5 Steps To Data-Driven Innovation in Finance

Organizations that analyze all relevant data and deliver actionable information stand to achieve $430 billion in productivity benefits by 2020 over companies that are not as data savvy, according to IDC.

Of course, analyzing all relevant data is a tall order because the data that's considered relevant today will very likely change, while new data will become pertinent as machine learning technologies come into play.

So where should CFOs and other business leaders focus? Rich Clayton, vice president of Oracle's business analytics product group, recommends five steps for data-driven innovation.

1. Automate Business Narratives
   
   As companies generate ever more data, executives don't have the context for all of this new information, which has resulted in increased demand for narrative explanations, says Clayton.
   
   "Knowing what happened is no longer enough. Understanding why it happened is critical—and it's the explanatory aspect of analysis that's actually most valuable," Clayton says. "Unfortunately, most processes to develop that narrative are very labor-intensive and manual."
   
   With people from different departments involved in writing these explanations, using tools like email and PowerPoint make version control a nightmare—and can affect the validity of the report.
   
   "We've always struggled with automating the integration of the narrative with the numbers and the facts," Clayton says. "With the right tool and processes to manage the creation of these reports, you can have more people contributing their expertise, which would increase the quality of the narrative and ultimately simplify the entire process."
   
2. Connect Plans
   
   Like the tale of five blind men who touch an elephant yet reach different conclusions about the experience, bias in business can be both true yet completely off-base.
   
   Every line of business, including finance, brings its own biases. "When you include more people in the discussion, you actually are de-risking the business forecast and plan," Clayton says. "You need an agile process and tools that enable you to create scenarios that are fully baked, with data and input from across the company."
   
   For example, a manufacturer whose systems for procurement, sales, finance, and human resources aren't linked together won't be able to plan for scenarios in which its new product becomes the must-have gadget of the season—or an earthquake disables its manufacturing plant.
   
   "Companies need systems that connect assumptions about all of these components that ultimately flow into a financial plan that goes through balance sheet and cash flow analysis," Clayton says. "Making those connections using spreadsheets is difficult."
   
   The more that lines of business can see how their assumptions affect the company's ability to execute, the more they'll buy in to achieve those goals, he says.
3. Beautify Your Insights
   
   Finance people are data people, but a 20-page spreadsheet is not meaningful to many employees.
   
   "Data needs to be made understandable, and rows and rows of numbers aren't quickly understandable for most people," Clayton says. "But by making insights visual and creating a story, people are more likely to learn and remember—and they're more likely to improve if they can understand. "
   
   Finance professionals need to go the final mile to make their analysis consumable—not just by making it visual, but by putting it into a story that people can follow. "Consider it the 'absentee test,'" he said. "If you got sick and weren't there to explain your analysis, could someone read it and be able to take action?"
   
   These visuals should incorporate analysis and present variances, enabling the reader to easily understand the options and the implications of each option.
   
   Unlike narrative reporting, where many people are creating and publishing a report, this kind of reporting involves few authors and many consumers and will increasingly tap machine learning technology, Clayton explains.
4. Reduce Bias in Decisions
   
   Automation lets companies amass data faster, and access to more data enables more informed business decisions. However, the quality of those data insights and human judgements will be reduced if the organization operates in silos.
   
   For example, assessing the customer experience involves analyzing data from across the company—from marketing to sales to customer service—as well as external data. If data can't be easily accessed from each of these areas, the company will continue operating with disjointed processes and strategies that may meet departmental objectives but lose customers.
   
   "Machines will increasingly augment our decision-making capabilities and enable us to de-bias our decisions—and that can lead to the creation of new products, services, and value we couldn't imagine before," Clayton says. "As business applications become more proactive, looking for and understanding patterns in the data from across the organization around the clock, companies need to have the right systems, processes, and culture to take advantage."
5. Create a Data Lab
   
   If you're not experimenting with data, you're falling behind, Clayton says. Modern tools make it easier for finance to experiment with a range of scenarios and business models and, as machine learning technologies are incorporated, test new algorithms to assess risks before they're are put into production.
   
   Such experimentation is what it takes to discover viable commercial uses for company data and should take place in a data lab, Clayton recommends.
   
   Data labs aren't the same as data warehouses or data lakes. "It's a platform where you can bring together sets of data, conduct tests, and use machine learning to identify hidden patterns," Clayton says.
   
   A data lab with the right analytics tools, internal and external data, and expertise—including analysts, subject matter experts, engineers, and scientists—can connect seemingly unrelated dots to potentially boost revenue and point to new business strategies.