Back to Newsroom
Back to Newsroom

Meet The Data Scientist Encouraging Companies to Rethink Data Governance

Thursday, 25 March 2021 03:15 PM

DALLAS, TX / ACCESSWIRE / March 25, 2021 / The rush for data is like the gold and oil rushes before it: the data is there for the taking, but knowing where exactly it is, how to get it, and what to do with it once you've got it is something else entirely. Every day, well over 2.5 quintillion bytes of data are created - the issue here is not one of supply, it's of management.

Companies that were once scrambling to collect as much data as possible are now taking a step back, nearly drowning in all of the facts and figures they've managed to accrue. The next frontier of data isn't data collection: it's data governance.

Enter Vincent Yates, Partner and Chief Data Scientist at management and technology consulting firm Credera. After receiving his Ph.D. in Statistics from the University of California Berkeley, Yates led data teams at Yammer, Microsoft, GE Digital, Zillow, and Uber - few have as much insider knowledge of how data is handled at the enterprise level as Yates does. The insider info has led Yates to a troubling conclusion: the vast majority of businesses are grossly mishandling the data they collect.

The statistics back him up on this one: 99.5% of collected data is never used or even analyzed. Even the data that does make it through analysis is often woefully underutilized, as, through 2022, it's predicted that only 20% of analytic insights will deliver tangible business outcomes. Companies generate huge amounts of value from their data already, so just imagine how much they could increase that value simply by taking greater advantage of the data they already have on hand.

Yates hopes to make this possible for more companies. "Many companies lack the strategic alignment to harness the full potential of data. To be a data-driven company, your data strategy must be the same as your corporate strategy," Yates says. "If that's not true, then you are not data-driven, and your corporate strategy is probably misaligned because you are not taking advantage of your data." Starting with an integrated data strategy will then inform the practical and tactical next steps for utilizing the data. In order to do so, Yates recommends:

1. Giving your data a purpose.

If there's one piece of advice Yates wants to pass onto businesses with budding data operations, it's that they don't just "do something." Data is a bit of a trendy topic right now, so the temptation can be to start compiling, translating, massaging, and otherwise toying with data just because it seems like the right thing to do. Once you have the data on hand, you need to have a real goal for it, not just an abstract hope that it will make your business better.

The best way around this is to start with an issue, a question, or an improvement you'd like to make. From there, collect the data relevant to this impetus. Once you have what you need and you know what you want from it, build just enough to get the insights from it that you want. Software developers suffer from "feature creep," and data experts can see the same thing happen to them. Repeat this process ad infinitum - it's a great starting workflow for getting your data to work for you.

2. Keeping things in sync.

Any organization that collects data collects from multiple sources and keeps it in multiple places. Harmless, right? Well, the truth is that decentralization will inevitably lead to data contradictions: if one of your customers changes their address, how many different datasets will need to be altered as a result? Do you have the infrastructure in place to make that happen?

The solution here is not to centralize all your data - different divisions of your business need different things, so keeping a single central ledger and updating it perpetually would be far more trouble than it's worth. What you should do instead is create fixes from the ground up. Identify databases that have big data turnover or close relationships and figure out how to maximize efficiency there. There is no catch-all solution for data governance: it's highly localized, and you should treat it as such.

3. Not leaning in too hard.

It's both tempting and exciting to create as many data algorithms as possible, setting up alert systems to let managers and executives know of new changes and patterns in your company's data. Lean in too hard into this practice, and you risk some serious data burnout. Think of it as the boy who cried wolf: eventually, you and your coworkers are going to stop looking at these updates as groundbreaking insights and start to ignore the lion's share of them - even if they have vital information. Create exactly the setup you need and expand as necessary, no more.

Creating an effective data governance regime cannot be done overnight. While it may take a while to see the results you're hoping for, the end product will not be disappointing - data, when used in the ways Yates recommends, can and will elevate your business to the next level.

CONTACT:
Sarah Bruner
[email protected]
972-759-1843

SOURCE: Credera

Topic:
Company Update
Back to newsroom
Back to Newsroom
Share by: