACM, New York, NY, 1777-1780. http://doi.acm.org/10.1145/1056808.1057020 Schonfeld. 2008. Nokia Plays Nice With Google In Mobile Search. TechCrunch. [Online]. http://www.techcrunch.com/2008/02/12/nokia-playsnice-with-google-in-mobile-search/ (Referenced April 08, 2008) Sonera Media Lab. 2002. Mobile Search Engines Whitepaper. [Online] http://www.medialab.sonera.fi/wo rkspace/MobileSearchEnginesWhitePape.pdf (Referenced April 08, 2008)