Nearly every day I read articles on Mobile Marketer about businesses launching new mobile initiatives: from mobile ad campaigns to iPhone apps and from enhanced vendor services to strategic partnerships. Even after cutting through the hype, its obvious enterprises large and small are scaling up their investment in mobile. Yet, as I read news feeds such as these and dialogue with customers, prospects, and partners one theme remains consistent: many mobile initiatives launch with little or no analytics and optimization strategy beyond very basic metrics like click-through rates or downloads. A colleague, Matt Langie, highlighted this theme in his blog six months ago in a post about mobile analytics.

Interest in Mobile Analytics Lags Mobile Web Usage
The divide is also apparent when comparing the percentage of analytic-related searches which include mobile keywords to the percentage of page views which are generated by mobile devices. While mobile analytic search volume is less than .5% of all analytic related searches, many websites are seeing between 1.5% and 10% of total page views originating from mobile devices. For websites heavily investing in mobile initiatives, page requests from mobile devices can be even higher—35%+ of the total page views in some cases. In other words, web usage from mobile devices is accelerating but many businesses haven’t fully engaged analytics and optimization efforts for the channel.

You Don’t Know What You Don’t Know
If your enterprise is launching (or has already launched) mobile specific initiatives, you should ensure an optimization strategy has been formulated and measurement for the initiatives has been deployed. Initial efforts in any “emerging” space have significant opportunity to build the business. However these efforts are more sustainable in the long term if the ROI of the investment can be demonstrated through analytics and then improved through optimization.

Your fledgling mobile initiatives may be wildly successful or they may need tweaking. Either way, the long-term success of such efforts may depend on your ability to measure, analyze, and optimize them.

The first step in quantifying and improving ROI for a mobile initiative is to measure the effort. The purpose of today’s post is to discuss measurement strategies for some of the most common channels including mobile optimized sites, mobile applications, mobile campaigns, and mobile video. The next several sections will detail how to approach the technical implementation of these initiatives so if you’re not a developer, now is a good time to reach out to your favorite engineer or tech lead.

Mobile Optimized Sites
When deploying measurement for a mobile optimized site, there are two main parts to the implementation: (1) the tagging methodology (i.e., JavaScript, manual page coding, server-side scripting, packet sniffing, log file records) and (2) the data transport mechanism (i.e., image beacon, XML, direct HTTP request, FTP—for log-file analysis).

The most common measurement approach for tagging and transport on mobile optimized sites is to leverage server-side scripting to generate a formatted request either through creation of an image sent back to the page which the device then requests from the data collection server or by generating an HTTP or XML request which is sent directly to the data collection server when the device requests the page.

In the SiteCatalyst v14.6 release, Omniture significantly enhanced support for this method of data collection by adding a class which handles tagging, formatting, and transporting requests. Implementation details for the new measurement library can be found in the Help area of SiteCatalyst. Just login and go to Help > Supporting Docs > Manuals > Measurement Library (PHP).

Essentially, the new class allows mobile implementations to closely resemble standard website implementations which reference a few lines of page code, utilize an external library, and ensure best practices are followed. The main difference being the standard website implementation leverages JavaScript whereas the new measurement library leverages a server-side language.

If the new class doesn’t meet your need, you can still deploy measurement by using a server-side script to generate an image beacon, a direct HTTP request, or an XML request. Once basic measurement has been deployed, I suggest reviewing the Mobile Implementation Manual to ensure you are leveraging Omniture’s best practices for implementing mobile sites. The guide can be found in the Omniture Knowledge Base under implementation manuals and is titled “Mobile Device Reporting.”

Mobile Applications
The 1.5+ billion iPhone apps downloaded have certainly caused a big stir. If you are involved in mobile initiatives within the organization you’ve probably considered or been involved in the release of a mobile application on one or more popular mobile platforms (iPhone, Blackberry, Android, Palm Pre, Symbian, etc.).

Although some customers have been measuring mobile application usage with Omniture for years either using the XML API or Data Sources, Omniture significantly enhanced support for measuring interactions with mobile apps in two releases this year (January & July). The new methodology allows customers to leverage Omniture libraries native to the platform the application is written in.

For example, to deploy measurement for an iPhone app you simply embed the Omniture iPhone components into the existing iPhone app, customize variables, and call the Omniture routine (track or tracklink) to send the analytics. The implementation is very similar for the other supported mobile app platforms. One advantage to this style of deployment is measurement can be deployed very quickly—less than an hour for a basic implementation in some cases.

If you’d like to review the specific implementation steps for the platform(s) you will be launching your app on, you can find step-by-step implementation guides in the Knowledge Base. Just login to SiteCatalyst and go to Help > Supporting Docs > Manuals > App Measurement for Android/Blackberry/iPhone/Symbian (implementation steps for Palm can be found in the last post on Palm Pre app measurement).

