A man in a suit, dismayed by what the analytics and data are telling him
PHOTO: Shutterstock

Despite the priority Chief Information Officers have given analytics and their deployment over the past three years, a recent survey carried out by Gartner shows that 91 percent of organizations are still struggling with analytics. The global survey asked respondents to rate their organizations according to Gartner's five levels of maturity for data and analytics.  

Analyzing the Findings

According to the findings of the Survey Analysis: Traditional Approaches Dominate Data and Analytics Initiatives report, most organizations have not yet reached a "transformational" level of maturity in data and analytics.

The report found that 60 percent of respondents worldwide rated themselves in the lowest three levels. "Most organizations should be doing better with data and analytics, given the potential benefits. Organizations at transformational levels of maturity enjoy increased agility, better integration with partners and suppliers, and easier use of advanced predictive and prescriptive forms of analytics. This all translates to competitive advantage and differentiation,” said Nick Heudecker, research vice president at Gartner.

The majority of respondents worldwide assessed themselves at level three (34 percent) or level four (31 percent). Twenty-one percent of respondents were at level two, and 5 percent at the basic level, level one. Only 9 percent of organizations surveyed reported themselves at the highest level, level five, where the biggest transformational benefits lie.

Gartner found improving processes was the most commonly cited reason for deploying analytics with better customer experience and new product development cited as the second most cited joint reason.

There were a broad range of barriers that prevent organizations from increasing their use of data and analytics. The research identified the three most common barriers as defining data and analytics strategy, determining how to get value from projects, and solving risk and governance issues. 

However, there is more to it than that. In fact, when we contacted enterprises and vendors we found the following 9 things slowing adoption.

Related Article: Data Scientists vs. BI Analysts: What's the Difference?

1. The Wrong Enterprise Culture

Manny Medina is CEO of Seattle-based Outreach, which develops analytics for sales engagement platforms. He said that the top reason companies have failed to become truly data-driven is that they have not built the required DNA into their enterprise culture. Declaring that you will operate as a data-driven company and requiring teams to present the data before making decisions is table stakes. “To really get this baked into your DNA you need to create a culture of iteration. Analytics is a triangulation game, there’s no silver bullet and it’s nearly impossible to get to true insight in a single shot. The problem is, in most organizations the ask is too big and the penalty for being wrong too high," said Medina

If you really want to reach transformational levels of data analytics, enterprises need to think fast and small. What are the micro-insights you can quickly and accurately identify, and how can you stitch them together to build the complete picture over time. Teams should be encouraged to operate this way and rewarded for every iteration. Only then are you on the path to transformation.

2. Vendors Over-Promise and Under-Deliver

He adds that the buzz around analytics and Artificial Intelligence (AI) has done more harm than good largely because the vendors promoting it have over-promised and under-delivered. “They’ve been promising a panacea and selling it to the highest levels but when it comes down to the troops on the ground who have to implement it the panacea quickly breaks down. They quickly realize they’re not close to ready to reap the rewards of these technologies,” said Medina.

Enterprises need to remember that there is a huge amount of heavy lifting to do around data preparation before they build the data pipe that will eventually fuel analytics. “This data preparation is low-level, heavy lifting that no one wants to do. But there’s no shortcut to transformation. The best way to accelerate your efforts is to provide all the support these teams need and celebrate every small win along the way,” he said.

Related Story: Top 10 Things to Measure in Google Analytics

3. Lack of Vision, Strategy and Champion

Niranjan Krishna is Head of Data Science and Innovation at Santa Clara, Calif.-based Tiger Analytics. He points out that there is more than one reason for slow adoption and no single bullet solution that can magically transform an organization. It is long journey that involves hard work along several dimensions. He offers this advice, Organizations need a coherent vision that clearly outlines what they want to achieve through analytics in the marketplace, a broad-brush strategy on how they are going to realize that vision and a champion who can drive and execute that strategy. Organizations often lack all these elements.

4. Bridging Analytics and Business

Organizations need a good bridge between the analytics side and the business side. The main function of this role is to demystify analytics for the business. “The CIO’s office is not often well positioned to play this role. IT leaders tend to be experts in creating, storing and moving data efficiently while battling issues around memory, speed and scale,” Krishnan said. Data analytics is a downstream function that kicks in after all these core issues are addressed to a reasonable extent. IT leaders are not adequately prepared for this. The data scientists organizations hire also do not typically rise up to this role.

5. Not Proving the Business Value

Data Scientists often struggle, or fail, to quantify the business value derived that can be derived from deploying analytic solutions. This is a big reason why many proof-of-concept solutions and prototypes do not get deployed as enterprise-wide scale solutions.

6. AI and Machine Learning Still a Black Box

Achieving “transformational” maturity in data and analytics necessarily requires the deployment of Artificial Intelligence and Machine Learning solutions, according to Krishna. Machine Learning and Artificial Intelligence involve a lot of “black box” techniques that do not provide transparency into their internal workings for business leaders. This often raises the barrier to adoption.  Ketan Karkhanis', General Manager of San Francisco-based Salesforce Analytics, viewpoint is this, analytics is no longer just about data, numbers, predictions, scores we also need recommendations and explanations, which leads to trusted transparency. A business user won’t trust recommendations if they don’t know the AI is suggesting an action. “At every customer interaction, every point of decision making, and every business process you need the answers to four questions: What happened, why did it happen, what is likely to happen, and what should I do about it. If you’re not answering all of those questions, you’re not using analytics to the fullest,” he said. In other words, transparency into these AI processes can help build trust in the system.

According to Krishna, often in highly regulated industries organizations need to provide full visibility into their decisions in order to be compliant. he offers this example, "An AI underwriting solution would predict that a particular applicant for a particular insurance policy is high-risk but would not provide precise details as to why." It's easy to understand how that would be troubling to an insurance company's customers.

Both highlight the point that transparency into these AI processes can help build trust in they system which will help improve adoption.

7. Protecting Analytics

Doug Bordonaro is Chief Data Evangelist at Palo Alto, Calif.-based ThoughtSpot.  In the past he has also led technical and data teams at IBM, Netezza, Disney, and AOL. He says that despite the investments companies make in analytics technologies, talent, and processes, most business people still don’t have access to analytics when they need them. Instead, BI and analytics teams treat data like gold, hoarding it and only doling it out to the highest levels of the organization. “For an organization to truly reach transformational maturity, they need to storm the analytics ivory tower, and ensure that employees on the frontline have the access they need,” he said.

8. Report Lock-In

Many CIOs and tech leaders think if they can establish the best process, create the right report, and deliver it on time, they're creating value with their analytics investment. “When most of your time is spent fielding angry calls from colleagues about long delivery times or SLAs being missed, it’s where you're going to focus, but getting too involved in daily delivery misses the larger point of analytics. Are you actually providing your teams with information they need to make decisions? Do you understand the business problems they're trying to solve? Could there be an alternative to a report?" said Bordonaro.

9. Thinking Bigger: Systems Vs. Customers

When organizations consider analytics solutions and programs, they still tend to look at it from a systems viewpoint rather than a customer standpoint. When an organization’s leader asks “why are we doing analytics,” it shouldn’t just be to have a better data warehouse strategy, Karkhanis said. Leaders need to be thinking much bigger, about how to improve business metrics such as margins and CSAT (Customer Satisfaction Score) to increase visibility and shape winning behaviors.