This is one of the leading criticisms of self-service business intelligence tools, for example. Critics of the technology say business users are simply too unskilled to be trusted with it. In Five Drawbacks to Self-Service BI, Matthew Gierc points out that those without an IT or statistics background are much more likely to commit data fallacies, thereby leading their companies astray. Nimrod Avissar worries that ad hoc data modeling will lead to “data anarchy.” Hyperbole aside, enterprises are beginning to see their well-intentioned data strategies stymied by low literacy.
SaaS companies offering embedded BI as part of their products are indirectly affected by low data literacy. If a customer is getting little to no value out of the BI solution embedded in your product, they’ll be more likely to downgrade their subscription or, worse, churn away to a competing SaaS vendor. For this reason, it is valuable for SaaS vendors to be aware of this BI roadblock and encourage enterprise data literacy education.
Gartner defines data literacy as:
“…[T]he ability to read, write and communicate data in context, including an understanding of data sources and constructs, analytical methods and techniques applied — and the ability to describe the use case, application and resulting value.”
From this, we can deduce that an enterprise is suffering from low data literacy if employees:
- Are unable to accurately interpret reports and/or data visualizations.
- Cannot create effective reports and/or data visualizations.
- Fail to provide adequate statistical evidence for their claims.
- Succumb to statistical fallacies.
- Are unfamiliar with the data they’re sourcing, either semantically or architecturally.
- Are unproficient in the data analytics tools used by the company.
Of course, data literacy levels in an enterprise can be difficult to detect organically; most are discovered as the result of a blunder. Rather than wait for employees to make costly mistakes, enterprises can have employees complete a data literacy assessment.
Low enterprise data literacy is caused by a lack of education, both formal and informal. Not only is data misinterpretation rampant in the popular media, as data visualization expert Alberto Cairo demonstrates in Arguing with Charts, but many Americans find it challenging even distinguishing facts from opinions. A 2018 Pew Research survey asked over 5,000 U.S. adults to categorize five statements as either fact or opinion, and “roughly a quarter got most or all wrong.”
Many also conflate “data” with “fact,” treating statistics as incontrovertible when, as a colleague of mine recently put it, “Data is never perfect; it provides a story based on an incomplete set of information.” Revealing data as inherently limited is a critical first step in data literacy education.
On an institutional level, low data literacy results from a lack of leadership. Enterprises are responsible for their employees’ training and should therefore ensure that all are equipped to navigate the company’s data, data repositories, and analysis tools.
Misinformation leads to misinformed decisions. Low data literacy, therefore, can cost companies revenue and adversely affect customer outcomes, further impacting the bottom line. In higher-stakes scenarios, it can even be hazardous.
As mentioned before, it can also result in BI-equipped SaaS products seeing low user adoption rates and/or low customer satisfaction where reporting and analytics are concerned.
According to the Harvard Business Review, “the responsibility [of teaching data literacy] has shifted from academic institutions to employers, where skills development programs are flourishing.” SaaS vendors offering BI should encourage their customer companies to take one or all of the following steps toward boosting data literacy rates among their ranks:
- Teach the tools. This is something SaaS vendors can (and do) directly impact. By providing BI training materials, programs, and sessions virtually and at live events such as user conferences, BI-capable SaaS products can ensure that their applications are properly used.
- Teach the data. SaaS companies can have an impact here as well. In addition to briefing BI trainees on how their data is organized, SaaS providers may provide data documentation for ongoing reference. Data dictionaries help BI users navigate data models, find related data, and combine fields in common calculations. Business glossaries explain what the data means semantically, and the two combined can profoundly improve literacy around proprietary data.
- Teach the fallacies. Even if SaaS providers don’t furnish this education themselves, they can provide the resources necessary for customer companies to train themselves. In any case, an awareness of statistical fallacies is critical to avoiding them when setting policies, designing campaigns, authoring reports, and interpreting reports. The more directly these fallacies are related to real business scenarios, the more effective the training will be.
- Encourage peer review. Editing is as critical to writing reports as it is to prose. Data visualizations, dashboards, and tabular reports should be critiqued by a peer or superior before they are published and distributed. SaaS providers can pass along report and dashboard peer review guidelines for customer companies to follow and/or practice peer editing as part of a user conference statistical fallacy training session. Some SaaS vendors may even opt to provide report editing as a billable service.
By taking an active role in customers’ data literacy, SaaS providers help ensure the success of their BI offerings while boosting customer satisfaction.
For links to online data literacy training courses, check out a recent episode of Data Talks with instructors Caitlin Johnson and David Langer.
Originally published with Software Business Growth.