Mobile Web Analytics

From Seo Wiki - Search Engine Optimization and Programming Languages

Jump to: navigation, search
Internet marketing
Display advertising
E-mail marketing
E-mail marketing software
Interactive advertising
Social media optimization
Web analytics
Cost per impression
Affiliate marketing
Cost per action
Contextual advertising
Revenue sharing
Search engine marketing
Search engine optimization
Pay per click advertising
Paid inclusion
Search analytics
Mobile advertising

Mobile web analytics studies the behaviour of mobile website visitors, in a similar way to traditional web analytics (or desktop web analytics). In a commercial context, mobile web analytics refers to the use of data collected as visitors access a web site from a mobile phone. It helps to determine which aspects of the website work best for mobile traffic and which mobile marketing campaigns work best for the business; this includes mobile advertising, mobile search marketing, text campaigns and desktop promotion of mobile sites and services.

Data collected as part of mobile analytics typically includes traditional information such as page views, visits, visitors, countries, but additionally includes mobile specific information related to mobile device (model, manufacturer, screen resolution, possible ad-hoc grouping based on device capabilities such as "smartphone", etc.), mobile network operator (carrier), preferred user language, etc. This data is typically compared against key performance indicators for performance and return on investment, and is used to improve a web site or mobile marketing campaign's audience response.

It is possible to browse websites on the majority of smartphones produced today. Some mobile phones are better designed for this such as the Nokia N95 or Apple iPhone which provide a more comprehensive browsing experience. The W3C Mobile Web Initiative gives best practices and technologies to help web sites better support mobile phone access.

Many companies have used the W3C mobile web guidelines and standards to provide specific websites designed to work best on mobile devices. This includes sites such World Wrestling Entertainment.


Why not use standard web analytics technologies?

Collecting Mobile Web analytics data is not as straight forward as traditional web analytics since many of the common methods for data collection either do not work or are at best unreliable. Using traditional tools such as Google Analytics may still appear to work but will actually provide misleading data that represents only a percentage of actual mobile traffic.

For example, when applying traditional analytics software on a Mobile Web site you may only see page views, visits, etc. counted only for cellular wireless (mobile phone) HTTP requests coming from the most advanced mobile browsers such as those found in the iPhone and other smart phones and PDAs rather than all the mass market devices actually browsing your site from a cellular mobile network.

Also, traditional web analytics software that uses server log parsing and associates different IPs with "unique visitors" may fail miserably in identifying unique visitors, since the IPs from which cellular wireless network HTTP requests originate from are the Gateway IPs of the network access providers (carriers).

Note that Mobile Web sites are usually open to access from any kind of network (fixed, Wi-Fi, cellular wireless, satellite wireless, etc.) so depending on where the HTTP requests are coming from, a traditional web analytics solution could be from pretty much accurate to mostly innacurate.

In addition, Mobile Web analytics involve metrics and KPIs associated with mobile device information (model, manufacturer, screen resolution) which can be usually assembled by combining device identification information taken from special HTTP headers (such as user-agent) with device capabilities stored in a device information registry (such as WURFL or DeviceAtlas. This is not provided by traditional web analytics solutions since it is only related to the Mobile Web.

Problems with tracking visitors, visits and clickpaths in the Mobile Web

Visitor Identification

This is probably the most important aspect of usable mobile web analytics and one of the hardest technical aspects to get right, primarily because JavaScript and HTTP cookies are so unreliable on Mobile Web (see below). As a result, some mobile web analytics solutions only detect or count user visits today. The best solutions provide a reliable, persistent and unique user identity for each visitor which permits accurate measurement of repeat visits and long term customer loyalty.

JavaScript page tagging

This notifies a third-party server when a page is rendered by a web browser. Assumes that the end user browser has JavaScript capabilities (which is not always the case in mobile handsets, especially the low end ones) and that JavaScript is enabled (which is not always the case since the handset user may disable it). Unfortunately most mobile web browsers at the time of this writing (October 2009) do not support JavaScript sufficiently for this to work.

HTTP Cookies

Commonly used in the World Wide Web to mark and identify visitors since it is a standard capability of all desktop web browsers, however HTTP cookies are only supported by a small subset of mobile phones, which results in visitors not being properly identified and thus visitor and visits related metrics not being calculated properly. Even when cookies are supported by a device, there is a probability that the device will automatically clear cookies between browsing sessions even if the cookie expiration time is set to a future date (also called "persistent cookies"),to save on handset memory or because the browser does not have access to persistent storage (in a small percentage of very old devices). The device user may also disable cookies.

HTTP Referer

HTTP referer information showing where a visitor navigated from is not provided on mobile in many cases. This is either because the device manufacturer has disabled sending of such information in the HTTP request to save bandwidth during network usage or because the mobile network operator's internet gateway is removing or altering the original HTTP header (because of setup of the gateway software or use of mobile web transcoding software).

Image tags

Using images for page tagging is impacted by handset caching mechanisms. In some cases caching of images on handsets is performed regardless of any “don’t cache me” headers output by the remote server.

IP address

The network address of the client machine will usually give some form of user identification and location on the desktop web. Unfortunately the client IP address on the mobile web refers to the internet gateway machine owned by the mobile network operator. For devices such as the BlackBerry or for phones using Opera Mini browser software, the IP address refers to an internet gateway machine in Canada or Norway owned by those companies.

Ways to collect mobile web analytics data

Collecting mobile web analytics data requires a different approach to traditional web analytics. There are a number of solutions available and the best results are obtained through the use of more than one technology.

Packet Sniffing

Also known as tagless data capture or passive network capture, this technique uses a tap between the mobile users and the web server to capture the full content of the client/server exchange. Tagless data capture techniques are increasing in their popularity for mobile web analytics because they capture all users, work with all devices and don't require javascript, cookies, server logs or plugins.

Image tags or beacons

Images can be forced to work on mobile providing the image is always unique. The level of information recorded depends on the architecture provided by the supplier and not all image beacon solutions are the same on mobile.

Link redirection

An important method of tracking mobile visitor activities. It is the only reliable way to record clicks from advertising, search and other marketing activities. It also records visitors clicking on links to leave a site - perhaps to a partner site. This is an important requirement that helps address the lack of http referer information on mobile.

HTTP Header analysis

Tells you a number of basic facts about the mobile phone and the browser. This can be used in conjunction with a device database such as WURFL.

IP address analysis

An operator database is used to identify both the operator and their country based on the IP address of their internet gateway device. IP address alone does not identify all operators and countries - some operators share their mobile network with a virtual network operator or MVNO, for example Boost uses the Sprint network. These two operators have very different customer demographics making clear identification critical for good mobile marketing campaigns. MVNO identification requires additional steps on top of basic IP address analysis. Additionally, some carriers share their mobile internet gateways, sometimes across multiple countries and many change gateways or introduce new ones on a regular basis. Additional steps are also required in all these cases to constantly give accurate results.

External links

Personal tools

Served in 0.147 secs.