4.1. technologies driving the evolution of mobile search
From a general perspective, there are three main technology families that have a direct impact on mobile search: enabling technologies, search technologies (in general) and specific mobile search technologies, as described in Table 9.
These different technologies are presented to show their expected evolution path and the emerging trends that can foster the use of contextual parameters (such as the user location or user profile), the rise of mobile social network(s), and the evolution towards a real multimedia search in mobile environments.
4.1.1 ENABLING TECHNOLOGIES
184.108.40.206 Network Technologies
From a European mobile sector perspective, mobile broadband technologies are present all over the EU, particularly 3G (UMTS) and 3.5G (HSPA), following the technology evolution defined by the ETSI/3GPP system architecture (also followed in other regions). The different releases show the network evolution and the future standards for mobile radio networks, with 4G being the next technological challenge for the industry, referred as Next Generation Mobile Networks.
The target architecture defined will be an optimised packet switched network architecture, which will provide a smooth migration from existing 2G and 3G networks towards an IP network with improved cost competitiveness and broadband performance.
The term 4G refers to the next level of evolution in the field of wireless communications. 4G systems will replace completely existing communication networks, providing a more comprehensive and secure IP based solution to users on an “anytime, anywhere” basis and at much higher data rates compared to previous generations.
Table 9. Technologies directly impacting on mobile search.
Figure 16. The roadmap from GSM over UMTS to NGMN.
Therefore, 4G can be viewed as a further step in the evolution of current industry efforts in the HSDPA, HSUPA, and EVDO arenas, enabling a personalised broadband access experience and consolidating the diversity of networks operated by mobile network operators.
In this context, Long Term Evolution (LTE) is the last stage from existing 3G to 4G, currently under discussion by the 3GPP. LTE is a set of enhancements to the Universal Mobile Telecommunications System (UMTS) introduced in 3GPP Release 8, basically focused on enhancing the Universal Terrestrial Radio Access (UTRA) and optimizing 3GPPs radio access architecture. The aim is to provide an average user throughput of 3 to 4 times the release 6 HSDPA levels in the downlink (100Mbps), and 2 to 3 times the HSUPA levels in the uplink (50Mbps).
The evolution towards a higher bandwidth in mobile systems and the subsequent reduction of both distance among cells and cell area, can enable a better user experience in the context of multimedia search (higher broadband capability) and context-aware search (more precise location) where the concept of femtocells (e.g. a UMTS femtocell) is a crucial milestone.
220.127.116.11.1. Mesh networks and cognitive radio
Progressive cell size reduction leads to a further step: mesh networks, in which every device can act as a network node and interact with nearby devices. The topology of a mesh network is stable and highly reliable, as each node is connected to several others. If one node drops out of the network due to hardware failure or any other reason, its neighbours can easily find an alternative route using a routing protocol.
In addition, cognitive radio technologies allow that either a network or a wireless node changes its transmission or reception parameters to communicate efficiently avoiding interferences with surrounding users.
All these features show the potential of cognitive radio technologies and mesh networks as enablers of context-aware and augmented reality applications and services, due to their capability of monitoring and adjusting several factors in the user environment, such as radio frequency spectrum, user behaviour and network state.
18.104.22.168 Sensor Networks
Context-aware search relies on technologies that provide trustworthy and reliable information of the user environment to further convert context information into services and applications.
22.214.171.124.1. Wireless technologies as context enablers
Main context-aware enablers in wireless technologies are referred to the advances in the devices and technologies based on RFID, Wireless Sensor Networks (WSN) and ad-hoc wireless networks which can be coupled to the mobile device and other enabler embedded technologies to the mobile device or network such as the location systems using the mobile network parameters.
The main trend in WSN is developing new communication standards that provide better location information and more bandwidth such as IEEE 802.15.4a. The family of standards IEEE 802.15 allows high aggregate throughput communications and low power usage within the scope of the user environment, defining the so called Wireless Personal Area Networks (WPAN).
Near Field Communication (NFC) appears as one of the most promising extensions of RFID technologies for mobile devices. As defined by the NFC Forum, “Near Field Communication (NFC) is a short-range wireless connectivity technology that evolved from a combination of existing contactless identification and interconnection technologies. Products with built-in NFC will simplify the way consumer devices interact with one another, helping people speed connections, receive and share information and even make fast and secure payments”.
Although NFC was expected to be used mainly for payment operations using the mobile phone, it allows different alternatives as enabler of mobile search applications based on the user context.
