Advanced analytics drive better business decisions
Part 1 of 2
Advanced analytics can guide the way to better decision-making—enabling organizations to increase revenue, decrease costs, become regulatory-compliant and so on. For example, as noted in its video, Sprint uses advanced analytics to “put real-time intelligence and control back into the network, driving a 90 percent increase in capacity.”
Realizing these goals requires an emphasis on the analytics conversation around three strategic pillars: focusing on business decisions, embracing an agile culture and investing in an ecosystem of talent. Based on continuous research by many leading firms and various advanced analytics consulting projects at IBM, this article offers a few key takeaways that address the complex questions posed in these three categories.
Focusing on business decisions
What does the key decisions matrix of your organization look like (short-, mid- and long-term versus strategic, tactical and operational)? Which decisions need improvement? What is the value of a better decision in each of these cases?
To embed analytics into strategic, tactical and operational processes, we must first define three core elements of decision-making:
- Context and cognitive biases: In an article from the Harvard Business Review, David J. Snowden of Cognitive Edge talks about a framework he helped develop called Cynefin (pronounced ku-nev-in). Government leaders and private companies around the world use it to understand the context in which their organizations operate and how the context relates to the decision-making process. He defines five categories of context (simple, complicated, complex, chaotic and disorder), but highlights the “complex” category because it is the reality of most business leaders today. For example, for the scheduling operation of a liquid pipeline, one might question if a manager should accept a late order from an important customer. Prescriptive analytics can optimize the trade-offs among all of the objective functions (that is, business priorities) based on rules and recommend rejecting or accepting the order. Daniel Kahneman, an expert in the area of cognitive biases and a Nobel Prize winner for his work in economics, describes the 12-question Decision Quality Control checklist in another Harvard Business Review article. It helps “unearth defects in thinking—in other words, the cognitive biases of the teams making recommendations.” These questions are categorized into three groups: bias of the decision maker, bias of the proposed action and the value of the proposal. Simulation and predictive analytics can create scenarios that are objective in nature to counter some of the inherent biases in decision-making. For example, a drug distributor company applied predictive analytics to reduce inventory cost. After iterating the models several times, the company concluded it is cheaper to fly high-value, low-frequency drugs from a central location than to store safety stock at individual distribution centers.
- Data and information: Based on years of trial and error with a number of internal analytics projects at IBM, the Institute for Business Value recently published a point of view on how analytics can most effectively drive business outcomes. One of the critical success factors is that instead of searching for perfect data, companies should focus on universally accepted and consistently used data. For example, a pipeline tool maintained by IBM sellers was used to measure baseline cycle time. IBM estimated an accurate impact to win-rate and revenue improvement; initial results matched projections almost exactly. This was very important for ensuring end-user buy-in and enabling actions based on the analytics output.
- Decision process and accountability: Michael Mankins and Lori Sherer of Bain & Company argue in the Harvard Business Review that advanced analytics can automate parts of the decision-making process (collecting facts > alternative courses of action > selection by decision-maker) by codifying the best decision-maker’s logic and taking the variances out. In addition to automation, systemizing “expert knowledge” into mathematical algorithms has another profound benefit: knowledge retention and documentation. This is especially important as more baby boomers retire over the next couple of years while creating significant knowledge gaps within organizations. This alone has been the fundamental component of the business case for many advanced analytics projects in Canada. According to a Bain & Company survey of executives worldwide from 760 companies, “decision effectiveness and financial results correlated at a 95 percent confidence level or higher for every country, industry, and company size.” Conducting decision audits, creating an open culture, aligning roles and responsibilities, incentives and organizational structure around disciplined decision-making can increase the execution effectiveness.
In the conclusion of this two-part series, we’ll look at the other two pillars of the analytics conversation: embracing an agile culture and investing in an ecosystem of talent.
To start your advanced analytics journey today, please visit the IBM advanced analytics resource page.