13 thoughts on “Daily Dose Daily Visitors”

1. Hi Dick,

   That’s the same pattern I’ve observed on my website for several years. My conclusion about it goes something like this:

   Friday’s number includes data from some geographic regions where it is already Saturday. Hence, it’s lower than for other weekdays.

   Sunday’s number is higher than Saturday’s because it includes data from some geographic regions where it is already Monday.

   Monday’s number is lower than the other days because people are catching up on their backlog from over the weekend. Globalization means someone somewhere is generating work for you even when you are away from work.

   Well, either that or it is the only day of the week when people actually work.

2. Excellent analysis, Tushar. I’ll buy that.

3. Insightful analysis, but I think the effect of other time zones is not so important. If I look at time-of-day stats, I see a peak in midmorning (east coast US time), a drop at noon to 2 pm, and another peak around 5 pm. The peaks are somewhat diffuse, so they may be the superposition of four sets of time zone data across the US, each with peaks at midmorning and midafternoon. There’s a smaller peak at around 5 am EST, which could be the midmorning peak in Europe; Europe’s midafternoon peak would be hidden by the North American morning surge.

   The overnight levels drop enough that I don’t think Friday is diminished much by it being Saturday somewhere else. I think Friday is more of a kick-back day, so less work gets done. Also, if a company has flex time, Friday is the usual day they stay closed.

   I agree that Monday is when people are catching up with their work, probably with what they blew off on Friday.

   Why is Sunday higher than Saturday? Again, I doubt it’s the early Monday Asian traffic. Maybe if you bring home stuff from work, you’re putting it off until Sunday, and that’s when you’re Googling for chart help.

4. I don’t recall where I just read this, but it was a study of similar data. The insight was that Mondays and Fridays are often not workdays. They are the usual day for US holidays that “float” — Veterans Day, Memorial Day, Presidents Day, MLK Birthday, Columbus Day etc.

   The study adjusted when adjusted for the “first operational day” vice day of the week, evened things out.

   Some Excel wiz can probably adjust the data for the occurrences of Holiday Mondays or Fridays ;-)

   …Michael

5. Michael: I imagine other researchers have found the same thing, but several years ago I surprised some college professors with the idea of “first working day of the week” in the context of a regression analysis for a consulting project. Of course, since at that time they were professors and I but a lowly doctoral student no way would they admit I thought of something they hadn’t. {grin}

   One place I mention the idea is
   Trendline Coefficients and Regression Analysis
   http://www.tushar-mehta.com/excel/tips/trendline_coefficients.htm

6. This might be off the wall but take into consideration sports, a lot of your collegiate games are played on Saturday. I know that football is played on Sunday’s, I bet if you would look at your logs or re-do your hit counter in the fall during NFL season you would see Sunday go down some.

   I do not have any statistics to validate this but it could be one of the reasons for the difference on Saturday and Sunday (I know my youngest does not like to leave the house on Saturday so he can watch all the basketball games).

   Signed by a true Steelers fan

   Fish

7. Tushar –

   And that was exactly where I read it! I had a need to get at trendline coefficients Friday and was on that very page stealing code ;-)

   Thank you for knowing where I’d been better than I could recall. Google couldn’t do that! I know, I was trying to find the reference …

   …Michael

8. I’m no guru — just a spreadsheeter in a finance office. In general Mon & Fri are busy days here crunching numbers and preparing reports — too busy to look for web solutions. Exception: here I am on Monday AM waiting for the DB to be updated & reports to run. I like the first workday theory, but wonder about slow web days because busy folks are crunching the new data & running routine reports. Also –what’s the ususal cycle for us crunchers? — I have both a weekly cycle and a monthly one — if monthly updates are most common, then look at the data by weeks in the month?

9. The reason Mondays are lower is that people come in to work on Sunday so that they can hit the deck running. Then they don’t need to look at blogs (nor do they have time to) on Mondays.

10. I think it’s what Sue says, Monday and Friday are busy not looking for web solutions. The Monday holidays would result in a lower average on Mondays, but I find the Monday and Friday pattern occurs every week, not just when averaged over several months with a handful of holidays.

    I expected a Saturday-Sunday thing where guys watch the NFL and that drives down Sunday site stats, but Sundays are ahead of Saturdays, and it seems pretty consistent week by week throughout the year.

    I suppose I could do a more detailed regression analysis, but it’s more fun to sit here and speculate idly/wildly.

11. Those numbers make me feel less special. I don’t feel like a member of an elite club any more when I have to share the site with 150,000+ other people. :(

12. Why Sundays? I guess people just can’t wait to get to grips with work on Monday again.

13. Tuesday through Thursday are the really productive days for work.
    Friday people take off, or are already on their way to taking off for the day… think? What’s that?
    Saturday you get the folks that remember something they should’ve checked on Friday.
    Sunday – noone wants to think about work on Sunday!
    Monday – recovering from the weekend. TEGO: the eyes glaze over. Don’t ask me to think on a Monday.

---

Posting code? Use <pre> tags for VBA and <code> tags for inline.