126.96.36.199.2. Environment monitoring
Environment monitoring is referred to those technologies designed to “tag” and “understand” the environment, which can be combined with wireless technologies or cooperate separately.
Examples include the Sekai Camera for the Apple iPhone, which combines the visual information from the mobile device camera with the GPS or 3G network location information and the stored information about the local environment where the user is located. Another example is the audio matching techniques to tag the environment with the sound received using the mobile device microphone for example, for a song search.These same techniques and technologies can be applied to obtain information about the user environment.
In addition, all the mentioned technologies for enabling search applications should perform in connection to a reliable information database.Apart from textual information, a major challenge in this context is the creation of complete audiovisual databases to support multimedia search.
Figure 17. Evolution of miniaturization and price reduction enablers.
188.8.131.52.3. Internet of Things
The RFID Working Group of The European Technology Platform on Smart Systems Integration (European Commission – EPoSS, 2008) says that the definition of “internet of things” can have different facets depending on the perspective taken. From a functionality and identity point of view it is defined as “things having identities and virtual personalities operating in smart spaces using intelligent interfaces to connect and communicate within social, environmental, and user contexts”. A different definition, that puts the focus on the seamless integration, is “interconnected objects having an active role in what might be called the Future Internet”.
From a general perspective the concept of internet of things could be considered as the ideal combination of technologies and communications systems presented in previous sections in which short-range mobile transceivers are embedded into all kind of gadgets and everyday items, enabling new forms of communication between people and things, and between things themselves.
The internet of things represents the real technological revolution and challenge for the present and future of computing and communications. As the ITU suggests (ITU, 2005), “a new dimension has been added to the world of information and communication technologies (ICTs): from anytime, anyplace connectivity for anyone, we would have connectivity for anything”.
In this context, RFID and related identification technologies will become the cornerstone of the upcoming internet of things, using a single numbering scheme to make every single object identifiable and addressable. Smart components would be able to execute different set of actions, according to their surroundings and the tasks they are designed for.
According to European Commission – EPoSS (2008), to reach such a level of ambient intelligence major technological innovations and developments will need to take place, amongst them governance, standardization and interoperability being absolute necessities to develop the internet of things vision. New power-efficient, security-centred and global communication protocols and sustainable standards must be developed, allowing vast amounts of information to be rapidly shared between things and people. The ability of the smart devices to withstand any kind of harsh environment and harvest energy from their surroundings becomes critical. Furthermore, a major research issue will be to enable device adaptation, autonomous behaviour, intelligence, robustness, and reliability.
One of the key issues of the internet of things will be related to trust, privacy and security, not only for what concerns the technological aspects, but also in terms of the education of the people at large. The growing data demand and higher data transfer rates will require stronger security models employing context related security, which in return will help citizens to build trust and confidence in these technologies rather than increasing fears of total surveillance scenarios.
184.108.40.206 Device Technologies
Big players are closely following the mobile platform due to its importance for designing the user framework in which services, applications and content will be based upon. Industry is analysing diverse options with regard to the openness of platforms, from closed (walledgardens) to open environments (Ballon and Wallravens, 2008). On one side, the closed approach has the advantage of offering search capabilities adapted to the specific platform, in order to provide a particular service only available through the corresponding platform. This is the case of the Microsoft OS platform for mobile devices, which is evolving towards providing a complete solution with embedded search capabilities. On the other side, platforms, on their side, are evolving towards unification in a single standard that could be implemented first on top segment devices. Major initiatives towards creating one open mobile software platform are Symbian OS, S60, UIQ and MOAP(S) (Nokia,Sony Ericsson, Motorola and NTT DOCOMO). In addition, Google Android and the Linux project are developing open source platforms with the support of device manufacturers including HTC and Motorola. It is expected that the number of platforms will be reduced and converge towards providing integrated search features, as a vast majority of mobile players agreed that platform openness is the key feature for promoting mobile search and other mobile data applications.
The development of location-based and context-aware search services are favoured by embedded GPS receiver and tactile screens for mobile web navigation, which facilitates an enriched user experience that has a positive impact on search applications. Embedded camera and sensors within the mobile phone (gyroscopes and accelerometers) and easy-touse mobile phones displays (bigger, tactile) are also useful feature for facilitating the use. These features are increasingly demanded, as smartphones sales showed for 2008, and offered by the main vendors and manufacturers. Table10 offers an overview of some of the main smartphone manufacturers.
