Nearly every financial planning and analysis professional understands the need to embrace big data and analytics. But how can firms move beyond basics to advanced analytics?
That challenge was on every attendee’s mind at the recent FP&A Advisory Board Meeting in London. There FP&A and business intelligence professionals gathered to discuss the state of BI and FP&A analytics with Nigel Geary, BI specialist at British Gas.
Key takeaways from the meeting:
- Cloud providers and Microsoft Power Pivot provide FP&A professionals more advanced capabilities to analyze data. It’s important that FP&A professionals train themselves to use these tools so that they can be self-sufficient and own the function and the models.
- To be more effective, FP&A professionals need greater insight and more flexible data and models. While there is plenty of data these days, it’s often hard to process. New computer languages like Hadoop are emerging to make handling large data sets faster and more efficient.
- Analytics, and the technology that supports it, typically is still driven by IT, not by finance, which creates a knowledge gap.
- There is a new era coming where users need to take back that capability in order to track, analyze and predict business performance.
“I want to understand what’s considered advanced analytics and what’s considered basic,” one attendee said. “I’m trying to understand what good organizations are doing with regard to applying best practices. To me, it feels that it’s in its infancy.”
Geary concurred. “What I see is very basic. Nobody has even thought about predictive forecasting,” he said. “That’s what the trend tells us—how a new product will affect sales and R&D; the next step is to take this together.”
For many companies, the first step into advanced analytics is constructing multidimensional decision support tools, commonly called “cubes.” In the past, these had to be created on mainframe computers. Now they can be created in the cloud, making them more affordable.
“[Cubes] allowed FP&A to build budgeting systems in their departments,” Geary said. “They also minimized the amount of IT support needed to run the models.”
Next analytic step: the hypercube
The next step in analytics, according to Geary, is the hypercube. Usually with budgeting, he said, FP&A pulls down data from the general ledger and creates with multidimensional data sets each month, refreshing the information and performing variance analysis. The hypercube breaks down this process to show users the profitability of different activities within the business. The hypercube ties all the ledgers together to calculate profitability by day, department, etc., and provide drill-down capability by numbers and transactions, and makes it possible to see any errors.
“You can’t do that if you have sales data in one cube and cost data in another cube,” Geary said. “The idea is to roll all of that up into a single cube bring it all into one place.”
To deal with the tidal wave of big data, data scientists have come up with a new data storage and language called Hadoop. Using Hadoop, companies can pull data out of new multidimensional databases that are not structured in the old columns and rows, providing greater flexibility.
The Practical impact
Intriguing as new advanced analytics tools may be, attendees noted that they often don’t have time to implement them. “In FP&A we talk about trends,” one participant said. “We talk about scenario planning: What would be our response in various cases. That’s a form of analytics.”
How advanced companies can be in implementing these emerging tools really depends on their current state of technological advancement. “If the planning is not great, they should be looking at the tools that perform scenario modeling and all basic tools today can give you that,” Geary said.
Companies can start by looking at historical patterns and the drivers of customers’ behavior to identify business drivers. Users should be able to look at reports in Excel and slice and dice the data to understand the business better. In the future, Geary predicted, companies will run more predictive models relying on bigger sets of data using new language like Hadoop.
One of the areas where advanced analytics plays a growing role is in integrated driver-base planning. Larysa Melnychuk, managing director of the FP&A Club, AFP, provided an example from her past work at a bank. “We used a ‘survival analysis’ technique for forecasting behavior of a complex portfolio of personal loans and credit cards,” she said. “This non-standard analysis—not typical for finance—helped generate business insights that significantly changed strategy on some of the products. It also helped us to improve profitability of the portfolio.”