Data Ingestion

Data ingestion is the stage where raw events are received from a source (such as a browser script) and taken into the processing system — the first gateway on the data's journey to a report.

Ingestion is usually invisible, yet the quality of every later report is tied to it, because anything lost here cannot be recovered elsewhere. Three issues are decisive at this stage: reliability of receipt (if requests fail to arrive due to the network or the browser, data is quietly undercounted), order and time (events may arrive out of order or delayed), and initial validity (a malformed event must be rejected or fixed right here). A point the tool's user should know is that part of the number gap between two tools forms precisely at this stage: if one tool's script fails to load for part of the audience, its ingestion is incomplete and its stats come out low, without any visible error in the report.