Thursday, April 05, 2007

Channel-Specific Analytics Are Doomed: Doomed, I Tell You

Did you ever have one of those crazy dreams, not quite a nightmare, where unrelated things get mixed up together? I felt that way this morning when I was looking at the Web site for one of the mobile marketing systems and saw they had alliances with Web analytics vendors. That rang a bell, but it took a while for me to realize that I had been writing about consolidation in the Web marketing space separately from mobile marketing.

The confusion is compounded by my recent look at non-Web analytics system including ClickFox (which gathers interaction logs from call centers and other systems) and Skytide (which gathers all kinds of data; I haven’t written about it yet).

There’s an obvious connection between systems that gather interaction data and those that manage marketing messages. As the Omniture / TouchClarity hookup I mentioned yesterday illustrates, some of the vendors are themselves bringing the two together. It’s no surprise that this would happen for Web systems, which tend to be internally integrated but isolated from other media.

Of course, the Web should not be isolated, and the trend is in fact towards cross-channel integration. Does it make sense, then, for Web analytics vendors to integrate tightly with Web targeting systems? You can see why an analytics vendor would want to do it—as a revenue-generating line extension and a way to help clients who lack an existing targeting solution. But the vendors (and I’m sure Omniture recognizes this) must also make it easy to integrate their systems with any other targeting product. Otherwise, they risk losing sales to prospects who already have a targeting solution and don’t want to change it.

From a broader perspective, though, interaction data from many channels needs to be combined for marketers to do the best job of analysis and targeting. This can be done by physically copying the data into a traditional data warehouse or by using some sort of virtual or federated structure. What’s important is that data from many sources must come together into a single location, where it becomes accessible to many execution systems. In other words—am I beating a dead horse here? —you don’t want direct connections between single-channel source and execution systems, such as Web analytics to Web targeting.

This has technical implications. In the cross-channel scheme, the role of the analytics system is just to gather and reformat data so it can be presented to the central storage facility. The actual analysis would be done in the central system or by a cross-channel analysis system that draws from it. This means that products which combine data gathering and analysis, like current Web analytics systems, need to decouple those functions and build open interfaces to reconnect them. These interfaces would allow users to substitute other products on either side of the relationship. In addition, vendors with specialized data storage technologies might offer a storage component with interfaces at both ends, one to accept feeds from multiple data-gathering systems and the other to allow access by multiple analysis and targeting tools.

This is not an appealing proposition for many vendors. Breaking their systems into components opens them up to more competitors and risks each component appearing to be a commodity. It also eases switching costs, placing further pressure on prices. In general, as I’ve noted many times, vendors seek to expand their footprint and increase integration, not the other way around.

But vendors who specialize in systems for one channel will increasingly find themselves frozen out of multi-channel opportunities. There are already many products to provide multi-channel data store, analysis and targeting. Data gathering still tends to be channel-specific, but that won’t last as new channels become better understood.

In short, vendors who seek to remain channel specialists are likely to find their business shrinking over time. This may seem like bad news, but the sooner they begin to adjust to it, the better off they’ll ultimately be.

1 comment:

James Taylor said...

I completely agree - unless you can understand the total picture of your customers and then act on it across channels, you are going to be left behind. You clearly need to be able to customize and tailor for specific channels but the data on behavior must be used to build a single, integrated picture.
More on decisioning in marketing