In one of the all-time movie classics, “Back to the Future,” teenager Marty McFly (played by Michael J. Fox) is thrust accidentally 30 years into in the past and seeks advice on how to get back to his present-day life. In one scene, Doc Brown gives prescriptive advice to Marty as he gets into the DeLorean car about to speed down the street: “Don’t worry. As long as you hit that wire with the connecting hook at precisely 88 miles an hour, the instant the lightning strikes the tower … everything will be fine!”
The parameters of the advice had all the signs of being too difficult to lead to the desired outcome. But Marty trusted Doc Brown. It reminds me a little of what’s happening today in the area of prescriptive business intelligence. Companies must trust the source data for prescriptive BI in the context of the decision to be made (context being the key word). And a modern decision management system must have flexibility.
I tossed a question out recently to the tribe on Twitter at one of my Friday #BIWisdom tweetchats. “As we all know, the business intelligence market constantly seeks higher ground, and prescriptive analytics certainly qualifies. How different is prescriptive analytics from decision management?” And how does it differ from predictive analytics?
The BI group tweeted several responses:
- Prescriptive is closely tied to predictive outcomes while decision management isn’t necessarily tied to the outcomes.
- Predictive analysis precedes the prescriptive, so it’s more far reaching.
- Predictive anticipates what, when and why. Prescriptive suggests decision options and shows implications of each decision.
- Predictive modeling sits mostly in the development side of BI. Prescriptive needs business process owners to do the modeling.
- Prescriptive BI forces you to go through the evolution from descriptive to predictive. Decision management in general does not.
- Prescriptive provides courses of actions and suggests options to situations.
- Prescriptive analytics allows organizations to manage “prophylactically” rather than reactively.
Nearly everyone in the group agreed that companies use prescriptive BI to generate suggestions based on a set of rules. But the company must predefine the rules, and the rules must be flexible to allow considering current outcomes versus when the company created the rules. Someone also observed in a tweet that “One rule begets the next rule; in reality, it is iterative.”
Prescriptive BI’s challenges
As one of the #BIWisdom tribe pointed out, prescriptive BI is intrinsic, and the typical organization is large with complex policies, politics and other organizational challenges.
In addition, prescriptive analytics works best in the context of well-established processes, but challenges arise when a company’s processes are still too manual to get the full benefit.
Data governance
Finally, prescriptive BI requires effective data governance. “Of course data governance is foundational to BI, but many companies today have BI and still have no data governance,” tweeted one of the group.
I asked the tribe who has implemented prescriptive BI successfully. One person tweeted, “I’d expect that successful prescriptive implementations will stay hidden for a while to preserve competitive advantage.” Another responded that he was working on a large program at the moment and all issues traced back to data governance.
Bottom line: According to the data in our 2014 Wisdom of Crowds® Business Intelligence Market Study, 30 percent or fewer of 1,283 respondents indicated their companies are doing any predictive modeling. It’s likely that five percent or fewer are attempting prescriptive BI. So the curtain is just beginning to rise for prescriptive.
I believe the success or failure scenes that will play out on the prescriptive BI stage will depend in great part on whether organizations first put in place the required foundational data governance.
It’s never too early or too late to plan a solid BI strategy. To assess whether your organization has an effective data governance program, here are three key signs that data governance is lacking, agreed on by the #BIWisdom group:
- Continual failure to really benefit from investment in BI.
- Inability to cope with unexpected situations and inflexibility of processes due to problems that unexpected situations cause.
- The proliferation of shadow IT. Data is in the hands of users without oversight.
When there are too many organizational-level issues, companies often blame data governance for the disruption. But implementing an effective program is the good kind of disruption. And as I frequently point out, leadership understanding and involvement is the key.
Howard Dresner is president, founder and chief research officer at Dresner Advisory Services, LLC, an independent advisory firm. He is one of the foremost thought leaders in Business Intelligence and Performance Management, having coined the term “Business Intelligence” in 1989. He has published two books on the subject, The Performance Management Revolution — Business Results through Insight and Action, and Profiles in Performance — Business Intelligence Journeys and the Roadmap for Change. He hosts a weekly tweet chat (#BIWisdom) on Twitter each Friday. Prior to Dresner Advisory Services, Howard served as chief strategy officer at Hyperion Solutions and was a research fellow at Gartner, where he led its Business Intelligence research practice for 13 years.