Bridging the Gap between Prescriptive Analytics and Trust

Note: this post was written for and first appeared at DATAVERSITY: https://www.dataversity.net/bridging-the-gap-between-prescriptive-analytics-and-trust/

Prescriptive analytics can amplify the decision-making ability of data-driven organizations by examining what has happened and using that information to recommend – or even implement – next steps. But this power depends on how much users can trust the data being mined, and that gap is where opportunities can all too easily be lost.

Becoming a Data-Driven Organization

Most organizations use some form of descriptive analytics, which organizes the data so it can be used to look at events and understand what happened and why. Predictive analytics goes further to spell out what is likely to happen based on what has already occurred. It can be applied to everything from lead generation and customer outreach to cost management and whether or not to expand into new markets.

We see this at work in multiple industries and scenarios. For example, marketing and advertising used to be driven by instinct and a sense of what is creative and appealing. Today, marketing is centered on measurement and performance analysis, which drives everything from where and when to promote a message to what format and content are most likely to succeed. This is why websites and brick-and-mortar companies capture so much data about searches, locations, habits, and purchases.

Prescriptive analytics takes that automation to the next logical level: If we can predict what the probable outcomes will be, why not use AI to suggest what to do next in order to reach a business goal? This can apply to big-picture tactics and micro-decisions that impact everyone’s ability to move forward. The capability is so compelling that, while 11 percent of large and midsize organizations use some form of prescriptive analytics today, Gartner expects that by 2022, this will grow to 37 percent.

The Trust Gap

Automating business decisions sounds great. But it only works if the data is trusted by the people who need to use it, which is dependent on several factors:

• Source: Do we trust where it came from (internally or externally)?
• Completeness: Are we missing anything important? Is the data up to the minute?
• Accuracy: Is the information reliable? Is the model believable?
• Relevance: Does the information have anything to do with the business or the decision we’re trying to make?

So, where is the trust gap coming from? For one thing, the Digital Analytics Association estimates that 44 percent of analytics teams spend more than half their time accessing and preparing data instead of analyzing data. This means it might miss the mark on all counts until the data is reviewed, cleaned, and sorted.

Data silos within the organization can also lead to incomplete information. Part of the issue there may be technical, as legacy systems may not be able to share data effectively with the rest of the organization. But, unfortunately, there’s a cultural component at work here, too.

A crucial part of being a data-driven organization is that everyone must be willing to share information. Some still take the maxim, “knowledge is power,” as a rationale to keep some data to themselves. Creating a data culture is as important as having the right technology, which can only happen if leadership embraces the idea.

Data vs. Gut Instinct

The thought of relying on machines to make business decisions is unsettling to many in leadership roles. Instead, they continue to operate on instinct – sometimes, flying in the face of what the data says.

A data-driven organization reduces risk by informing decisions through rational analysis; it doesn’t eliminate the human factor. Leadership still gets to determine the organizational goals as well as the parameters of the decision. The people who are ultimately responsible for the outcomes get to choose which actions will be automated.

But here’s the risk: opportunities could be lost by not trusting the data. In the time it takes for people to gather and analyze information, form committees, and conclude, the market may have shifted, prices and availability may have changed, or a competitor may have taken the lead. Automation can provide faster, more accurate decisions that deliver real, measurable ROI.

Prescriptive analytics doesn’t have to mean that all decision-making is turned over to the machines. But it can provide unbiased, rational conclusions vital to the organization’s health and growth. Because humans are involved, building a culture that leads to trust will take time. Leadership’s responsibility is to live it and support it throughout the enterprise. For data-driven organizations, making that leap – from descriptive to predictive to, finally, prescriptive analytics – can bring a huge competitive advantage.