AdWords is now Google Ads. Our new name reflects the full range of advertising options we offer across Search, Display, YouTube, and more. Learn more

Ads
5.2K members online now
5.2K members online now
Improve your Google Ads performance and boost your ROI, CTR, and Quality Score
Guide Me
star_border
Reply

working with statistically significant data

[ Edited ]
Visitor ✭ ✭ ✭
# 1
Visitor ✭ ✭ ✭


I'm having trouble understanding when my clicks and other data has enough statistical significance that I can bid at the keyword level instead of the ad group level.

 

My avg conv rate is 0.85%. Pretty low, but it's still profitable.

 

I've read strategies for keyword bid management, but the keywords in the examples sometimes have only 20 clicks or so. Shouldn't I wait to evaluate a keyword when it has at least double the number of avg clicks to conv? In my case that's 235 clicks. Most of my keywords never reach that many clicks.

 

And the same with ad groups. What volume is data is enough to justify me bidding at the ad group level, rather than bidding the same for all ad groups according to the account's overall stats?

 

What I've been doing is mostly going by the last 6 months of data for each ad group, and doing very little about keyword performance unless there are keywords that are getting over 200 clicks with no conversions. And I feel weird about that.

1 Expert replyverified_user

Re: working with statistically significant data

Participant ✭ ✭ ✭
# 2
Participant ✭ ✭ ✭

Hi

 

For that I use a tool like this: http://splittester.com/

Re: working with statistically significant data

Participant ✭ ✭ ✭
# 3
Participant ✭ ✭ ✭

But... it´s not only a question of quantity, but also of qualitative differences, categorization and judgement.

If the ad groups / keywords are similar, you can handle them aggregated. If they are different, you will have to wait longer in order to gather more impressions for each segment.

Re: working with statistically significant data

Top Contributor
# 4
Top Contributor

Hi there;

I agree ; If you want to do that, you'll have to do analyze the data yourself, since Google doesn't disclose the errors associated with data. (e.g confidence intervals, standard deviation...) - and in statistics, values without quoting the error - are meaningless.

 

For keywords and ads, I use the AdWords experiments. In experiments you can see differences in performances that are statistically significant.

Moshe, AdWords Top Contributor , Twitter | Linkedin | Community Profile | Ad-Globe
Did you find any helpful responses or answers to your query? If yes, please mark it as the ‘Best Answer’

Re: working with statistically significant data

Visitor ✭ ✭ ✭
# 5
Visitor ✭ ✭ ✭

I like this tool for determining statistical significance:  http://www.usereffect.com/split-test-calculator.