Follow Datanami:
January 7, 2019

Building Data Culture a Priority as AI Investments Ramp Up

(Rawpixel.com/Shutterstock)

It’s not yet a do-or-die situation, but organizations are definitely under the gun to demonstrate positive results from their considerable investments in big data and AI. The pressure is on chief data officers and other leaders to show results from their big data and AI investments in 2019. They’d do well to focus some of their efforts on building a data culture.

The strategy consulting firm NewVantage has been tracking the ins and outs of big data with its annual survey for five years. While there has been some improvement, the firm has yet to record widespread success. “Companies are investing in Big Data and AI, but they are not seeing commensurate results,” the firm writes in its new report, “Big Data and AI Executive Survey 2019,” which was issued last week.

In the forward for the latest report, NewVantage CEO Randy Bean and Tom Davenport, the co-founder of the International Institute for Analytics, wrote that the glass is half full on AI and big data. “While there are still signs of emptiness, over all we see a glass that is half full and filling up slowly,” the two write.

On the one hand, companies are bullish on the prospects for turning data into competitive advantage. The survey of 60+ tech executives predominantly from big financial services and healthcare firms shows more than 90% of organizations are increasing investments in big data and AI. A similar percentage of executives say there is a greater urgency to invest in big data and AI than in past years.

Cultural Shortcomings

However, succeeding at big data and AI is easier said than done. While more than 60% of executives in NewVantage’s survey say they’re getting measurable results, fewer than 50% say they’re “competing” on data and analytics. Fewer than one-third say they’ve transformed their businesses into data-driven organizations, and only 28% say they have created a “data culture.”

The dismal success rate of big data and AI projects is something that everybody in the industry should be concerned with. In 2017 Gartner analyst Nick Huedecker pegged the failure rate of big data projects at about 85%. There’s something to be said for failing fast and failing often, if it eventually gives rise to real success. But the data behind big data failures suggests something else may be at play.

So what’s the big hold up? Like with most things, it’s a complicated equation, with multiple variables impacting the final outcome. The growing technical complexity of the systems put in place to manage and analyze data has been pegged as one of the reasons why these projects more often end in failure than success. But companies are increasingly turning away from heavier Hadoop-style architectures that require a lot of custom development and tuning to more nimble cloud-style platforms that have already been built and can just be used.

This trend shows in the NewVantage survey. Only 5% of execs in NewVantage’s survey say technology is the big challenge. The biggest challenges actually stem from other factors, like organizational alignment, overall agility, and resistance to change. What’s more, culture was cited as a big challenge by 95% of survey-takers, which is a strong indicator of what needs to change before organizations can truly succeed with data.

The need for a bigger, better, and stronger data culture will be a recurring theme for big data and AI in 2019. But just what is a “data culture” and how do you build one?

Data Culture Defined

According to McKinsey and Company, data culture can be an elusive thing to achieve. “You can’t import data culture and you can’t impose it. Most of all, you can’t segregate it,” write three McKinsey partners in the September 2018 article “Why data culture matters.”

Buy-in from the top is important for building a data culture (Monkey Business Images/Shutterstock)

The McKinsey partners did a good job describing what a data culture is and showing examples of how it works. They identified seven key aspects of building a data culture, as follows:

First, a data culture is inseparable from a decision culture. The only thing that data is good for is for making better decisions. Data is essentially useless by itself, and only has value when it’s paired with the ability to impact decision-making.

Second, data culture must be supported from the top. For some items, having a grassroots effort that originates with lower level employees is a good thing (see next item). But for a data culture to truly take hold at an organization, McKinsey finds that there has to be buy-in and support from the CEO and the board of directors.

Third, the democratization of data is important. Getting employees at all levels of the company involved with using data for decision-making is another important aspect of building a data culture. Removing the barriers that prevent data from flowing is a good way to stimulate data democratization.

Fourth, don’t overlook risk as you build your data culture. The potential to lose sensitive data or to build unwanted bias into algorithms are real threats that can stymie the growth of a data culture, not to mention damaging a brand. Putting clear policy and process lines in place gives data practitioners the confidence to know exactly where they can operate.

Fifth, find “culture catalysts” who can bridge the new world of AI and data science with existing business operations. These catalysts, who are rarely digital natives, are necessary to set a positive example for the rest of the company and to lead the charge into new analytic endeavors.

Sixth, don’t assume you need outside data for every project. While there are cases where a rising tide can lift all boats (such as supply chain or logistics), McKinsey throws cold water on the idea that the path to a sustainable data culture is to participate in one of a number of emerging data markets or “ecosystems.”

Seventh, look inward for data talent. While bringing in an outside expert is necessary in some cases, McKinsey stresses the need to maintain a balance between “injecting new employees and transforming existing ones.” This makes a lot of sense when you realize that building any kind of culture takes time and can’t be created by fiat.

Just as a good data culture can help you move forward, a bad one can set you back. The most important thing to realize, the McKinsey partners write, is that a data strategy must be informed by the business strategy and core operations.

“But when excitement about data analytics infuses the entire organization, it becomes a source of energy and momentum,” they write. “The technology, after all, is amazing. Imagine how far it can go with a culture to match.”

Related Items:

How to Build a Data-Driven Culture

The 3 Roles Needed for the Modern Data Team

How to Build a Big Data Culture

 

Datanami