It’s been far too long since I’ve posted about web analytics. When I started this blog back in 2003 that subject was the initial focus and it led to some of my most popular posts such as the annual Complete Guide to Web Analytics.
I wish I would’ve kept that going, but unfortunately my focus changed and I wasn’t paying as much attention to the space so it didn’t feel right to do the guide again. However, throughout this time I’ve continued to consistently use the various public website tracking services such as Compete, Alexa, Quantcast, and Google AdPlanner to do research, check out competitors, and watch the sites that I’ve built over time.
I’ve generally been very frustrated with the services, and from everything I’ve ever read I’m not alone. Also according to Compete, they are all experiencing less usage over time as well.
Oh, except if you use Quantcast to check all three of them or Alexa to check all three of them you’ll mostly see trends that have them holding steady. Wait, actually on Quantcast you’ll see that Alexa is dropping. Except Google AdPlanner agrees with Compete that Compete and Quantcast are dropping badly, yet it thinks Alexa is on the rise? And on Alexa you’ll see that they don’t have enough traffic data to rank Alexa? That’s just laughable.
The takeaway is that not only are the numbers themselves not that accurate, even the trends vary wildly from service to service. Even though they all use slightly different ranking methods, it’s clear that they aren’t catching even the trends right.
This is tough though. It’s not like these services haven’t tried over the years to get accurate from using toolbars, to panels, and more. Out of the bunch, I like what Quantcast has tried to do in offering publishers the ability to put their tag on their site so that it can be directly measured by Quantcast. The benefits being that at least one service has your traffic being somewhat accurate, and that perhaps that will help with advertisers and provide demographic information.
Unfortunately that offer isn’t compelling enough to get a majority of publishers participating. In an unscientific browsing of their rankings of publishers, you can see that near the top of the rankings they get about 30-50% of publishers participating, but once you get down to the the ranking of this blog you get about 1-2% participating. There just isn’t much incentive for smaller publishers to participate.
Unless that incentive can be figured out, it seems like our best shot at getting a relatively accurate gauge would be for Google to do it. They have the largest number of publishers using their Google Analytics product, most of the largest publishers use DoubleClick for ad serving, and through Google search and other methods they probably have some sense of traffic and trending even if publishers aren’t using Google Analytics or DoubleClick.
I believe some of these things are data inputs for their Google AdPlanner tool which gives traffic estimates. However, even though I use Google Analytics and DoubleClick for ad serving on this blog, they don’t enough traffic data to rank ConversionRater.
Of course, there are data privacy issues which I believe keep them from being able to use Google Analytics or DoubleClick data directly without the publishers giving permission. But why not give publishers an option in Google Analytics to let their visitor/visits/pageviews be public if they want? Again though we have the incentive problem, why should publishers participate in making this data public?
I don’t know the answer today. If someone could figure this out though, they could create a very valuable service. Yes, I know Comscore exists as well, but since it costs significant money to access I’ve never been a big user. Plus, it has had no shortage of complaints and publishers are often disputing Comscore’s statistics with their own internal data.
All I know that is that it’s 2011, and it’s still really difficult to get not only accurate numbers for websites, but even seeing accurate trends is difficult. That’s just not right.