Hi Julio,
I have been into this topic from many weeks and i dont know if i could explain you better.
Inefficiencies are pre-defined within Celonis Starter kit and after running transformation job for manual inefficiency table it gets loaded in data model. Inefficiencies are tagged within Knowledge model for their use cases.
Impact: KPI value with their respective inefficiency is compared with Global Value of KPI and what is % change (positive/negative) after comparison.
You can also modify/create Impact calculation as per your need. Please refer this link for more information: Configuring an execution gap list in legacy views (celonis.com)