Forum Moderators: martinibuster
Looking at the day-of-week statistic (which computes the share and totals of page impressions, ads, clicks, and revenue for each weekday), I see that Thursdays are very very slow, especially CTR and EPC are way below weeks average (can't go into details here, but the deviation is not in the typical 2-3% range which is acceptable). On Fridays things seem to go back to normal again, and CTR and EPC are in the normal range again.
Okay, I know that one day of the week has to be the "worst day", but my Thursdays are so far off that I think I am missing something here (not that I could do anything about it, though).
My only idea on the reasons for this is, that advertising budgets are running out on Wednesdays, then all of the sudden 'bad ads' (e.g. aff ads) are starting to show up (hence the lower CTR) - and these guys do not pay very much (hence the lower EPC). But I wonder why this just affects Thursdays...
Any insights or ideas why this may be the case?
What's you experience? What's your "worst day"?
-- Mark
It's a travel site, and I see the majority of traffic during weekdays (guess people are researching where there next business trip should go to), with Mon/Tue being best and Fri being worst of the working days (Thursdays come in as #4 though). Saturday is slowest in the week, Sundays are a bit better than Satursdays, but not as good as Mondays.
I am really interested to see your experience in AdSense clicks/revenues compared to week average.
-- M.
On my Gallery, traffic and AS revenue is usually lowest on a Sunday, which continues to astonish me as I expect my demographic to be home surfers but I must be getting a lot more "professional" users than I planned for.
I do get the occasional very very good weekend for no obvious reason.
Rgds
Damon
Total traffic:
Monday to Thursday 100% - Varies little
Friday 80-90%
Saturday & Sunday 50-60% depending on weather
The busiest day is usually Monday or Tuesday, though Wednesday grabs the top spot now and then.
Interestingly, Monday was always my busiest day a few years ago. I have an editorial travel-planning site, and I always figured that people were going back to work on Monday morning, looking around their cubicles, thinking "I gotta get outa here," and typing the name of an appealing vacation destination into Google as they sipped their first Starbuck's coffee of the week. The day-to-day variations have leveled out a bit in the last couple of years. Why? I'm guessing it's because more people have broadband, so they're using the Web for research at home to a greater degree than they were in the past.
Another interesting factoid: Although I see a fairly big drop in traffic on Fridays and Saturdays, earnings from affiliate sales and AdSense don't drop nearly as much. That's pretty easy to explain: Armchair travelers and people with vague, off-in-the-future travel plans may not bother to surf for tourist information on the weekend, but a person with a vacation coming up will take time off from other leisure activities for trip planning.
Donster
I think what you have noticed in timing of traffic peaks may be a more general trend. I'd say that the peak has moved from Mon/Tue to Tue/Wed over the last 10 years or so (back to my 9600bps analogue leased line and even UUCP rather than IP a bit futher back than that...).
But the evening peak is still around the confluence of EU hometime and US lunchtime, which suggests to me that US workers still surf from work more than home and in the EU vice versa, but what do I know?
Rgds
Damon
But the evening peak is still around the confluence of EU hometime and US lunchtime, which suggests to me that US workers still surf from work more than home and in the EU vice versa, but what do I know?
It would be interesting (if only for fun) to see a chart of which countries users are coming from at any given hour of the day. One thing that fascinates me is how global a Web audience is, compared to traditional media. For example, one of my hotel affiliate partners now shows which countries bookings are coming from, and I see bookings from places like Moldova and Nepal in my affiliate reports. It's like being back in high school, when I was an amateur radio operator and it was exciting to exchange transmissions with somebody on the other side of the world. (Hmmm....Maybe I should ask my affiliate hotel partner to issue a QSL card for every new country in my bookings report!)
Of course, my main site (75% of my total traffic) is a Christian tablature site. On Saturday people are starting to print out songs for worship on Sunday, on Sunday people are printing like mad, and I guess on Monday a lot of people are trying to find chords for the songs that they just heard on Sunday. :)
Mickey
Saturdays are the slowest days, but very frequently, perhaps as much as 50% of the time, Thursday, for no apparent reason, goes into the dumper.
I have a few niche sites.
Traffic patterns are different that this though.
Monday is the busiest day, falling by about 5% per day through Saturday (so down about 25% on Saturday compared to Monday). Sunday traffic is up about 5% from Saturdy.
It's lunchtime, so I did some very simple stats looking at my logs for the last few days. Here are clickthroughs from US(-ish) IP addresses by hour of day (GMT). These numbers are NOT absolute, but from a representative sample.
The first column is clicks, the second is GMT hour of day:
35 00
15 01
13 02
17 03
13 04
08 05
11 06
07 07
06 08
07 09
03 10
04 11
11 12
15 13
14 14
12 15
18 16
24 17
20 18
19 19
13 20
19 21
27 22
17 23
So there *is* a spike around US lunchtimes in this data, though not as spectacular as the midnight spike (which *might* be a mistake in my processing).
Rgds
Damon
WARNING: Statisticians may choke on small samples.
And here are some matching stats for UK-ish IP clickthroughs...
Left column is number of clicks (NOT absolute number); right column is hour of day (GMT).
01 01
04 02
00 03
03 04
00 05
00 06
00 07
02 08
07 09
08 10
13 11
12 12
06 13
07 14
09 15
16 16
12 17
03 18
04 19
11 20
02 21
04 22
01 23
This does look like it supports my "bored around hometime" theory, but possibly split into "just before I go home" and "after I've eaten" sub-peaks these days, as well as a lunchtime spike.
Each sold separately. Available where sold. Batteries not included. All liability disclaimed, including but not limited to: inability to use egrep awk sed and sort correctly, inability to count beyond 10 even with socks off, having a site with different demographics to yours, not being as clever as you, not having expensive lawyers.
Rgds
DHD