If the native libraries aren’t the right fit, you can still deploy a measurement solution using XML, Data Sources, or the new server-side measurement library. Details on how to implement these methods can be found in the knowledge base.

Mobile Campaigns (Banner, SEO, SEM, Barcode, Messaging, Widgets, Bookmarks, E-mail, etc.)
Deploying measurement for mobile campaigns almost can’t be oversimplified. Measuring mobile campaigns which drive users back to a website can be split into two parts: (1) providing a unique url which distinguishes the campaign from other entry methods, and (2) creating a mechanism for sending this campaign identifier to Omniture.

To accomplish #1, it’s usually best to settle on a unique URL query parameter which identifies campaigns and pair that with an identifying value unique to the campaign. For example, a link encoded in a barcode might look something like the following:

bar code

Once the QR barcode has been scanned with a reader, it’s decoded to a URL that looks like this:

The URL can then be accessed on the device to take the user to

Once the user accesses the link, you’ll want to ensure measurement has been deployed which picks up these campaign codes and places them in the corresponding Omniture variable (usually the campaign variable which appears as v0 in the query string). If you access the Omniture link above, you will notice the value ‘qr’ shows up in the campaign variable (v0).

To modify your tracking beacon and accomplish #2 with server-side code you could deploy something like the PHP code below (similar capabilities are available in other server-side languages). Often the business will have already standardized on a single campaign identifier and it’s a best practice to use the same identifier for mobile campaigns. The example below assumes ‘cmp’ is the query string parameter being used to uniquely identify the campaign.

Sample PHP


echo "<img src=\"".$campaign."\" width=\"5\" height=\"5\" border=\"0\" alt=\"\" />";

Sample PHP Using New Measurement Library


echo $s->track();

Adding measurement for SEO efforts is slightly different since you will need to pass along the page’s referrer. If your deployment will be using the server-based measurement class, collection of the referrer will be done automatically. If not, here’s how to pass along the referrer as part of your measurement beacon.

Sample PHP


echo "<img src=\"".$page_referrer."\" width=\"5\" height=\"5\" border=\"0\" alt=\"\" />";

Mobile Video
A recent market study by Transpera and InsightExpress showed users of mobile video exhibit behaviors that make them a key demographic for advertisers. The study uncovered other interesting tidbits which show the value of this uniquely engaged demographic. Regardless of the industry, if your business is providing mobile video, there is a significant opportunity to maximize the impact of video on KPIs (i.e. segment mobile video users, optimize video usage, improve ad sales positioning, etc.).

Regardless of the desired optimization, the first step is to ensure usage of mobile videos is being measured. While several viable approaches exist for capturing mobile video usage depending on device capabilities, delivery format, and video access method, we’ll just touch on the most common use case—accessing video through a link on a mobile-optimized site. Most of the technical steps to deploy video measurement in this scenario have already been covered in the prior post on Measuring mobile user actions in SiteCatalyst.

While you could just use the method covered in the post to increment desired custom variables and the custom link report, why not use some of the out-of-the-box video reporting capabilities? To leverage the standard Omniture video reports, you need additional video parameters not covered in the prior post.

To customize the link redirection method, you’ll set the page event type (pe) to indicate the type of user interaction being measured is media (pe=media) and you’ll set the media parameters with their corresponding attributes of the video being viewed. (mediaName, mediaLength, mediaPlayer, mediaSession). In this case, we don’t know what the user’s session will ultimately look like so we set mediaSession to 0-0.

Here’s a working example:

Original site link: <a href=”″>SiteCatalyst Product Tour</a>

Rewritten link: <a href=”″>SiteCatalyst Product Tour</a>

The usage of the dynamic variable flag (D=) in the example above will result in the video url being used as the name in the video reports. To customize the reports to show a more friendly value, simply replace the value in mediaName with the desired name.

Note: Some of the Omniture video functionality will not be available when using this method. For example, since the actual seconds of the video consumed are not available at the time the link is clicked, the following reports (video segments viewed, time spent on video) and the following metrics (video views X-100, Average View Time) are not available.

Get Code & Docs
We’ve covered basic analytics implementation steps for mobile optimized sites, mobile apps, mobile campaigns, and mobile video. However, as you begin implementing the various solutions covered, I recommend accessing technical documents and code for each measurement methodology you plan on deploying. The technical docs mentioned in this article can be found by logging in to SiteCatalyst and browsing to Help > Supporting Docs > Manuals; and code libraries for mobile optimized sites and mobile apps can be found by logging in to SiteCatalyst and browsing to Admin > Admin Console > Code Manager.


Thank you, that does explain how this is possible and the benefits of using QR barcode for more details campaign tracking. The question is how many people are a going to utilise this and does offering bonus incentives such as mobile content help increase those using the QR barcode and how does this affect the campaign conversion rates. Watching interactions with people they still seem to be using mobile internet devices such as iphones/palm sitting next to their laptop or computer. This could spike campaign results as they are not technically the same visitors so its very hard to filter out.