I recently read an email on a mailing list asking about "different ways of measuring blog readership/audience for a broad range of specific blogs". There are lots of different ways of measuring blog readership and audience, depending on what you want to measure and what sort of access you can get to the statistics.

There are various sites that gather data about websites, and the first few that were mentioned on the mailing lists were Quantcast, Compete and Google's Adplanner. They have different means of gather data and as a result different levels of accuracy.

Quantcast uses a pixel to gather data for participating sites and makes estimates for everyone else. For large sites and for participating smaller sites, I've always really liked Quantcasts reports. If you take a look at the Quantcast report for Orient Lodge you can find a lot out about my readership. They also provide very up to date data.

Compete uses panels to gather data and do not seem to be quite as reliable as Quantcast. They use tracking code to gather audience profile information. However, they are pretty expensive to get to the interesting data. Here is the Compete site analytics for Orient Lodge.

I haven't played with Google Adplanner much, but they tap into data gather from Google Ads. If you authorize it, they supplement the data with Google Analytics data. They provide information about other sites that people visit. Here is the Google AdPlanner data for Orient Lodge. I hope to explore the affinity calculations in a later blog post.

What was not mentioned in the list was Alexa. They've always seemed a bit random and while some people claim they are getting better, many people don't trust their data.

If you can get more direct access to a sites traffic data, either through Google Analytics or server statistics, you can get much more interesting information. What percentage of the traffic bounces, or visits one page and leaves without visiting other pages? For those that do stick around, how long do they stick around? Where are the readers coming from? What are they using for browsers? How did they find the site? Direct links? From where? Keyword searches? What keywords?

This leads to the next question about what you are trying to measure anyway. I've often suggested that for my site, I'm not concerned with bounces. I want people to find what they are looking for on the first page they visit. If I were running an online store, I would be more concerned about bounces. I hope that people spend time reading and thinking about what I write, so for people that don't bounce, I hope to have a high time on site.

Related to this, people on the mailing list suggested that other metrics, such as the amount of engagement is what really matters. How often do people comment, link to the site, retweet messages about an article, save a page in a shared bookmarking service?

RSS feed subscriptions were also mentioned as well as Feedburner and their email option. Personally, I haven't used my RSS reader accounts in ages, although I'm still subscribed to hundreds of blogs. Messages on Facebook and Twitter get a much higher priority for me.

So, why are we concerned about these metrics anyway? The biggest issue is probably advertising. Much of the focus has been on getting an increase in page views or impressions, so you can sell more impressions. This has raised a concern about journalists trying to write article that will get the most impressions. However, not all impressions are created equal. Writing more esoteric articles may result in fewer impressions that reach a much more desirable advertising demographic. Journalists writing simply to get the most impressions may end up doing themselves a disservice as more and more advertising inventory goes unbought or sold at remnant prices while high quality impressions from specific audiences become more valuable.

There are a lot of different tools for measuring readership, and the best answer to which is best is that it depends on what you're trying to measure, why you're trying to measure and what sort of access to data you can get.