Note: The numbers below are dated. Late 2011 number have roughly 3000 visitors per day and the same level of engagement. Visitors spend a combined 13000 minutes/day on the blog.


I do not find normal blog rankings based on popularity or traffic like Wise Bread’s list very useful. Ranking by popularity or traffic is like ranking companies based on stock price and trading volume. You can generally see a strong correlation between a sustained professional writing effort combined with targeting the largest consumer segment and being (randomly) discovered by an established source early on.

In investing ratings are a lot more interesting that absolute numbers as they provide a means of scaling the numbers relative to the size. Nobody in their right mind would think that a company trading at $100/share would be worth twice as much as a company priced at $50/share. However, by dividing with the earnings, say the former has $10 in earnings and the latter has $5 we find that both have a P/E of 10 and thus they are similar in terms of earnings efficiency. In investing you can build many more ratings than that to try to tell companies apart. You will find that companies in the same industry tend to have the same P/E values. If not, something interesting is going on and you want to look closer.

However, nobody seems to be doing it for blogs. Yet I’m sure advertisers would like to know some of these numbers, so here’s a business idea for some enterprising entrepreneur.

  • Total eyeball time per day. To calculate, we use sitemeter’s summary page. Today, I have 682 visitors on average per day and they stick around for 4:26 minutes (266 seconds). I calculate 682*266/3600 (3600 seconds in an hour) to get 50.4 hours per day. This is how much time people have spent reading this site each day on average for the past week. I think this is a better measure for total impact than unique visitors.
  • Page views per visit. This shows how sticky your blog is. Do people stay around after coming to your site. If the number is high, it could indicate that your content is interesting and that you have dedicated readers. It could also indicate a growing blog, since new readers tend to spend more time reading old posts. Or it could indicate that readers visit rarely and then have to read backposts for the past few days. If the number is low, your core readers are either dominated by google searches or they simply don’t stick around, which of course could be because they visit often instead like on heavily crosslinked blog networks. Regardless, this number does say a lot about reader behavior. My number is 2.5 views/visit.
  • Visitor growth. Growth can either be organic or explosive if you get on digg or get a mention in a major newspaper. Here I tend to look at the histogram for visits and pageviews for the past 12 months. For my blog, you will see a spike for Sep08. That is when I got mentioned on MSNBC’s blog, twice. However, it seems few of them came back and the other months show a steady progression. To calculate the growth, you take the geometric average. I’m lazy, so I just take the first month and the last and calculate monthly growth as the twelfth root of 19996/10022 minus 1, that is, exp(log(19996/10022)/12)-1, so 5.9% growth/month. You can use the law of 72 on that. I have a doubling time of about a year.
  • Feedburner numbers?. I do not find feedburner numbers that enlightening, primarily because I do not believe the numbers from the google reader component. As far as I understand google reader numbers are cumulative. Once someone subscribes via google that count never goes down. You have the person counted for life even if the person no longer reads along. You can get some insight in the impact of this, because occasionally feedburner fails to count google readers. You will then see the feedburner numbers drop substantially, like 30-50% for a few days. This is a better reflection of the true number of subscribers.
  • Comments/post. I think this is a reasonably good way of gauging active reader “reach” or the intensity of your community relative to your writing. In other words: Are readers getting it? I get my numbers off of my dashboard. I have written 494 posts and they have gotten 4058 comments. Hence, I get 4058/494 = 8.2 comments per post.
  • Comments/subscribers. This better gauges the form of your fan base. Do readers actively engage with your content or do they just click, read half of it, and then go away again. I use the 4058 comments from above and divide with my feedburner number of 1369, so I get 3.0 comments/subscriber.

Keep in mind that ratio building is not an precise science at all. You are not interested in knowing whether 2.9 is higher than 2.7. These are effectively identical. However, 7.4 is an outlier compared to 2.9. That is interesting.

For those bloggers who are reading this, let me know your stats either in a comment or by sending them to jacob@early… and I will put them in a list and link to you.


I have not found much enthusiasm amongst my fellow bloggers for making such a list :-( , but it looks like I’m not the only one wanting to build alternative measurements. @kdpaine sent me this.
Take a look. It’s pretty cool!