Mobile is much different
The mobile revolution is in full swing—if you want to stay competitive, you have to invest and change your mindset.
Anyone who gets to grips with the mobile web is soon reminded of the Wild West, and the world of the desktop web appears to be an island paradise. When it comes to producing content and measuring reach, it is difficult to build on previous methods. Those who want to secure or expand their market success will have no choice but to invest in adapting their entire production processes. Informal discussions in connection with the WEmad.org initiative (see separate section) indicate that dealing with the mobile web is still far from routine everywhere. This is because much is new and different about the mobile web:
1 What has proven itself on the desktop does not necessarily apply to mobile. Although the mobile ecosystem is based on the desktop web, mobile devices with their special characteristics have new weak points and lead to corresponding new best practices. Even with the latest smartphones, the limited transmission bandwidth must always be taken into account. And Web 2.0 features such as JavaScript/AJAX should only be used with caution, as they can put a constant strain on the processor and thus drain the battery quickly. In addition, mobile phone operators can interfere with the transmission and, for example, reduce the size of images or block access to websites completely.
2. mobile web users have new usage patterns and navigation methods. They are extremely goal-oriented and cost-conscious. The fear of high costs prevents them from surfing comfortably, and they would rather interrupt the transmission (of an unnecessarily overloaded website, for example) than risk a low bill. People also only surf on the move where the network coverage is good, as a poor signal can lead to costs if the transmission is incomplete, but not to the desired gain in information.
3. mobile web browsers are new, numerous and full of problems. While there are two dominant browser programs for the desktop web, there are over a dozen in use on mobile devices. They all have their own advantages, but also quirks and corresponding workarounds. Established web standards are often only partially and incorrectly implemented, meaning that incorrectly prepared websites can easily cause the browser or even the entire device to crash. And there is no improvement in sight: with the flood of new mobile devices, the variety of displays (e.g. black and white eInk screens on eReaders), navigation (e.g. voice-supported as on the Nexus One) and therefore browser variants will continue to increase. Web publishers must be able to automatically recognize the capabilities of the devices (using WURFL or DeviceAtlas) and adapt the output much more flexibly than before.
Mobile analytics also measures well-known key figures such as views, visits and visitors. However, conventional traffic analysis tools fail with the mobile web or are unreliable at best, as they often only record a fraction of the actual mobile data traffic. This is particularly true when determining unique visitors. As a rule, a visitor is assigned to a network address, but mobile users access the Internet via an access node (gateway) of the respective mobile network operator and appear with the network address of this computer, meaning that they cannot be differentiated using the conventional analysis method. For example, all Blackberry users appear to come from Canada, and anyone surfing on their cell phone with the Opera browser appears to come from Norway, because the corresponding gateway computers are located in these two countries. And anyone surfing on a moving train or in a car, for example, often changes their network address along with the mobile phone antennas they are using.
In order to identify visitors on websites, cookies and special Javascript solutions are used on the web. Currently, many mobile web browsers do not support JavaScript, or users have disabled this feature. And although cookies are accepted, they are not stored permanently. If permanent cookies are missing, visitor identification between sessions is not possible.
And the one-pixel image tags also fall victim to the technical constraints on mobile devices: the web server tells the user that the image file should not be cached so that it can be counted again the next time the page is called up. However, this instruction is ignored by certain mobile browsers because performance can be improved by avoiding a further retrieval. And so another analysis tool loses its significance. Those who are active in online marketing (SEA/SEO) also often miss the origin information (referrer) in the mobile area, which shows from which page a visitor came to their own page.
Collecting mobile web analytics data therefore requires different methods to conventional web analytics. There are a number of solutions for this, which are best used in combination. Pixel tags can be delivered with a different name each time, web traffic can be redirected to intermediate computers for analysis and the web traffic itself can be intercepted and analyzed accordingly. Many of these techniques are available in commercial web analysis tools. But there are also a number of free offerings for mobile analytics. For example, the widely used Google Analytics tool announced corresponding extensions in November, and providers such as Mobilytics, Admob Analytics, Amethon Analytics and Bango Analytics offer further solutions.
