The demand for data has never been higher, but it is important to remember not all data is quality data. Making business decisions based on bad data is like living off a diet of Big Macs—death will come sooner rather than later. As Doug Bonderud explains in an article for Spark, benchmarking data can be an effective way to clean up data and create a more competitive business.
Whether across silos inside the business or across industry verticals, people are increasingly understanding the value of sharing data. Especially in budding analytics initiatives where there is low maturity, it makes sense for organizations to pool their data in order to understand what “normal” looks like. This is how best practices develop. Bonderud continues to say this:
While information governance and taxonomic requirements remain critical, [Marc Rind, vice president of product development and chief data scientist for ADP] says you can come back to that as long as you are collecting the information. With data volume and velocity on the rise, this is good advice. Collect what you can, when you can, or risk losing its value. Rind also advocates for the value of protection, especially when it comes to collecting HR data. While businesses and providers are trending toward an open source model to help empower collection and discovery, it’s critical to ensure “everything is stripped of identifiable data,” says Rind. This is key for benchmarking. If you’re not anonymizing HR data, you may be behind the curve already.
Ultimately, “cleaning” data refers to uncovering errors in the way that it is collected. For instance, if data is redundant, mislabeled, or generally vague, benchmarking can bring these issues to light. However, if you find that your data is pretty healthy already and yet problems are persisting—consider that it could be the business that is the problem.
You can view the original article here: https://www.adp.com/spark/articles/collection-issues-or-corporate-problems-the-value-of-benchmarking-data-10-1423