Adult Video Clips - Adult Video Search

Sponsor Test

Adult Video Clips - Adult Video Search

Purpose:
To test the hypothesis that sponsors are counting referral hits correctly.

Procedure:
Create a java script program that will open four browser windows at the same time. The program will open three sponsors and a counter. Using the counter as a control, data will be collected, and compared to the control.

Observations:

Date Sponsor A Sponsor B Sponsor C Counter
07/09/98 25 5 13 24
07/10/98 30 7 10 28
07/11/98 29 6 15 18
07/12/98 35 8 7 28
07/13/98 35 8 4 9
07/14/98 26 11 13 11
07/15/98 36 6 15 21
07/16/98 42 13 11 26
07/17/98 21 6 8 7
07/18/98 30 12 11 22
07/19/98 24 11 8 10
07/20/98 24 8 5 16
Totals: 357 101 120 220
Variance from control: +137 -119 -100 0
% Variance 162% 46% 55% 100%



Intermediate Conclusions:
Day 1 (7/9/98)
Total hits on the control indicate that a very small percentage of daily visitors participated in the test.
Hits on "Sponsor A" validate the control, and suggest an honest sponsor.
Hits on "Sponsor B" were rediculously low, indicating either different recognition of what is considered a hit, server error, or possible fraudulent counting of hits. More testing is necessary.
Hits on "Sponsor C" although low, do not vary considerably to warrant suspicion of fraudulent counting. More testing is needed.

Day 2 (7/10/98)
Once again, total hits on the control indicate that a very small percentage of daily visitors participated in the test.
Hits on "Sponsor A" continue to validate the control, and suggest an honest sponsor.
Hits on "Sponsor B" still rediculously low.
Hits on "Sponsor C" are also low.

Final Conclusions:
Hits on "Sponsor A" were extremely high, indicating that there were other factors contributing to the hit count. Further research revealed that the sponsor used also had a banner displayed on a previous page. This made the results invalid, and should therefore be disregarded.

Hits on "Sponsors B & C", although varying from day to day, are statisticly identical, revealing an average % variance of 51%. These results indicate to me either that the control doubled the count, or the sponsors discount referred hits by ~ 50%. Personally, I think its the latter.

Based on these results, I am forced to reject the null hypothesis. It is therefore concluded that the sponsors tested are not counting referral hits correctly.

WEBMASTERS are you interested in the names of the sponsors, send me some mail.

If you didn't participate, don't forget to join in.


Adult Video Clips - Adult Video Search





Adult Video Clips - Adult Video Search