Add Data Science Skills to Corporate Finance Decisions

Every hour, every minute, every second, organizations capture and consume massive amounts of data used to make strategic and tactical decisions. Yet, as the volume of data grows, few finance teams are using that data to provide business partners with actionable insights. Those that do create significant value for the business by using data science skills to predict a series of outcomes over time and use those predictions to identify the best course of action.

“Data science and machine learning (ML) provide the means for organizations to predict a future that looks uncertain,” says Carlie J. Idoine, Senior Director Analyst, Gartner. “They provide prediction insights to drive action, but the appropriate actions are neither obvious nor easy to sort out.” 

Business rules can be complicated given competing resources, numerous constraints and continually changing variables. This makes it hard to evaluate all the alternatives and quickly decide on the optimal actions for achieving business objectives using an easily repeatable approach. Finance leaders should bring together two distinct but complementary data science and ML techniques to drive high-impact decisions. 

Predictive versus prescriptive

Combining predictive and prescriptive capabilities is a key first step in solving business problems. 

Predictive analytics addresses the question of “What is likely to happen?” It relies on techniques such as predictive modeling, regression analysis, forecasting, multivariate statistics and pattern matching.

Prescriptive analytics addresses the questions of “What should be done?” and “What can we do to make “X” happen?” This technique includes graph analysis, simulation, complex-event processing, recommendation engines, heuristics and, increasingly, neural networks and machine learning.

Bringing together forecasts (predictive) with action (prescriptive) enables an organization to explore how changes to different variables are likely to affect the outcomes or alter the relative trade-off. “This combined approach gets to the heart of the task of adding business value — of proactively making decisions that drive action and influence the future course of an organization,” says Idoine.

The nature of the problem guides CFOs’ choice of whether to use prediction, forecasting or simulation for the predictive analysis component

For some industries — like manufacturing, transportation and financial services — predictive and prescriptive modeling have long been part of their processes. However, even with the known power of these capabilities, they are often isolated and trapped in specific, dedicated departments instead of being used in tandem with others and leveraged across the entire organization.

In the age of digital business, flexibility, agility, collaboration and responsiveness are key. In effect, the ability to predict outcomes, quickly assess innumerable alternatives and take action is something that can no longer be reserved for the few.

Select the optimal mix

One approach is to use a predefined framework or rules for choosing between alternatives. The other approach is to use an outcome-driven, constraint-based evaluation of an interdependent set of options or optimizations. The resulting actions can be recommended to human decision makers, used for decision support or fully automated as part of a decision management system

The nature of the problem guides CFOs’ choice of whether to use prediction, forecasting or simulation for the predictive analysis component. The complexity of the solution guides their choice of whether to use rules or optimization for the prescriptive analysis component.

Build in-house skills

To jump into predictive and prescriptive analytics, CFOs need to develop their team members’ skills to improve decision making. Gartner has identified three ways finance leaders can develop and expand their teams’ in-house skills for predictive and prescriptive analytics.

1. Change hiring priorities. Organizations that take on data science often start with hiring a data scientist, one of the most in-demand jobs of the 21st century. While not invalid, this can be difficult and come with a hefty price tag. Many organizations already have people who operate in a data-scientist-type role, even if they don’t have that title, who are often specialists isolated in individual departments or workgroups. To expand the use of predictive and prescriptive analytics more consistently, finance leaders should find ways to leverage these existing resources throughout their organization before they look to outsource.
2. Tap into the business citizen community. A flood of accessible, user-friendly data science products has enabled employees without data science degrees to obtain advanced analytical capabilities. More products now offer detailed analytical guides, or augmented analytics, that can be easily understood by business analysts. These “power users” or “citizen data scientists” are typically positioned well enough to take on the role of citizen data scientist. Although this role cannot replace an accredited data scientist, it can complement and extend an organization’s analytics universe.
3. Partner with experienced vendors. In some cases, organizations can partner with outside consulting groups or software providers to spearhead an analytics initiative. For organizations that are new to analytics, and data science specifically, or with no in-house skills to build on, this is a model approach. Enterprises in this situation should first look to providers with not just predictive and prescriptive proficiency, but that also possess knowledge of their particular industry. Leaders should also ensure the partnership model provides for transfer of skills and lessons, and be cautious of the potential loss of intellectual property with this approach.

This article has been updated from the original, published on December 15, 2017, to reflect new events, conditions or research.