Table 10. Major smartphone manufacturers (2008).
220.127.116.11 Cloud computing.
Cloud computing is on-demand computing service; the software does not reside at the users’ device. The computing resources are owned and managed by a service provider and the users access the resources via the internet. Cloud computing is a highly important phenomenon influencing network and computing architecture, thus also setting the framework under which mobile data services and applications will be developed. As in this architecture user’s files and folders are stored in the “cloud”, users can access their data and applications everywhere and at any time only requiring a mobile device with internet access. Current cloud computing applications in the mobile realm include mobile email, mobile search, and navigation apps. Among the benefits of cloud computing, the location independence is an obvious one. Device independence is equally important and cloud computing may pave the way for further convergence of PC and mobiles as services regards.
As could computing ‘only’ needs a browsers–which is are already provided with anysmartphones (and more alternative browsers available for download)– this technology could free-up processing and storage power of handsets. Thus, it could become a “standard” in the way mobile applications are built and run and allowing developers to create a single version of their applications, promising greater future compatibility.
Could computing will shape the way doing business both in the PC and the mobile world.
In the mobile environment, cloud computing is a potential way to bypass mobile applications that are tied to a certain carrier or manufacturer, and may contributing to opening the market to alternative providers. The value chain may also shape up differently. Future mobile data applications, like mobile search, are online services likely to be provided over the internet through a web browser, while the software and data are stored on the servers. Mobile search tools would benefit from cloud computing as the user information is centralised on a number of “cloud” servers, thus simplifying search operations. A major challenge will be to keep a real time update of the user’s context and device data in order to provide accurate mobile search services, particularly context-aware services. In this context, data security becomes critical. Security typically improves with centralization of data operations, but raises concerns about the potential loss of control over (sensitive) data (personal data, location data, etc). This calls for a transparent and secure manner to guarantee user’s privacy.
4.1.2 search technologies
18.104.22.168 Multimedia search
Currently the most common process to conduct multimedia search is to retrieving meta-data annotated to the audiovisual content. Relevant metadata to the content can be annotated either automatically through algorithms or by personal/social interaction (an example is Google’s Image Labeller game). A review of annotation techniques can be found in Kompatsiaris (2008). The research progress is steady in overall terms, but with differences by topics, given the complexity of the research challenge Ideally, the most useful schemes for multimedia retrieval would not require previous annotation or done on the spot, i.e. the retrieval would be based on direct visual and/or audio search. Considerable research effort is spend in this world wide; at Europe scale by the ICT programme of the European Commission (DG Infso).
In visual search, a common request is finding an object embedded in an image or a video clip. An example would be receiving information about an object captured by a mobile camera. A typical procedure for “augmented reality” or “reality mining” applications would be taking a picture of a mountain landscape and getting back the names of the peaks indicated on top of each of them. It is also conceivable to search for codes or characters embedded in the query image are conceivable (Xie et al., 2008). Despite technological developments, search in video files will take long time to be operational (perhaps 5 years or more). There are no commercial developments with remarkable results yet, but there are two areas with clear progress and already in some practical use. The first one is the so-called “2-D bar codes” where a logo or a specific pattern composed of points are captured by the mobile device camera and compared with a pre-existing database. This kind of application is being used for marketing purposes and for ticketing (events, travel, etc). In the second area, is an augmented reality browser that uses the images captured by the mobile device camera (for instance a skyline of a city) to retrieve some information about the objects around (for instance, information about a building). Side information, like the geographical location and orientation of the camera, is used to improve the search process to supply more relevant results. Main applications stores (i.e., iPhone’s and Android’s) contain an increasing range of instances of such browsers.
Audio search is definitively gaining momentum and further improvements are expected in the coming next 3-5 years. The number of applications is increasing and search technologies are becoming more effective and reducing the error rate. There are several techniques for audio search based on the process of low-level characterisation of the audio signal (signal low level coefficients) and the matching of similarities between signals. Typical examples in use are looking for ringtones (Lie et al., 2008) or trying to find the title of a song played in the user environment using the mobile device (e.g. the application “Listen” installed on the Apple’s iPhone).
