The way I understand it in the case of missing metrics we should pass in while with defaulting to empty strings we would pass which could taint the data.

Recommend checking with @TravisLong also on this.

suggestion:

AttributionMetrics( 
 source = source.takeIf { it.isNotBlank() }, 
 medium = medium.takeIf { it.isNotBlank() }, 
 ... 
)

question: defaults to "Mozilla" so we'd report this for every organic user. I see allowing for a value so wouldn't this be more appropriate for the scenario of organic acquisition?

Empty strings will end up in the dataset, that's correct. In most cases just omitting the value is enough for Glean to keep this as "null", you shouldn't have to explicitly set it to .

This is a good question. Thought about this for a while and I think we should just stay with recording the actual distribution ID and use query in the backend to distinguish them.

Main reason here is if we have another distribution ID that's also organic (similar to what we had before), we would still want to distinguish them in the backend.

Thanks for bring this up though. Good call out.