Hello
There are variations in cost values provided by Google Ads API (Data Transfer Job) and the Google Database consulted directly in the google ads platform or in looker studio.
I made an alaysis comparing the information extracted through the API against the information from the Google Ads database exported through Looker Studio.
I took different ranges of days and compared the total cost per campaign. The difference between campaigns presents a range of variation that goes from 0% to a 34% maximum, however, when making a general sum of the cost for each source of data, the resulting variation between them ranges from 2.39% to 5.38% as you can see in the following table.
Days compared | Total Campaigns | Range of % variation between campaigns | Consolidated % variation |
65 | 336 | 0 - 20% | 2.39% |
30 | 283 | 0 - 11% | 2.51% |
15 | 255 | 0 - 26% | 4.40% |
7 | 245 | 0 - 34% | 5.38% |
Subsequently, another analysis was done taking the same range of days but this time the comparison was by date, resulting in a variation ranging from 0% to a 10% maximum between dates and when consolidating the cost of the different periods compared, The resulting variation ranges from 2.39% to 5.38% as you can see in the following table.
| Days compared | Range of % variation between days | Consolidated % variation |
| 65 | 0 - 10% | 2.39% |
| 30 | 0 - 10% | 2.51% |
| 15 | 0 - 10% | 4.40% |
| 7 | 0 - 10% | 5.38% |
In both cases, it is noted that the larger the period compared, the less the variation will be between the data sources. This is because for each campaign there can be the same amount of variation, only that on one day the number is positive and on another day the number is negative so when doing the general sum the values compensate each other, finally showing a small variation in the comparison of results.
For example, for one campaign on April 17th the difference between the total cost from the Google database and the information extracted in the data transfer through the Google Ads API is 8.881784 and on April 18th the difference is -8.881784. On March 14th the difference is 8.881784 and on March 15th the difference is -8.881784.
But when we consolidate the data from February 14 to April 18, the values offset each other giving a small variation of 0.10273300.
Has anyone else had this issue? Why is this happen? Is there a way to solve it?
LinkedIn
Twitter
