fivetran / fivetran/dbt_twitter
[Feature] Add documentation on differences among aggregations across different grains
- Dominant language
- No language data
- Stars
- 2
- Forks
- 5
- PR merge metrics
- No merged PRs in 30d
Description
*Copied from [fivetran/dbt_ad_reporting #92](https://github.com/fivetran/dbt_ad_reporting/issues/92).*
### Is there an existing feature request for this?
- [X] I have searched the existing issues
### Describe the Feature
Some customers have brought up concerns on why sums are different across different grains, for example why spend is not the same summed up across the ad level versus the campaign level.
Example from Joe regarding a recent customer inquiry:
```
"The reason for this discrepancy is due to not all ads being served at the ad level. There are some ads that are only served at the ad group/campaign/etc. levels. This means that you can have ads sent at the campaign level and that counts towards your total spend. However, since the ad was never sent at an individual ad level, it will not be included.
This was a large reason for us breaking the ad reporting package into separate hierarchical end models. Because we found when we only used the ad level, we were missing data as some ads were not served at an ad level, but were served at a campaign level."
```
We should add more context like above to our ad packages READMEs as a proactive measure.
### Describe alternatives you've considered
_No response_
### Are you interested in contributing this feature?
- [X] Yes.
- [ ] Yes, but I will need assistance and will schedule time during your [office hours](https://calendly.com/fivetran-solutions-team/fivetran-solutions-team-office-hours) for guidance.
- [ ] No.
### Anything else?
_No response_
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by locating the ad packages' README files and reviewing how they describe reporting grains and spend. Add the provided explanation of why aggregations can differ across ad, ad group, and campaign levels, then confirm the README context clearly explains why lower-level reporting may omit data.
Written by the indexing model from the issue text.
Assessment
- Domain
- documentation
- Issue type
- Documentation
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100