Iuvo

Frustrated with Adwords because you can’t keep track of the little changes you make or can’t drive down the keyword changes you made and their impact on your campaign as a whole?  Feel like Adwords is a cauldron that you are throwing ingredients into but not really sure of their affect on the brew?  Well then, Adwords Experiments might be what you need!

Adwords Experiments is an addon to the Adwords interface and is currently in Beta (like all cool Google products).

Experiements allows you to split test multiple variations with your campaigns, such as Adcopy, URLs, new keywords, max bid changes etc.  For this example I will explain an experiment I did with changing the max bid for a particular keyword.

The Experiment options are found at the bottom of the settings page for the campaign you would like to run the experiment for.  See below..

The experiment is a 50/50 split for one keyword.  The experiment will have a 25% increase in max bid while the control of course will stay the same.  Each time the keyword is eligible for an auction, 50% of the time it will run with the original bid and 50% of the time it will run with the 50% increased experiment bid.  This is how you can see the difference between the control and the experiment.  Once you have setup your experiment in the settings, you need make the changes to your guinea pigs.

For keywords, go to the keywords tab, find the keyword you want to experiment on and click on the “segment” drop down menu and click on experiment. (See below)

You now should see the keyword details expand and the split of the keyword bidding into control and experiment.  Now you need to alter the bid for your experiment.  Click on the experiment max cpc bid and you will see the below window pop up…

Make the changes as required to meet your experiment and save the changes.  Depending on the amount of traffic that keyword receives, i would recommend letting the experiment run for at least a week before making any changes.  When the experiment sees a significant change, positive or negative, it will present a graphical representation.  Here is an example of the keyword I experimented on..

This example shows that by increasing the bid by 50%, the keyword received an additional 53 clicks, 2150 impressions, increased it’s CTR by 0.96%, increased its average rank from 2.2 – 1.5 and received twice as many conversions.  These all seem like very positive results, and they are, but there are some less optimal results included in the mix.  The Average CPC went up from $1.10 to $2.35, the cost obviously increased greatly with the number of clicks, the conversion rate went from 20% to 6.25% and the cost per conversion $5.80 to $38.02.  It would be great if positive results carried over perfectly to the experiment but they don’t.  Having a 20% conversion rate is amazing so wanting that keyword to get more clicks is justified.  For this experiment I will keep it running for another 2 weeks and see what happens to the results, more data = better decisions.  The cost per conversion is still below the clients expectations so this in turn brought them more conversions within their goals.  Also, the increased CTR should also have a more postive impact on the QS and therefor CPC, and overtime could improve the results even more so.  Conclusion – I will keep this experiment running for longer and work from there.

I really enjoy using experiments as it helps me test assumptions and ideas while keep the results well documented and easy to access.

Have you tried using Adwords Experiments?  Need help with Google Adwords Experiments?

Give me a call — (646) 351-0659 or email me at hugh@iuvoonline.com

Hugh Prysten: Consultant & SEM Strategist at Iuvo Online Media Consultants