Forum Moderators: martinibuster
This is an ecommerce site with approximately 275 Vis / day.
I made the common mistake:<WRONG>Put more ads == make more money!</WRONG>
I then streamelined my site to one leaderboard and one click unit with background and border to match my site.
There was still something missing, I wanted to tell which time of day would get me the best ctr and ecpm. Especially since adword advertisers are able select which time of day their ads run.
Here's what I did.
Create 24 channels named 0 to 23 one for each test hour
make a directory on my site
in that directory,
create one file
0.php with the leaderboard code for hour 0
1.php with the leaderboard code for hour 1
2.php with the leaderboard code for hour 2
....
create a file called time_slice.php
<?php
$tim=time()-(3*3600)+5*60*60*24; // adjust for your time zone and day of the week
$ss=$tim%60; // variable for seconds
$tim/=60;
$mm=$tim%60; // variable for minutes
$tim/=60;
$hh=$tim%24; // variable for hours
$tim/=24;
$dd=$tim%7; // variable for day of the week
$slicefile=$hh . ".php"; // set filename to hh.php
include($slicefile);
?>
Simply include the file time_slice.php to invoke the leaderboards.
-------------------------------------------
So far, It appears that better ads are appearing in the leaderboard and creating a larger ctr and larger ecpm in my URL channels which have still active. I saw somewhere that somebody reset his channels every time he sees fit. This is effectively what I did.
The test is not 24 hours old so I will post results in a week.
Thanks to Ann for awakening me to Smart pricing and to all who contribute to this forum
[edited by: martinibuster at 5:08 am (utc) on Aug. 7, 2006]
===============
That was me. I just do it manually when the ads stop generating enough income.
Mike
The object of this experiment was to find out if I had time of day which didn't click through. So far on my site I'm finding that 3:00 am to 9:00 am local time ( atlantic time) are not very productive - at all. During the "dead times", I could run different adds if I wanted to. Even self promotion of the web site would be an added bonus or affiliate program.
I created a channel for every hour of day.
0, 1 , 2, 3, ... 23
made Got all the code for each hour and pasted it in it's own file
0.php
1.php
2.php
3.php
...
22.php
23.php
The code checks the time, isolates the hour variable and includes the appropriate channel which is effectively a time channel. All this without an "if" statement or "switch" statement
Over time, though, it may average out.
I'm certain that my window will vary with the day of the week also. It can be compounded with a particular day of the week which we already have data for.
It is my hypothesis that during a certain time of day, people are more prone to surfing thus clicking more on ads.
Alain Pelletier
How often do you reset your channels?
what do you do with the existing channels? Do you de-activate them or just remove them?
Great idea for an experiment - thanks for sharing it.
TJ
This means that I could increase ecpm by 30% by not showing ads during that time. It wouldn't necessarely change the bottom line though.
Alain Pelletier
How often do you reset your channels?
what do you do with the existing channels? Do you de-activate them or just remove them?
=============================================
I just change the AdSense code to one that has NO channel. Everything else just stays the same at Google, I don't turn anything off or delete anything as far as they are concerned. I just run an AdSense block for 24 hrs. that has no channel hooked to it. For whatever reason that seems to reset their system so that when I put back the regular AdSense block of code it runs better paying ads.
After a week of monitoring, my test was a success.
I have determined the time of day which ads are showing with no or little ctr.
It is no surprise that this would occur for me in the dead of night. I was able compare ctr with what my visitors look for (I also log referals in a database and monitor daily what my visitors come through SE, link etc.)
The time of least CTR activity for me was between 2:00 eastern and 9:00 eastern. I guess people at that time don't feel like surfing the net but rather look through SE to get what they need.
By Wednesday (4 days of experiment) it was very apparent that a pattern was immerging.
-------------------------------------------
Please remember that this is an ecommerce site and your site may have significantly different patterns.
-------------------------------------------
The aim was to increase ecpm and CTR without compromising revenue.
This was achieved by changing the distribution of the AS adds. Instead of having 100% AS, I varied the distribution with by using different weights.
during my strongest AS CTR, I displayed the adds 100% of the time.
during my weakest AS CTR, I displayed the adds 50% of the time while the other 50% I displayed self promotion banners.
The other times, I checked the revenu over the amount of days (in this case 5 days) and varied the distribution weight as I saw fit... 75% AS/25% in house - 90% AS / 10% in house... etc.
here is the random distribution code in PHP.
<?php
$maxrnd=2; // set Adsense distribution ration: value of 2 is 2:3 Adsense ratio - Value of 9 is a 9:10 Adsense distribution
if (!rand(0, $maxrnd))
{
include("banneradsnew.php");
}
else
{
?>
<script type="text/javascript"><!--
google_ad_client = "pub-12345687"; // real number surpressed
google_ad_width = 728;
google_ad_height = 90;
google_ad_format = "728x90_as";
google_ad_type = "text_image";
google_ad_channel ="12345678"; // real number surpressed
google_color_border = "CED7E7";
google_color_bg = "CED7E7";
google_color_link = "0000FF";
google_color_text = "000000";
google_color_url = "008000";
//--></script>
<script type="text/javascript"
src="http://pagead2.googlesyndication.com/pagead/show_ads.js">
</script>
<?php }?>
It is really nice as well for you to include the code to make it easier. It's helpful webmasters like you that make WW one of my first stops when I get online. Thanks again!
