In our last blog article Putting the “A” back into FP&A: Becoming a trusted provider of business insight, we discussed how the role of finance professionals was evolving. Once seen primarily as providers of information about past results, finance professionals are now becoming essential deliverers of business insights that help management with forward-looking decision making. 

Analytics have played a key role in this evolution. Valuable business insights lie hidden in the enormous volumes of data that businesses collect. The challenge, of course, is digging it out. Today we’ll discuss how finance teams—and other business professionals as well—can use data discovery and exploratory analytics to extract insights from all that data and share it with those who need it. 

To get useful insight, data needs to be examined in all its dimensions, such as time period, geography, product, customer, sales channel, etc. It’s equally important to consider how the different dimensions relate to each other, reflecting factors such as seasonality, regional preferences and more. Dimensionality is required to accurately model the business structures, rules and relationships of the business.  

A recent study by the analysts at FSN revealed that more granular models help finance teams create more accurate revenue forecasts. The study noted that “Technology has enabled enormous strides in PBF [planning, budgeting and forecasting], which means it is now feasible to build much more granular and complex business models...”1 

How does technology help with the analysis? Let’s consider a common business question: how to identify the predictors of wins or losses from an opportunity sales pipeline.  Sales opportunity data can be pulled from the sales force automation system, along with data about sales methods and practices, and data about the sales representatives themselves, pulled from the HR system, including tenure, levels of technical training, and more. 

Ideally, the system would first analyze the quality of the data set – identifying duplicate data, empty fields, outliers and so forth – to ensure the data is robust enough for statistically significant analysis.  Statistical modeling techniques could then be used to uncover patterns from the data.  In this example, a technique known as decision tree modeling can uncover predictors for the likelihood of whether a sales opportunity will result in a win or a loss.  

In the diagram below, the system discovers correlations in answer to the question “Which fields influence sales?” The system discovers that the level of technical training of sales reps, compensation, years of sales experience, number of client visits and average call length are all predictors of average sales. We see that years of sales experience has a 60% predictive accuracy from the callout in the center.




In addition to decision tree modeling, a number of other modeling techniques are available. To create forecasts from time-series data, there exists a whole range of statistical modeling techniques from simple linear extrapolation to complex regression techniques. The idea is that given a sufficient amount of historical data and indicators of probability, a statistical model can generate a highly reliable forecast.  Monte Carlo simulation is another approach used for range and scenario planning, which enables analysts to incorporate risk factors explicitly into their plans. 

So, which tools are best for performing these kinds of exploratory analytics? Spreadsheets may be the default choice for most finance people, but they offer limited data management and workflow capabilities, and little ability to handle large volumes of dimensionally rich data. Spreadsheets, however, can serve as a user interface within a modern planning, analysis and forecasting system. That way, users can enjoy automated enterprise workflow and process management, along with robust multi-dimensional modelling, integrated reporting and analysis capabilities, all within the context of a familiar interface. 

To learn how this type of analysis works in the real world, watch this video  about a Brazilian company that is a market leader in the fast moving, highly competitive world of cosmetics. Learn how this company uses sophisticated data analysis to track style trends and provide its customers what they want, when they want it.



1 Gary Simon, The Future of Planning, Budgeting and Forecasting, Survey 2016, Insights from the FSN Modern Finance Forum on LinkedIn, FSN, 2016






  • 2017-02-06 13:09:43
  • Spencer Lin
  • Future of the Finance Function, Future of planning budgeting and forecasting, IBM