Impression count UI vs API - gaps of a factor of ~2

[Tal Lipits]

[Note: I do know that this is not the first time the topic is raised, however I didn't see in the replies a solution for my problem, so I'm raising it again]

Hi,

 

Comparing the impression count I get using UI (keyword planner - forecast) vs the API (TrafficEstimatorService), I notice gaps – roughly a factor of 2.

For example:

• Keyword=’55 inch tv’                      UI Impressions= 4,282                    API impressionsPerDay (average of max and min)= 1965
• Keyword=’baby stuff’                     UI Impressions= 2,313                    API impressionsPerDay (average of max and min)= 899
• Keyword=’converse boots’          UI Impressions= 1,548                    API impressionsPerDay (average of max and min)= 906

 

Config:

• UI:
• Locations: United States
• Language: English
• Max CPC:10 (my currency)
• Date: July 2, 2020
• And also:
• Search networks: Google
• API (campaignEstimateRequest):  
• Locations: United States
• Language: English
• microAmount: 10,000,000
• Date - I executed the script that accessed the API today (July 1, 2020)
• And also:
• matchType: BROAD
• In both cases I use the same client account

 

Questions:

1. In the API - I didn’t specify: date. I understand that the API returns the impression results per day. Is that correct?
2. Is the API based estimation made per the subsequent date of current date (i.e. impression estimates for tomorrow)?
3. In the API - I didn’t specify: search networks. Can this be a cause for the gaps?
4. In the UI - I didn’t specify: matchType (may be I missed this setting elsewhere). Can this be a cause for the gaps?
5. Any idea why I see these gaps?
6. How do I create report definition to see some more data?

Thanks a lot,

Tal

[Google Ads API Forum Advisor Prod]

Hi Tal,

Thank you for reaching out. The keyword planner UI and TrafficEstimatorService actually run on different back-end systems so it is expected that their results will differ even with the same settings. With that in mind, I will answer your questions below:

1. Yes, that is correct. Is is the estimated metrics per day.
2. No, it is not necessarily the estimate for tomorrow or any specific day. Just the projected average daily amount.
3. You should use the networkSettings to specify your desired search networks. If not set, it will default to just targetGoogleSearch.
4. You can specify the match types in the UI by using the symbols explained here.
5. Making sure the network and match type settings are the same should help but as I mentioned, it is expected that the values would be different.
6. There are no reports for the TrafficEstimatorService. You can only see the values returned in the TrafficEstimatorResult.

Regards,
Mitchell
Google Ads API Team

ref:_00D1U1174p._5004Q21AoHC:ref

[Tal Lipits]

Hi Mitchell,

Thanks a lot for a quick reply.

Looks like indeed the settings are all identical.

I ran another test, with matchType=EXACT and got gaps with similar factor of 2:

Keyword=’55 inch tv’                      UI Impressions= 2,592                    API impressionsPerDay (average of max and min)= 1135
Keyword=’baby stuff’                     UI Impressions= 988                    API impressionsPerDay (average of max and min)= 389
Keyword=’converse boots’          UI Impressions= 1508                    API impressionsPerDay (average of max and min)= 748

1. Is there anyway to check that the settings are identical on the server side too? In other words - how do I make sure that there aren't any gaps after all. I could only compare settings I'm using on the client side (whether it is the API or the UI).
2. I understand that these are different systems, and gaps are expected. Just to make sure I'm inline with the expected behavior - Is it also expected to see a factor of 2?

Thanks a lot,

Tal

[Google Ads API Forum Advisor Prod]

Hi Tal,

For your first question, even if the settings match, it is expected that the results would differ. There is no way to check the server-side settings.

For your second question, there is no expected factor difference between the values. As long as you have specified the desired settings, then the values returned are the expected values.

[Tal Lipits]

Thank you for the clarification.