Two final notes on multimedia search. The first one relates to the increasing relevance of multimodal queries (Xie et al., 2008) that allows to query by different types of content (annotations, metadata, audio and video) and that takes into account additional side-information (location, context, etc). Multimodal queries are more complex to carry out and this performed in practice by dividing the search task into several processes which are later combined to supply the result of the query. The second note refers the long-standing difficulties in performing multimedia search on how to bridge the “semantic gap” between high and low level descriptions of multimedia content. As an example, the analysis of an image using algorithms is different from what a human can understand by looking at and analysing the same image: this is what is called the semantic gap. In this sense, evolution in multimedia search has been conditioned by an appropriate talking of search, considered in the next sub-section.
22.214.171.124 Semantic Search
Semantic search refers to the process of using semantic ontologies for retrieving results with meaningful concepts and sentences rather than just independent terms, which often may even not related with the real aim of the query. An ontology specifies the formal representation of a set of concepts within a domain and also the relationships between those concepts including the definition of classes and functions.
The goal of semantic techniques is to develop an interconnected database of information and documents (the web 3.0 Internet in its most ambitious incarnation) accessible through the “natural” way humans ask questions and obtain responses. Applying this goal to search, it implies obtaining and presenting results related to what the question really means and not only a “blind combination” of the terms of the search.
As in the case of multimedia search, progress in semantic techniques relates also to advances in artificial intelligence. Both methods share the objective to improve the man-machine interface; the search process and the presentation of results in such a way that it resembles the way humans “interpret and interact with the world”. In the particular case of the mobile domain, introducing both kinds of techniques could enhance the usability and usefulness of search since users have customary routines with multimedia information (voice, pictures) and the input and output means are limited.
Some companies (mainly from the USA) like Yahoo!, IBM, and Google are making important efforts towards providing a more human-adapted search. In Europe there are some research initiatives funded at the level of the European Commission and at the level of EU Member States, more notably the French QUAERO project or the German THESEUS.
126.96.36.199. Cognitive technologies
Cognitive technologies refer to technologies that “understand” the information captured from the user environment in a similar way to what humans do, and therefore they are able to process it and attribute it with some meaning. They belong to the general field of artificial intelligence. Cognitive technologies for mobile search find already some modest implementations for at least three purposes, namely to profile the user, for recommendation and priority ordering, and for processing the environment .
188.8.131.52.1. User Profiling
To enable personalised mobile search (and also contextually adapted search) it is necessary to capture the user profile and process it to transform it into information linked with user interests and desires and, therefore, able to improve the results of search.
Profiles can be constructed from various types of information sources, including user’s tastes, user’s behaviour inferred from the consumption of mobile services, the social networks to which they belong, etc., or a combination of all. Services consumption is normally measured using audience measurement techniques, and further analysed through data analysis statistical methods. Lancieri and Durand (2006) describe some methods for internet user behaviour analysis based on access traces and its application to discover communities based on a self-similarity model. Other authors (Murata and Saito, 2006b, 2006a; Murata, 2004) extract audience information and user’s interest from the routine visits and web log data. Over the past years, audience measurement technologies have evolved to cover several services platforms. User’s profiles have been created and user’s behaviour has been modelled based on such data in several research studies, e.g. in Álvarez et al. (2009). A method for inferring identity from user’s behaviour using Bayesian statistics can be found in Carey et al. (2003).
184.108.40.206.2. Prioritising and Recommanding
Cognitive technologies are employed to offer recommendations on content supplied to users based on their past behaviour on content consumption. They are mostly based on content filtering tools (Xie et al., 2008) and can be clustered into three main methods: content-based filtering, collaborative filtering and hybrid methods, which are briefly described in the following.
Content-based filtering is a technique related with user profiling, where the user’s preferences are inferred from the consumption of mobile services. Collaborative filtering acknowledges the fact that for privacy reasons or for marketing purposes it is highly desirable to characterise the profile of homogeneous communities of individuals. It is lately been improved thanks to the success of user communities and social networks. In this area two major groups of algorithms can be distinguished: memory-based (Yu et al., 2004) and model-based approaches (Melville et al., 2002). In memory-based approaches, a rating prediction is made upon the ratings of other users with similar interests. A modelbased collaborative filtering technique (Melville et al., 2002) first completes a statistical model on the community of users. Then it predicts the ratings based on the acquired model parameters. Hybrid methods (Boutemedjet and Ziou, 2008; You and Wong, 2007) use input from multiple services and applications to build a recommendation prediction. They are mainly used for textual search, and are now expanding also into other multimedia search.
220.127.116.11.3. ‘Processing the Environment
Once a profile is created and recommendations are proposed, the results can be combined with the user context, which need also to be captured. It is also possible to directly tailor the results from the recommendation engine could be tailored to the location and other information on the surroundings of the user.
