<< Back

Garbage In, Garbage Out: Why the quality of your data matters more than its speed


Many years ago (I won’t tell you how many), my high school computer science teacher first introduced me to the concept of GIGO. It stands for Garbage In, Garbage Out and what it means is your results are entirely dependent on the quality of your beginnings. If you put faulty information into a computer program, you’ll get faulty results coming out the other end, no matter how sweetly designed the program itself is.

The same holds true for business information systems.

If the data you’re putting in to the system is less than high quality, you definitely won’t be getting high quality insights at the other end.

Often times you will see the business intelligence community getting all starry-eyed over the idea of real-time analytics. The idea that a sales person in Vancouver can close a major deal in the morning and you’ll know how it affected sales, inventory, and your bottom line before the ink is dry on the customer’s signature is an enticing one… but is it entirely practical?

Yes, if you are 100% positive that the data being input into your BI system is absolutely, perfectly, squeaky-clean high fidelity data. But if it isn’t, you just might be better off slowing things down a little.

Spending time getting your data model right, building reports, and developing the most elegantly insightful dashboards anyone has ever seen won’t mean a thing if you don’t also spend time making sure the data being generated by your business processes is of the very highest quality. And if for some reason you’re forced to work with suboptimal data, you should also be spending time making sure you understand how that data is contaminating your system and how you can adjust your systems to clean out the garbage and start giving you quality results.


You can use one of these four ways to address the problem.

  • Make an exception. If you know that this is a one-time thing and you’ll have the problem fixed soon, you can simply treat it as an exception this time to work around the problem.
  • Fix the data after it gets to you. If you can’t change the data at it’s source, adapt your analysis to take into account the data discrepancies. This might mean taking the data and running it through a few clean-up procedures before passing it to your BI system to eliminate duplicates and get rid of formatting discrepancies that make the data about ACME Industries and A.C.M.E. Industries appear to be 2 different things instead of one and the same.
  • Fix the data at the source. By far the best solution, by ensuring your data sources are providing you with the highest possible data to begin with, you can be confident that the information you glean from your data is the furthest thing from garbage.
  • Have a data audit. While anyone can tackle a data source and eliminate duplicates, inconsistencies, and other common data integrity issues, only those intimately familiar with your operations and your clients will be able to determine if the data you’re collecting is complete enough to be meaningful for the types of things you want to do with it. This isn’t something that can be done by data cleaning software, either. It needs a thorough mind to determine whether or not you’re collecting all the data you need and if it is performing adequately. And it never hurts if they can anticipate future needs, too—you might not need that particular data point at the moment, but could it be useful in the future? Only a savvy data analyst will not only understand how to tie your current data to your business goals, but how to ensure you’re collecting the right data for future business goals, too.

All of this falls under the larger heading of “data governance” and as we talked about last week, we believe data governance is going to be more and more in the news this year when it comes to business intelligence. As more and more businesses embrace BI, there is going to have to be a corresponding increase in awareness of the importance of good data governance, too.

Like this post? You might like some of our other posts too!

Don’t miss another blog post – subscribe today! Enter your email address in the sidebar to the right and click Subscribe, or follow us on Facebook or Twitter.