(Part IX of a Series on Web Analytics for Search Engine Marketing)
I’ve come at last to the conclusion of this series on Web Analytics for SEM – which isn’t to say that I’ve exhausted the topic. There are so many different ways to measure SEM programs using web analytic approaches that it would take a book (god forbid) to really provide comprehensive coverage. I hope, though, that this series has been enough to provide compelling evidence that if you are relying on your Agency or your Bid Management Software for SEM Analytics you are missing the boat. The view that the SEM Buyer has (and will probably always remain focused on) is simply too narrow and misses too much that is essential to understanding the overall success of a SEM program.
Before I go on to other things (my plan is to tackle some CampaignTracker issues, some Content Match Issues and then do an extensive series on "Functionalism." This last is a new methodology that represents our current best thinking on how to do web analytics and I hope will be prove very interesting!), I wanted to step back and talk about something I’ve often noticed with web analytics and which I think is very evident in the web analytics for SEM series.
Consultants and experts in web analytics (and any other discipline) have a built-in stake in appearing to be much smarter and much more knowledgeable than everyone else. We play this game, like everyone else – because it’s expected and a necessary part of doing business. Indeed, I fully believe we are both smart and knowledgeable. But that being said, I’ve noticed that there is tendency to act as if the only way to get valuable web analytics is to do really complicated, really hard analysis.
I’ve written elsewhere about why I don’t think this is true (see, for instance, this blog which I also published as an article sometime back). Many of our most useful analytic projects have been strikingly simple. And as I review the various SEM Analytics I talked about, I think it’s interesting how easy most of them are.
Measuring Organic Cannibalization is trivial – it can be accomplished by simply going dark on key terms for a short period and measuring the before-and-after click volume sourced organically.
Measuring PPC Self-Cannibalization is slightly trickier - it’s supported directly by some tools (also depending on how you configure campaigns) but can also require some basic visitor segmentation. Still, it’s hardly rocket science and it can be trivially analyzed using Active Segments or Data Warehouse in the most common web analytic tools.
Setting up Conversion Proxies is usually just about as easy. If you need fancy proxies, then there may be some effort involved. But most conversion proxies are easily created with visitor segmentation tools and the analysis is hardly more complicated than a cross-tabulation of visitor counts by keyword (or Ad Group) landing and achieving the proxy.
Finding SEO Holes is even easier in most tools. You just need to check SE entries by engine by site area. This can sometimes require some sweat equity, but the level of thinking is minimal.
In my experience, the more complicated an analysis, the less likely it has been to provide genuine improvement to a business. It’s true that web analytics tools sometimes make measuring very simple things hard. But just because an analysis is hard doesn’t mean it’s especially likely to be valuable. The SEM Analytics techniques I’ve discussed are all readily understandable, readily measurable with almost any decent web measurement tool, and well within the range of most any competent analysts. That doesn’t make them less valuable or compelling – it makes them more so. Really, there is no good excuse not to be doing this level of measurement on your SEM program!
Happy 4th!
Comments