A/B testing for position preference

Replies to This Discussion

Permalink Reply by Jamie Murphy on May 8, 2009 at 9:55pm

Hey Ken,

A/B testing can prove very interesting. If possible go beyond CTR/CPC and track conversions, time on sight and other website metrics.

Cheers, Jamie

▶ Reply to This

Permalink Reply by Ken Thomas on May 9, 2009 at 8:17pm

Oh-oh.
I've run this test, with two identical campaigns except for different position preferences, 1-3 and 7-10+ (and with different campaign names).
Result?
The 7-10+ campaign, so far, has averaged at position 3.5 - with higher CPC than the 1-3 campaign!
I wonder if there is something in AdWords that makes it 'blind' to position preference (and campaign name) as the only differences between otherwise-identical campaigns?
I'll let this run a little bit longer and see if the results start to look a bit more logical.

Regards,
Ken

▶ Reply to This

Permalink Reply by Jim Jansen on May 23, 2009 at 7:26am

generally not a good idea to duplicate adgroups, but what the heck, see what happens

▶ Reply to This

Permalink Reply by Ken Thomas on May 23, 2009 at 10:25pm

Thanks Jim.
"Not a good idea" is exactly what I learned. The preference settings didn't result in a proper test - the 'low position' campaign continually ran ads in high positions.

▶ Reply to This