A data audit is an assessment of the state of the data a company holds. It precedes every major project – deployment of a new system, analytics, and artificial intelligence alike – because the quality of the inputs determines the outcome. Several dimensions are examined at once. Completeness, that is, how many records have their key fields filled in. Accuracy and correspondence with reality, which is verified on a sample. Currency, that is, how old the data is and when it was last changed. Consistency between systems, where the biggest surprises tend to be, since the same piece of information differs across three systems. And accessibility, that is, whether the data can be reached other than by an export from a single person. The output is a list of problems with an estimate of effort and the order in which to address them.
See also: Master data management, Data normalisation, Company AI Audit.
