Forum Moderators: martinibuster

Message Too Old, No Replies

How many impressions is a good sample?

How many impressions is a good sample to determine effectiveness?

         

ken_b

6:44 pm on Feb 14, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



How many impressions is a good sample to determine the effectiveness of adsense layouts and color options?

I recently changed the adsense color/border scheme on my pages and am wondering how many impressions is a reasonable number to determine if such changes are postive or not.

I left the changes in place forabout 50,000 impressions (that's normally 2-3days on my site) and the made an adjustment. But I wonder if that is really a big enough sample.

europeforvisitors

8:24 pm on Feb 14, 2004 (gmt 0)



It's probably an adequate sample if you noticed dramatic differences in CTR, but not if you noticed only slight differences that could have reflected day-to-day variations.

Also, if your CTR tends to be higher on some days of the week than on others, you might want to leave the changes in place for a couple of weeks so that you can compare Monday to the previous Monday, Tuesday to another Tuesday, etc.

onfire

9:12 pm on Feb 14, 2004 (gmt 0)

10+ Year Member



A week at min, a month to get a bigger snapshot and enough data to check against, but just when you think your seeing a pattern - change - improvement and all seems settled, bang...... it changes.

jomaxx

12:10 am on Feb 15, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



It sounds like enough, but the more important statistic is how many clicks you got during that period (which you can't tell us because it would give away your CTR - haha).

In other words, 250 clicks isn't really enough. 2,500 clicks is probably a pretty good sample to make decisions on.

ken_b

12:42 am on Feb 15, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Jomaxx

>>how many clicks<<

Well, yes, I was thinking about the number of clicks involved. Let's just say the change was noticably negative, even on a relatively short term trial.

Thanks.