• by Steve Smith , Staff Writer @popeyesm, October 31, 2008

The challenges of tracking, tagging and targeting mobile users is legendary. As Ed Moore, Product Marketing Strategist for Openwave, explains, the technical differences between standard Web-based browsing and mobile access to data are substantial. Openwave has been a longtime supplier of infrastructure solutions to major carriers, which gives the company a unique ISP-level view of all the usage patterns on a network. Its Mobile Analytics package gives carriers a range of tools for segmenting and targeting audiences for advertisers. This month, Openwave released a "Behavioral Targeting" module that adds familiar BT tracking and segmentation to the mix. While many wireless operators are still trying to master some of the basics of demographic segmentation in mobile advertising, tools like Openwave's are luring them to the next stage in mobile.   

Behavioral Insider: As mobile users embrace the Web on small devices, what patterns of use are you seeing?

Ed Moore: The way that people browse on mobile is very aligned with how people access the Web on the PC. The concept people are rapidly moving towards is, they just want to have a device and use the service they are familiar with, the destinations they are happy using. This is one of the key drivers. So analytics in that vein performs a similar function -- whether it is for the content owner, a broadband network when you are tracking and targeting or behavioral targeting, or analytics in general. What we find in mobile operators, though is that there are still limitations or special subtleties around how mobile traffic flows. Analytic data can be gathered in different ways. The techniques currently used on broadband networks and PCs aren't necessarily directly equivalent in the mobile world.

BI: Explain the technology differences between standard and mobile Web. What tracking methods are not possible here?

Moore: A good example is tracking with things like Google Analytics. The standard way to do that is drop a javascript in a Web page. When that gets loaded onto a PC browser the script executes and reports back the analytics to the server: who has accessed, when, and how, etc. In the mobile world there is no concept of javascripts executing on the fly. Those techniques don't function, and it is why in some cases you see analytics that are skewed to the very small number of devices on the mobile Internet that do support javascript. So it is not the case that browsing is not happening from a wider group of devices. It is just not being reported in the same way

BI: What kinds of misreporting are you seeing in the market as a result?

Moore: An iPhone gives back information, because it can run a javascript. You will see statistics of the overall iPhone traffic on some analytics platforms as a much higher percent than they actually are because they can report and other handsets can't.

BI: How can mobile tracking technologies work around this limitation?

Moore: We find with people doing targeting from mobile analytics that they kept the same approach. You rely on the content supplier to put a tracking image of something on a content site. But it gives you a limited approach and it opens itself up to strategies operators can put in place to maximize data network efficiencies by blocking access to ad servers or ignore images below a certain size. But you can only get analytics information across your own site. Openwave's infrastructures are deployed within the operators' network, so we handle every transaction going over Web sites from every user. Therefore we can correlate analytics data across any number of publishing sites and any number of different user behaviors without requiring the addition of extra images or javascript.

BI: What key metrics are carriers looking for?

Moore: What is of interest is breakdowns of customer usage by day, week, nation, URLs, top sites, by age. For operators we can look at the efficiency and throughout going to the various advertising networks. An ad network can track its own usage -- but an operator finds it interesting to see which ad networks are being the most efficient or generating the most revenue.

BI: How does you behavioral module work in relation to the typical behavioral targeting engine we know online?

 Moore: Our view of targeting is exactly the same. Our targeting module is functionality that sits on our core analytics platform. We have access to all the transactional records for users coming through that operator's network and can analyze the behavior and usage and see which destination sites or which pages particular people have been viewing. From that we can categorize the user and build a picture of their behavior over time and then segment them any way you like. Our targeting modules, when they are deployed into an operator's network are then available to augment requests for ads.

BI: So are you indexing context down to the page level of mobile content?

Moore: Depending on the complexity and popularity of the site we can categorize all the way down to specific pages. One of the destinations where people go is the operator's portal, so rather than say [this user] is a portal user we would break it down and see which areas they went to. For the large popular sites it makes sense for us to see if they went to sports or home entertainment, etc. We break it down to that level. Our infrastructure and gateway platform handle transactions, whether it is on deck or off deck, and we combine all of those pieces of activity into a single set of segments.

BI: Is there a group of segments you use? In online BT an ongoing argument seems to be the number and kind of necessary segments.

Moore: We do --  but from my personal view, I would like to see the industry as a whole come up with segments that everyone understands. For us, [BT] is not our core system, so it makes sense if we are helping the operator to encourage as much standardization as possible. Because the mobile operators do cover a lot of the market with a small number of companies, there does seem to be some activity in industry groups to try to do standardization.  I think that would benefit everybody, whether media buyers planning campaigns or just looking at reporting information. It would be helpful to get a picture over a couple of operators all in the same way.

BI: Are the carriers ready for BT?

Moore: No. It is really early stage as far as the operator is concerned. They are still looking at the mobile ad market and wondering where it might go. They know there is money to be made. But it is early stages and obviously targeting is a means to improve the returns on advertising that you are serving. But there is interest growing in that whole area. The other thing which we are beginning to see is third parties such as the ad companies, content owners, who would like to see targeting information themselves and see what the operator has. There are others beginning to ask, ‘Can we get this?'  Six months ago there wasn't much.

BI: What level of ad targeting are they interested in, broad demographics?

Moore: Yes. But they know the benefits of targeting and it is something they want to do. There is a certain amount of education. One ad network I spoke with recently sheepishly admitted they couldn't properly follow the number of click-throughs. They knew what they were serving up but they couldn't get an accurate count from the company as to what was clicked for. We asked that question of our analytics group. They picked one operators and one ad network and took the data for the month we had. They looked at how many ads the network was serving, how many were clicked on. They had data on revenue rates, and that gave us a figure on how much that network should be making. We have more data just from serving the traffic than the networks who serve the ads.