Due to the diversity of potential context information, the cognitive systems do focus on a limited set of environment variables, later to be used in the extraction of meaningful information about the context. Among the different types of information currently available in the mobile environment, location is the prime example.
Typical information about the surrounding objects is amongst the most used. Here, information about the location is captured either by GPS or through mobile signal information processing. In the first case, the GPS provides an excellent precision of the location but consumes a significant amount of batteries energy. The second option consists in locating the terminal via the cell ID of the mobile network or using other signal processing techniques, e.g. the signal strength or the time-of-arrival (Gustafsson, 2005). The precision is lower, but also the energy consumption.
Location is just one piece of information. The increasing network of sensors and emitters in the environment and embedded readers mobile device is additional way to obtain information on the context. Other examples of relevance include gyroscopes (for orientation), accelerometers (for in-device movement tracking), or weather measures (temperature, pressure, wind, etc). Cognitive technologies to use this information in the mobile environment and improve search result belong to the context-awareness techniques discussed below.
4.1.3. Mobile search specific technologies
Context awareness is regarded as an enabling technology with a high potential for mobile data applications, particularly in field of mobile search. It refers to all technologies concerned with the acquisition of context (using sensors to collect information about the surroundings or environment), the abstraction and understanding of context (matching a perceived sensory stimulus to a context), and application behaviour based on the recognized context (triggering actions based on context). These systems capture and use contextual information in dynamic way as to optimize, change, or create communications flow and business processes. Contextual information can be collected for any mobile asset involved in a business process, and this includes not just devices and products but also people.
18.104.22.168. Augmented reality
Augmented reality (AR) is a field of computer graphics research that deals with the combination of real-world and computer-generated data (virtual reality), where computer graphics objects are blended into real footage in real time. At present, most AR research is concerned with the use of live video imagery which is digitally processed and “augmented” by the addition of computergenerated graphics. There are many applications of AR currently developed in different fields, i.e. in advertising, medicine, navigation, emergency services, prospecting in hydrology, ecology or geology, visualization of architecture, enhanced sightseeing, flight simulation, and entertainment. AR-enabled interfaced will enrich context-aware technologies, by rendering contextual search more precise and intuitive.
4.1.4. The way forward
One of the purposes of the expert workshop on “Mobile Search Prospects” (Sevilla, 16-17 April 2009) was to identify key issues to take into account for a technological roadmap for mobile search. Participants see a couple trends that are going to influence technological challenges. First, there is a real and increasing need for pure mobile search applications: “find a timetable”, “get me home”, etc. These applications will need technologies and developments that are tailored specifically for the mobile environment and are distinct from the general PC-based search environment.
Second, mobile search does and will coexist with popular dual usages PC/mobile. Although the technical requirements are clearly distinct,
Figure 18. Expected availability and relevance of important technologies for mobile search (2010 – 2016).
people want to perceive the search as a seamless experience across media. This is currently not the case and actions should be taken for consumers to perceive a similar quality of service. Therefore, there is an interest in “convergence” of technologies. This leads to the final challenge: the most important “technical” issue in mobile search is interoperability. In fact, the interconnection and interoperability of technologies was considered a key factor for success; much more than developing specific “hard-core” mobile search technologies.
Based on this general framework, the panel of experts who took part in the workshop in Seville analysed the relevance for search and the time of market appearance for the technologies discussed in the previous chapter. Figure 188 plots the eight aforementioned topics with respect to these both criteria; those on the upper left corner are more important that those on the bottom right. Figure 18 highlights the importance of mobile broadband connectivity, and the availability (and adoption) of smartphone devices. These two factors are necessary pre-conditions to boost other technological implementations.
A good deal of context-aware (#7) and cognitive technologies (#6) are already available (or in an advanced prototype stage to be operational in the near future), but yet show little presence in commercial services and applications.
Once the connectivity (#1) is broadly assured and smartphones (#3) are deployed at large scale, it is expected that the technology relevance drawn would be re-adjusted after 3-4 years, considering that the identified preconditions would have achieved at that time a sufficient level of mass adoption that foster context-aware systems, services and applications usage.
4.2. Mobile market trends: baseline scenario
The same assumptions, notation and definitions made in the introduction to section 2.1 are used in this section of the report.
22.214.171.124. Mobile worldwide subscriber base
The subscriber base experienced a fast growth during the last years and is expected to continue to grow (Figure 19) with nearly 5,000 million subscribers worldwide in 2012.
