July 24, 2007

MineThatData Traffic Analysis

Click on the image to enlarge it.

You're probably wondering how many folks read this blog, right?

Or, maybe the topic is only of interest to me.

Either way, you get to learn all about blog audience dynamics today.

According to my statistics, I have about 3,900 visitors per month (via Google Analytics) and an average of 450 daily subscribers (via Feedburner).

The image illustrates how we get to subscribers, and the true amount of monthly traffic generated by the blog.

During an average month, 1,873 visitors arrive via search (almost all via Google). 976 visitors arrive from other websites, about eighty percent from blogs referring folks to MineThatData (thank you all!). 200 visitors type in the URL.

Of these visits (3,049), about two percent (60) become what I call "Loyalists", folks who frequently visit the URL, or subscribe by RSS or E-Mail.

I estimate that each month, I add 60 subscribers, and lose 30 subscribers, yielding a net increase of 30 subscribers. To date, about 450 of you subscribe via RSS, E-Mail, or type in the URL frequently. You are my "Loyalists". Thank you!!!!

The 450 "Loyalists" drive the relative popularity of this site, visiting about 13,650 times a month. There are 854 actual site visits, 5,460 readings in Google Reader, 1,092 via E-Mail, and 7,098 readings in other RSS Readers.

In total, this yields 14,504 actual readings from my "Loyalists", and 3,019 visits from folks who weren't impressed with the site, or got what they needed from one visit and moved on.

In total, this yields 17,523 real "visits" each month.

There are flaws with the analysis --- there are probably many more "Loyalists" than measured by Feedburner --- BlogJuice suggests I have twice as many Bloglines subscribers as Feedburner suggests I have. So there might be as many as 1,000 actual "Loyalists", each visiting half as often as I have estimated.

Still, the number of real "visits" is a reasonably accurate reflection of what happens each month ... yielding about 210,000 annual visits.

2 comments:

  1. Anonymous4:12 PM

    Kevin,
    Might I state something that is overlooked?
    Same user, different access points. I would venture to guess that your "loyalists" may be logging in from minimum 2 locations(work and home). Others might even travel and pick you up "on the road".
    Also, how is IP determined if visitors come from different hotspots(starbucks or an airport)?
    That is when I wonder how well we can really track user trends/patterns without more invasive means besides cookies--i.e. spyware.
    How did you account for this in your previous experience?
    K

    ReplyDelete
  2. We'd have to ask Feedburner how they deal with the issues you discuss. I watch my Feedburner log files --- and I can see where they are de-duping various RSS requests.

    At Nordstrom, we did have to deal with the challenges you outline.

    We linked multiple e-mail addresses to cookies across multiple computers via a many-to-many relationship.

    Then, we linked e-mail/cookies to individuals at a specific name/address (also a many-to-many relationship).

    Then, we linked individuals to households (many-to-one relationship).

    Then, we analyzed the business, for the most part, at a household level.

    It may not have been the best merge logic available, but it was better than 90% of the things our competitors were doing.

    ReplyDelete