AdWords
1.8K members online now
1.8K members online now
Improve your AdWords performance and boost your ROI, CTR, and Quality Score
Guide Me
star_border
Reply

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Visitor ✭ ✭ ✭
# 1
Visitor ✭ ✭ ✭

Hello everyone,

 

just wanted to solicit some feedback...not really sure there is a solution...but still...wanted to see if I can get some feedback.

 

I used the keyword tool to design a search word campaign. I provided about 20 Keywords and then had google suggest another 200. I reduced the number of keywords down to 109 , used phrase match and for the given budget and duration I ended up with somewhere between 1150 to 1400 clicks with somewhere between 35 to 43 k impressions. I have attached the screenshot.

 

Average position as per the estimate is 3.45 to 4.22

 

 

So far so good....problem is that when I I set the campaign live..and I am getting  basically zero impressions (3 to be exact within the first 20 hours).  

 

When reviewing the campaign it becomes blatantly obvious what is wrong...the average first page estimate for any of the keywords I am using has basically gone up by more than 1000 %..some are as high as 3000 % (give or take). 

 

None of the 109 keywords I have set up even shows on the first page....and it looks to me that unless I increase the CPC by at least 1000 %...the campaign is done.  

 

Campaign Health Score is 100 %....I obviously have no data yet to Quality Score but the ads are spot on.....compared to other ads.  

 

I am by no means a pro in search (last campaign I set up is 18 months ago)...but I am just floored by the difference in the estimate and what I am getting now. The way I look at it..the keyword estimate/campaign estimate is completely and 100 % worthless when the real life results differ to the degree I am experiencing now.

 

I have been in touch with support..and their take was to give it a week and see what "happens". (geez)

Clearly..that was not the answer I had been looking for...hence I am here now.

 

Anyone has suggestions / comments...ideally I am missing something, I flip a switch....and the campaign performs. 

 

Any feedback is appreciated....

 

Cheers

Estimate.png

 

1 Expert replyverified_user

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Top Contributor
# 2
Top Contributor

Well... This is statistics. Sometimes the margin of error is so broad and the estimates don't work. Mainly  due to  more intense competition and lower  QS than estimated.

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’

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Visitor ✭ ✭ ✭
# 3
Visitor ✭ ✭ ✭

?....this is statistics?...are you referring to it as "within the margin of error statistics" ?

Just because I hold some advanced degrees..what is going on here ... is not "statistics". 

 

The competition and bid environment is known to Google and is part of the algo which computes the forecast. QS hasn't even computed yet...and it is known at the time of the forecast for other ads within the space...so I am pretty sure they use a median value.

 

I actually consider the whole thing everything but a statistical anatomy....

 

I am pretty sure if the keyword/campaign planner is routinely off by 99 % there would have been significant push back from the community and Google would have discontinued the use of the product altogether (and/or had corrected it). 

 

Any further feedback as to get campaign results closer to forecast would be great.

 

Thanks

 

 

 

 

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Top Contributor
# 4
Top Contributor

 This is a multi variable statistics, many variables can change during the live auction,  especially your campaign performance (metrics). For example:  If the system estimates a QS of 6 and yours is 4, then there could be a huge difference

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’

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Visitor ✭ ✭ ✭
# 5
Visitor ✭ ✭ ✭

Honestly...since you feel the above is an acceptable explanation  for a discrepancy of about  99.75 % of live results  versus simulation ....I am actually at a loss of words.  

 

I appreciate you taking the time to respond....clearly I was hoping for more than just "that's just the way it is".

 

 

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Top Contributor
# 6
Top Contributor

I have seen this  many, many times.  The key here is  that statistics (and AdWords is a statistical platform) is a science of scale. 

If your metrics / variables  are outside the variance   (or standard deviation - depending on model applied), then the statistics does not "represent" your campaign performance. (I gave an example how this could easily happen).

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’

Live Campaign Results are off by a factor of 99 % versus keyword and campaign esimate

Visitor ✭ ✭ ✭
# 7
Visitor ✭ ✭ ✭

Well...I am not trying to pick your answer apart....but you are actually making my point that something is not right.

 

I am not introducing variables here....all of my input is defined.  (keywords (most of which are suggested by Google anyhow) , CPC, budget) The only entity who can introduce  a variable is in fact Google....

You simply cannot shrug it off and say..."this is statistics"...it really is not..this is not statistics.

 

Anyhow....if in fact, as you stated, you have seen this plenty of times...then it is what it is I guess.

 

If anyone else on this forum has a constructive idea if I can change something on my side to get real life performance closer to the simulation...I would love to hear it.

 

Thanks to all of you for reading.