Christoph J. Walther
> Comparison between the desktop web and the mobile web (pdf)
1 What has proven itself on the desktop does not necessarily apply to mobile. Although the mobile ecosystem is based on the desktop web, mobile devices with their special characteristics have new weak points and lead to corresponding new best practices. Even with the latest smartphones, the limited transmission bandwidth must always be taken into account. And Web 2.0 features such as JavaScript/AJAX should only be used with caution, as they can put a constant strain on the processor and thus drain the battery quickly. In addition, mobile phone operators can interfere with the transmission and, for example, reduce the size of images or block access to websites completely.
2. mobile web users have new usage patterns and navigation methods. They are extremely goal-oriented and cost-conscious. The fear of high costs prevents them from surfing comfortably, and they would rather interrupt the transmission (of an unnecessarily overloaded website, for example) than risk a low bill. People also only surf on the move where the network coverage is good, as a poor signal can lead to costs if the transmission is incomplete, but not to the desired gain in information.
3. mobile web browsers are new, numerous and full of problems. While there are two dominant browser programs for the desktop web, there are over a dozen in use on mobile devices. They all have their own advantages, but also quirks and corresponding workarounds. Established web standards are often only partially and incorrectly implemented, meaning that incorrectly prepared websites can easily cause the browser or even the entire device to crash. And there is no improvement in sight: with the flood of new mobile devices, the variety of displays (e.g. black and white eInk screens on eReaders), navigation (e.g. voice-supported as on the Nexus One) and therefore browser variants will continue to increase. Web publishers must be able to automatically recognize the capabilities of the devices (using WURFL or DeviceAtlas) and adapt the output much more flexibly than before.
Mobile analytics also measures well-known key figures such as views, visits and visitors. However, conventional traffic analysis tools fail with the mobile web or are unreliable at best, as they often only record a fraction of the actual mobile data traffic. This is particularly true when determining unique visitors. As a rule, a visitor is assigned to a network address, but mobile users access the Internet via an access node (gateway) of the respective mobile network operator and appear with the network address of this computer, meaning that they cannot be differentiated using the conventional analysis method. For example, all Blackberry users appear to come from Canada, and anyone surfing on their cell phone with the Opera browser appears to come from Norway, because the corresponding gateway computers are located in these two countries. And anyone surfing on a moving train or in a car, for example, often changes their network address along with the mobile phone antennas they are using.
In order to identify visitors on websites, cookies and special Javascript solutions are used on the web. Currently, many mobile web browsers do not support JavaScript, or users have disabled this feature. And although cookies are accepted, they are not stored permanently. If permanent cookies are missing, visitor identification between sessions is not possible.
And the one-pixel image tags also fall victim to the technical constraints on mobile devices: the web server tells the user that the image file should not be cached so that it can be counted again the next time the page is called up. However, this instruction is ignored by certain mobile browsers because performance can be improved by avoiding a further retrieval. And so another analysis tool loses its significance. Those who are active in online marketing (SEA/SEO) also often miss the origin information (referrer) in the mobile area, which shows from which page a visitor came to their own page.
Collecting mobile web analytics data therefore requires different methods to conventional web analytics. There are a number of solutions for this, which are best used in combination. Pixel tags can be delivered with a different name each time, web traffic can be redirected to intermediate computers for analysis and the web traffic itself can be intercepted and analyzed accordingly. Many of these techniques are available in commercial web analysis tools. But there are also a number of free offerings for mobile analytics. For example, the widely used Google Analytics tool announced corresponding extensions in November, and providers such as Mobilytics, Admob Analytics, Amethon Analytics and Bango Analytics offer further solutions.
Christoph J. Walther
> Comparison between the desktop web and the mobile web (pdf)
