The ‘Democratization of Data’ is Here – It’s Just Not Evenly Distributed

democratization data
The ‘Democratization of Data’

For years, a compelling vision of the future was the attainment of the “democratization of data,” in which data analytic power is available at all levels of the organization – from the CEO’s office to the call center to the production floor.

Many of the data services and platforms that have come on the scene in recent years emphasize the accessibility of data to a broader spectrum of users – through self-service portals, online analytics, and even mobile apps. So, it’s clear we’re making great strides in getting data out to decision makers outside of the executive suite or the analyst compound. But, are decision makers making the most of this new capability?  Do they even know how?

Sam Ransbotham, writing at the MIT Sloan blog site, says we’re not quite at a state of data democracy – it’s more a data “meritocracy” – with advantages still going to smaller groups of professionals who understand the inner workings of data analytics.  “This particular democracy offers no guarantee that you’ll get to participate — suffrage is far from universal,” he opines.  Instead, he refers to what many enterprises have as a “Potemkin” data democracy – “an illusion that appears when we incorrectly equate the ability to access data with ability to use data.”

In terms of sheer access, progress has been impressive. Ransbotham cites recent research from MIT Sloan Management Review, which finds the democratization of data is on the rise. A majority of respondents, 77%, report an increase in access to useful data since last year.

However, while more data is accessible, still only handfuls of analysts know how to employ it to advance their businesses. The ability to use data is still wanting – caused by lack of data knowledge. This “hamstrings people in two ways,” says Ransbotham. “First, they are unable to use the readily accessible data well themselves, and second, they are unable to tell when others are using data poorly or disingenuously.”

People to keep an open mind when it comes to learning new data skills, and encourage training to develop these skills. A challenge is to promote the idea that a data-driven culture is open to all, versus what Ransbotham refers to as a “meritocracy” of skilled individuals. In many enterprises, there are two camps – the data analysts and scientists in one, smaller camp, and the rest of the business in the other, larger camp.

Corporate incentive systems also need to be adjusted to allow employees the time and space to learn and experiment with data, and then be appropriately rewarded for bringing new findings or ideas to the enterprise. There needs to be some degree of independence and self-initiative to make this successful. At the same time, instilling a data-driven culture requires deep collaboration – and understanding — between data analysts and business end-users. Businesspeople need to be encouraged and trained to think more like quants, and quants need to be encouraged and trained to think like businesspeople.

Democracies can be messy and uneven, but they work because all citizens have a stake in the process and share in decision-making; this rule applies to budding data democracies as well.