By Simon Bittlestone, CEO of financial analytics firm Metapraxis
“Data science” is becoming a bit of a buzz-phrase in today’s business world.
Packaged up as the golden bullet for an organisation’s forecasting and insight challenge, it’s a potentially enticing option for finance functions, who are feeling the strain of increasing demands from their investor, board and management stakeholders, to prove that they are effectively planning for the future.
As a result, many businesses find themselves jumping in at the data deep end, investing in data lakes and hiring expensive teams to “mine” them. But what happens when it’s years down the line, and the finance function is no closer to hitting gold?
WANT TO BUILD A FINANCIAL EMPIRE?
Subscribe to the Global Banking & Finance Review Newsletter for FREE Get Access to Exclusive Reports to Save Time & Money
By using this form you agree with the storage and handling of your data by this website. We Will Not Spam, Rent, or Sell Your Information.
Going back to basics
In our rush to mine data, we are losing sight of that fundamental concept that has guided empirical study for hundreds of years – the hypothesis. Too often teams of data scientists are picking up their proverbial shovels and uncovering correlations in data that are interesting, perhaps, but more importantly, irrelevant, and even worse, obvious. Only recently one of my own colleagues told me that they had uncovered an “exciting” correlation in the data of a major US retailer, which was between store square footage and, wait for it, sales volumes… needless to say, he had his shovel confiscated.
Understanding what data matters has never been more important. To do this the finance function must start with the business model, not the data. Financial results are essentially downstream outcomes of operational decisions. By focusing on what actually drives value for their organisation, FDs will, in turn, establish a clear connection between the operational or market based non-financial indicators and the financial results. Whilst the data itself is of course important, it is this first step of identifying what data matters that will ensure data-driven success for the business.
Unifying the organisational silos
This driver model can help hugely in breaking down the organisational silos that exist across business functions. Too often around the boardroom the FD is left fuming at the worsening operating cash flow, while the CEO sits comfortably in the knowledge that they have a positive profit story for the market and the Head of Sales sports an arm full of Rolexes after landing the largest ever contract on a sign now, pay later sweetener. Clearly, operating cash flow is influenced not only by profitability, but also by net asset movement – which itself is driven by capex and working capital movement. Put all these into a simply presented diagram and you have a much greater chance of unifying the management discussion on what needs to be done to improve the financial performance.
Having determined the data that matters, applying predictive and prescriptive analytics to those data sets to help answer the questions “What is likely to happen?” and “What should we do about it?” can provide enormous value to an organisation. However, here it is vital to remember that analysis is only one step in the business performance cycle. Results need to be measured and they need to be analysed, but if that analysis is not communicated effectively, the dialogue required to generate commitment to a new course of action will not take place and results won’t improve. The information matters, but the way that it is presented to decision makers is just as key.
Taking advantage of technology
That is not to say that success is that simple. Finance still faces an enormous challenge in achieving all this, in the form of legacy systems. Data is typically disparately located, inconsistently tagged and occasionally even lacking entirely. Investment in data sources is important, but before you dive head into a data warehouse programme, it is important to ask three things:
- What is our most important data?
- What business value will this data warehouse provide?
- What happens when our business changes structure?
More often than not data warehouse programmes don’t succeed, generally because they don’t focus on the data that is really important to the organisation, understand what they will do with that data when it is neatly stored, or lack the flexibility to evolve with the business.
Finance also has a guilty pleasure that doesn’t help: Microsoft Excel. Excel is not designed to run enterprise planning, analysis and reporting processes.Its great strength – absolute flexibility – is here a major weakness, and it is limited in the dimensionality needed to adequately reflect the complexities of an organisation. Finance must embrace the category of financial analytics which is designed for exactly this challenge and which can provide a platform for planning, analysis and reporting, to avoid one Excel formula error bringing down an entire business.
Finance has a key and exciting role in providing the right information for decision making in a world where businesses are becoming more complex and markets more uncertain. Data is important, but if it is the new gold, then finance analysts that can model their business, understand how to apply data science techniques appropriately and communicate effectively are worth their weight in it.