It’s easy to embrace new technology without realizing what it takes away from you, like GPS did to the Inuit people.
Nicholas Carr, in his 2015 classic, The Glass Cage, recounts how the Inuits developed the skill of “way-finding” over thousands of years. Crossing dark and dangerous ice landscapes that look the same in every direction isn’t easy, so they learned to use subtle cues like wind direction, sounds, or even animal behavior to help them find their way. Then GPS technology came along, and way-finding became a lost art. Today, many more Intuit people get lost or injured as they navigate these spaces.
If you started working in data sometime over the last 15 years, you might have the same problem as the Inuit. Solving old business problems with new technologies can destroy your ability to find your way around the data. Last week, I showed you the features of a good data warehouse. This week, I want to remind you how easily new technologies can cause you to forget this lost art.
New Names, Same Problem
IBM researchers in the late 1980s originated the idea of storing data specifically for use in decision-making. They repurposed existing database technologies and called it a “data warehouse.” Systems today move, store, and retrieve data far faster and at much higher volumes than the ones I used back in the 1990s, but they haven’t changed that original idea: a data warehouse is a warehouse. A data warehouse stores data.
People miss the fundamental purpose of a data warehouse by thinking it only exists to support analytics. A good data warehouse does much more than that by taking responsibility for all a company's data needs. Just a few examples include improving data quality across the entire company, synchronizing master data to other applications, and migrating data from one application to another. At all the companies I worked for, the IT department understood this value and included data warehouse activities in all their project plans. Services like these make your whole company more agile with its data.
Vendors try to give you more attractive names for data warehousing. “Data lake,” for example, communicated something less structured but more fun and refreshing. Gartner coined the term “data fabric” to describe how systems of the future would stitch data together like a blanket. Then I read about Microsoft’s latest solution, “Fabric OneLake”. I felt like I was hearing about Taco Bell’s latest menu option, the Cheez-It Crunch Wrap, which is really just the latest reincarnation of the Doritos Volcano Taco.
But in the end, warehousing the data isn’t a technical solution; it’s a management strategy.
Don’t Lose Your Bearings
Like those Inuit hunters, I fear that all these technological advances could cause you to lose your bearings. A data warehouse is not primarily a technical strategy; it’s a business strategy. But companies spend a lot of money on technology solutions that don’t make their data more accessible or their decisions better. Investing the time and resources needed to plan, organize, and fill up a data warehouse doesn’t fit into the project timelines that most analytic projects expect. Many data teams just skip this effort and go straight into developing reports for business metrics.
If your boss asks you for the ROI of a data warehouse, tell them that it will help their teams avoid getting lost while hunting for data in the cold, dark data wasteland surrounding them.