Blog Logs

Note 11/17/07: This is obsolete, but potentially interesting to someone. I’ve removed most links; some no longer exist, and the others didn’t seem worth fixing.

Some of the things I find in the server logs for a dabbler’s journal.

Web server log reading is not exactly a science.  My webspace provider furnishes me two views of my logs–one uses Webalizer and gives a simple summary of traffic, basically arranged by month; the other uses NetTracker and permits some fancy filtering.  Although they show generally similar information, in detail they can be quite different.  For instance:

  • According to NetTracker, saw 751 hits yesterday, and served 492 page views.
  • According to Webalizer, saw 600 hits yesterday, and served 333 files.

Like I say, not exactly a science; since they’re reporting from the same data source, I expected them to track a little more closely than that.  On the other hand, I’ve used several weblog readers over the years and know that there’s a lot of room for interpretation.  Although the logs record discreet events, it’s the patterns in those events that’s interesting.  Summarizing those patterns requires judgments about how things should be classified.  It ain’t just counting.  So when we see these disagreements we’re seeing different programmer’s interpretations of the data, and I’m sure both sets of assumptions can be defended.  Trusting that those choices have some basis in reality, consistency over time is probably more important than getting the details “just right.”  After all, I’m trying to understand why folks are reading me.  Both tools give me clues.

Followup:  For August 2, both reported 949 hits.  That’s better.

Who Reads Dabbler?

I’ve come to believe that the most interesting number NetTracker delivers about this site is “Total repeat visitors.”  The numbers there suggest that I’ve got about 25 “daily readers” (user agents, for the most part), and perhaps a hundred readers overall who check the site regularly.  This has been pretty stable since mid-May, when my readership suddenly doubled over a three-week period.

  • Some of the regular readers are looking for Midwest League news, because I am treated by many as an MWL information source.  I’d guess that’s half of the 25, and perhaps a third of the hundred.  (This shows because those journal entries generate click-thrus which show in the detailed counts.) 
  • The other obvious source of traffic on the site is Google.  I’ll discuss that below.
  • There are subscribers to the PictureShow and Bureaucratic Whimsy feeds, but I can’t tell much about what those readers are looking for from their website traffic patterns.  With the photo blog it’s largely a noise issue–the design of the site automatically drives the photo hit rate up, and some of the Google traffic is looking for pix.  I expect Whimsy to hit critical mass some day, but it’s not there yet.  Might be wishin’….
  • The most common reader is me.  If I’d stop tinkering with the design, the most common reader would be Bloglines, which presumably isn’t just me but certainly includes me.  There’s enough other traffic that I’m sure I’m not doing this only for myself.
  • At least one friend watches the site just because she’s my friend (thanks, Nan).  My sister checks in sometimes, too.

One last note:  Last month’s long vacation seemed not to affect the readership pattern.  I’m a little surprised.

What Pages are those Readers Reading?

Trivia: One of the most-served “pages” on the site is the ever-changing Daily Picture thumbnail.

The most popular pages on the site are the RSS feeds.  Here’s how they rank (all numbers which follow are from June 15–the day I rebuilt the RSS feeds–thru August 1 and exclude me when I could reasonably do so):

  • Excerpts (~1,700)
  • Bureaucratic Whimsy (~1,000)
  • Full Text (~1,000)
  • Midwestern Diamonds (~1,000)
  • PictureShow (~1,000)

Truth:  I didn’t expect these numbers to be so similar.  Note that for convenience I’m ignoring the “discontinued” feeds, which show a similar pattern.

Home:  The Front Page is the other “routine” way to check the site for changes.  That page has also had about 1,000 hits in this period.

After the feeds, there’s an enormous dropoff–which means that most of my page hits are from feed readers, and tell more about the tools than about my (human) readers.  The highest-count page which isn’t a feed or Home is the “Chrono” page (another index, of course [not currently extent]), with 88 hits in this time period.

The site has eight indexed categories.  Here are the hit counts for those indexes:

  • Midwestern Diamonds (84–now combined with CrankSpace)
  • Bureaucratic Whimsy (78–gone)
  • The CrankSpace (67–odd, there’s far too little content here to justify significant traffic)
  • War Stories (65)
  • Gadget Bag (27)
  • PictureShow (24)
  • Miscellania (22)
  • Commonplace (20–now combined with Miscellania)

Pages, jowo.  Tell me about pages.  These are the most popular pages, and some comments:

  • Gracenotes CDDB (80 hits–not sure what they’re looking for, but my page ranks second for this search term on Google–and first for Gracenotes iTunes)
  • Fifty and Five (75–mainly people checking me out; this page is linked as “about joel” from every page on the site)
  • Perfection (62–there’s some distortion in this count, as one person seems to have viewed it a bunch of times)
  • Home Run Derby (53–largely traffic from folks looking for information about the 2004 major league derby; probably attracted by the Sandburg/Dawson reference)
  • Eric Duncan (57–folks looking for information about the Yankees’ 2003 first-round draft pick.  Surprisingly few of these readers found that I also have a picture of Duncan on the site–so I added an explicit link.)
  • Yes, Drill Sergeant (57)

Second tier.

  • Heavy Weather (38–an obvious break in the data)
  • New River Bridge (34)
  • QC: Swing! (34)
  • Suite Judy Blue Eyes (33)
  • A Beautiful Day (31–my Eric Duncan picture)
  • Opening Day (26–the box score did it; most of these hits were from folks who were looking for a specific ballplayer)

Well, you get the idea.  There’s really no one page which attracts a lot of readers, and those which do attract people are not the (bureaucratic whimsy) folks I’m doing the most serious writing for.

For what it’s worth, a new page generally attracts five to fifteen hits within a few days; I take these to be mostly folks reading the excerpt feed and clicking through.  (See what I mean about interpretation?)  Page views after that are largely driven by Google.


Of the 6,600 or so page referrals to since June 15, only 925 show a referring website in the log–that is, I’ve had around 5,700 hits from bookmarks, user agents, and other non-reporting link sources.  Of the 925 hits which reported their referral page, 563 were referred by Google; the next highest total for a “search engine” is Yahoo at 64.  AOL Search generated 46 hits; all other sources of traffic are insignificant, except in aggregate.

What did they search for?

  • 2004 home run derby (24)
  • eric duncan (23)
  • gracenotes database (11)
  • qc swing (10)
  • qcswing (8)
  • drill sergeants (8)
  • manny trillo (8)
  • john o’donnell stadium (8)
  • lansing parade of homes (7–I know why they found me, but I’m not sure why they were looking, especially on July 22)
  • the remaining search strings were used six or fewer times.

The first keyword set which could be called a “Bureaucratic Whimsy” search is Frank Patrick (3).  No other search for a “Whimsy” page has more than one hit.

Oh, yes.  I got a link from Robert Scoble during this period (to my comment about Spolsky on Microsoft); it generated 15 hits.  Not exactly a Slashdot impact.

Final note:  How do folks expect to find anything useful when their search term is “passable”?  Or “frequent”?

Another 11/17/07 note: This entry drew a response from Effern/Ethan Johnson on his then-blog The Vision Thing (no longer extent in the same form), which provoked a followup from me. Effern’s entry is gone, and my response will likely not be reposted.

This entry was posted in Dabbler Notebook. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.