Glen Rabie is Co-Founder and CEO of Yellowfin, an Analytics and Business Intelligence company that helps businesses understand their data.
getty
There is a rumor going around in IT and big data circles, encouraged by some very respected organizations. It goes like this: The only way to make employees fully data literate (and yes, everyone in a modern enterprise should be data literate at some level) is to impose major organizational change on the corporate culture and implement training until every last person is fully versed.
This approach is well intentioned and even well reasoned. But it is a huge ask and, in most cases, unnecessary. More importantly, it often doesn’t succeed.
The reason it doesn’t work is that not all employees require the same standard level of data literacy. Those who analyze or wrangle data for a living obviously require a much higher level of literacy than those on the front lines of business who consume it. Moreover, those who use data only tactically have different literacy requirements. Many of these individuals simply need sufficient literacy to do their particular jobs — which means that extensive, enterprise-wide data literacy programs can result in overtraining, not to mention an onerous responsibility for CDOs and their staffs.
Yet there is another element in this push for broad data literacy that needs to be addressed: the role of software vendors.
Instead of embarking on massive training programs, organizations should put pressure on their vendors, demanding that they provide tools that don’t assume or require deep literacy. Any training that is implemented should be focused on how employees can better use data within their jobs, not on how to extract a report out of their software.
BI Is Entering A New Phase
Enterprises are far beyond the first phase in the data analytics software revolution when platforms were built for professional analysts. They’ve even surpassed the era of “user-friendly” BI platforms that enabled business users to interact with data through reports and dashboards. When those second-generation applications arrived, nontechnical personnel still needed a certain level of skill in data wrangling in order to prevent adverse outcomes. As a result, the need for data literacy actually went up.
Today, BI platforms are quickly moving into phase three — a period in which users no longer have to battle their way through analytics. Given the advancements in data automation, embedded contextual processing, natural language queries, machine learning, collaboration tools and the ability to create business narratives within the platform itself, solutions are lessening the need for advanced data literacy instead of requiring it.
Sufficient breakthroughs are here for vendors to create applications that assist, accompany and automate data. The goal should be to help the process along, making it easier for users to digest, interpret and consume. And for the CDO and their organization, the mission should be to select or even demand tools that help the enterprise better understand data and how to use it.
There is one caveat to this discussion: the acknowledgment that there are different use cases for BI software. CDOs need to select the right platform for the right deployment. Someone doing advanced predictive analytics will need to be very well trained; on the other hand, solutions for business users should reflect their skill sets. A marketing specialist isn’t an expert in data science — nor should they have to be. They should only need to know enough about data to be an expert in marketing. By pairing software with the right disciplines and skill levels, organizations will be able to focus their training investments where they’re needed most.
The Need For Nuance
As organizations move through their digital transformation initiatives, there’s no doubt that data is shifting to the very heart of business activity. This is what’s driving the self-service BI movement. Yet the data literacy mandate, despite its current emphasis, needs to be nuanced.
The future belongs to vendors that provide — and organizations that demand — platforms that deliver sophisticated analytics to business users without the need for advanced data literacy training. These users should only have to consume the necessary information at the point where it’s needed in a way that makes sense to them. Training ought to be focused on how to apply the knowledge that’s gleaned from data instead of how to obtain that knowledge in the first place.
Those responsible for data literacy should make sure that training is tailored for the role and the competency required. Training should certainly include tools — but with a focus on those specific tools people need to access and analyze their particular data.
Best practices may also include insights regarding situational bias. When evaluating a business problem, it's easy to allow bias to creep in, so it’s wise to make people aware of inherent biases, as well as how data can be used to overcome them. Finally, measures should be in place to ensure that data is leveraged most effectively (i.e., systems are used, appropriate data is applied and people do not revert to gut-level decision-making practices).
If organizations require improvement in their BI performance, it shouldn’t have to come through huge investments in data literacy training. Whether through automation, contextual analytics or better situational storytelling, the real answer will come through tools that actually advance the way the enterprise leverages its data — an improvement that will truly make everyone in the organization smarter, and without all that instruction.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
