Should your data be perfect before starting with analytics and drawing knowledge from it? Let’s hope not because we might end up solving all Data Quality issues and spend all our budget on that. We should even start using analytics as soon as some sources become available. Data Quality is positively correlated with the its use, so better start right away.
Analytics can help make the distinction between data errors and interpretation errors. Explorative research hence is a perfect start, followed by monitoring of the data already know, to avoid it going out of spec again is the way to move forward. If metrics remain within their confidence boundaries, there is no need to panic. We can easily quantify what’s good enough and furthermore we can install early warning tresholds that trigger signals if we are drifting out of spec. What can we expect more?
Quantifying what good enough means…