This strong growth is expected to be driven by the major emerging economies of Asia (particularly in China, India, Indonesia, Pakistan), Latin America (Brazil, Colombia), Europe (Russia, Ukraine, Turkey) and Africa (South Africa, Algeria, Nigeria). Regional growth is presented in Figure 20.
Figure 19. Mobile subscriber base worldwide (2007 – 2013).
Figure 20. Mobile subscriber base by region (2008 – 2013)
126.96.36.199. Mobile broadband
By the end of 2008, over 362 million years, representing more than 50% of total 3G subscriptions were estimated globally, connections by the end of 2013 (Figure 21). representing around 10% of the 3,686 This is a major factor for mobile internet and million users worldwide. The forecasts is that mobile 2.0 services development, and hence broadband mobile connections (3G, 3.5G+) also for mobile search.
Figure 21. Forecasts for 3G and 3.5G access worldwide (2008 – 2013)
Figure 22. Mobile internet users vs total internet users (2008 - 2013)
188.8.131.52. Mobile internet
Mobile internet adoption is expected to exhibit a fast growth in the following years, achieving up to 40% of total subscribers in 2013 (Figure 22).
Figure 23 presents regional differences in mobile internet users. The Far East and China region will continue to be the largest mobile internet regional market. Mobile internet usage is already high in pioneering countries such as
Figure 23. Mobile internet users by region (2008 – 2013).
Figure 24. Mobile operators’ average revenue per user (2008 – 2012).
Japan and Korea, and China continues to show strong growth in subscriber numbers. Africa and Middle East, and the Indian Sub Continent, will also experience strong growth, as the mobile phone provides the most viable medium for obtaining internet access (due to lack of fixedline infrastructure). Growth will be moderate in more saturated markets such as North America, and Eastern and Western Europe, although it is expected that mobile 2.0 will boost the mobile internet services take up.
4.2.2. Industry revenues
184.108.40.206.1. Average revenue per user
Charging for increasing data traffic has long been sought by mobile operators as the long term solution to declining total average revenue per user (ARPU), mostly voice. However, existing consumption of mobile data services is failing to offset the decline in total ARPU, and despite its growing contribution, this tendency will be maintained during the following years.
Figure 24 shows the forecast that total ARPU will decline, although data will increase its importance, until reaching almost 30% of total by 2012.
220.127.116.11.2. Mobile 2.0
Out of the 30% of operator’s total revenues that data represent, mobile 2.0 applications will become a primary source of revenues, reaching up to 50% of data revenues. Currently main income generators are messaging (SMS, MMS), mobile email, basic contents (like ringtones or wallpapers) and information services. By contrast, nascent mobile 2.0 will be based on social networking and user-generated contents, instant messaging and mobile search.
As Figure 25 shows, although instant messaging currently is the primary source of mobile 2.0 incomes, social network services / user generated content (SNS/UGC) will become the leading mobile 2.0 application. The main assumption considered is that mobile search is expected to be widely adopted from 2010/2011
Figure 25. Global revenues for mobile web 2.0 by application (2008 – 2013).
on, as context-aware technologies improve mobile search engines and results delivery and accuracy.
18.104.22.168.3. Mobile search
As presented in Section 2.2.2 and specifically referring to mobile search, Chard (2008) expects the advertising and user profiling contribution to the total mobile search revenues to grow from 30% in 2008 to around 40% in 2013, being the total mobile search revenues aggregate growth (CAGR) of 27% for the same period (Figure 26).
Figure 26. Mobile search revenues by advertising and data charges (2008 – 2013).
Figure 27. Mobile search revenues by region (2008 – 2013).
According to the Mobile Entertainment Forum (MEF), advertising revenue split ratios will likely be similar to internet ones with about one third for the search solution provider and about two thirds for the publisher, including as a main difference with the web a residual percentage up to 10% for other players in the mobile value network.
According to Chard (2008), Far East and China, Western Europe and North America represent the largest markets for mobile search in terms of total revenues, representing 27%, 21% and 13% of the overall market, respectively, in 2013 (Figure 27). It is important to highlight the leading position of Europe within global market, as the combination of Western and Eastern Europe forecasted revenues surpasses the leading single region, Far East and China. For this forecast to come true, total revenue in Eastern Europe in 2013 should be about five times higher than the 2008 value, reaching a level not far from that of North America.