SLA on reporting and fixing bad profiles
J
Juniper Hare
User Story: As I data app/contact management provider, I want a 7 day SLA for PDL giving an answer on why a problem profile was merged/wrong.
For example, if Mary Jones and Mary Smith are joined in the same profile, the SLA would require PDL to get an answer within 7 days on why/how the profile was joined and whether it was correct/not correct.
The best case scenario for this feature would also be that a reported profile (that is bad) would be broken up/unmerged when a profile is clearly a false positive merge.
An alternative "less than best" use case is we submit a PDL ID and the data is removed from the PDL dataset so it doesn't get sent to us again, and then we don't resell it to our customers.
L
Loud Stork
With over 3 billion person profiles, we want to ensure that our approach to data quality issues is scalable. For this reason, when we come across false positive merges, rather than applying our engineering resources toward fixing that particular profile, we use that profile as a research data point to try to find other false positive merges, determine overlapping root causes, then implement longer term solutions that solve much larger swaths of profiles.