The characteristics of leading state organizations
What do leading organizations do that's different from the rest? AFP took up this question in the FinNext Asia session entitled, “The Roadmap for FP&A Transformation.” Larysa Melnychuk, founder and CEO of FP&A Trends, laid out an opportunity for FP&A to reclaim and repurpose time; Michael Coveney showed what change looks like; and Takeshi Murakami shared a case study of that change at Microsoft.
Melnychuk began with a familiar refrain from FP&A Trends research: 40% of organizations still report data of poor quality; 74% use spreadsheets intensively for planning (more than three-quarters of their process), and only 12% are using cloud technology. Optimism for efficiency is on the horizon, however, as 11% have started using artificial intelligence (AI) and machine learning (ML) for planning, with 57% of organizations planning to do so in the near future.
The characteristics of leading state organizations
Today’s typical organization has a 12-month focus when it comes to setting a budget and tends to have separate strategic business and operational plans. “And those practices work okay, but once it's outside of the planning cycle that we're used to, all sorts of challenges present themselves,” said Coveney, head of research at FP&A Trends and author of Budgeting, Forecasting and Planning in Uncertain Times.
Coveney cited six areas in which leading state organizations differ from the pack. The first is leadership. Organizations in the leading state recognize they can't predict the future, and therefore change the cultural mindset to focus on answering the following questions: How do we change what we're doing at the moment? What are the analytics telling us? What are the facts telling us? And how do we adapt?
The second difference is that leading state organizations have integrated processes — the strategic plan, business plan, and different operational plans are all aligned and combined as one. This requires FP&A to spend time with the different operations rather than create a process that simply changes numbers in a cell.
In leading organizations, “What we find is that typically FP&A will work alongside, whether it be logistics, manufacturing or HR, and they will talk with those organizations, asking what information they need to make decisions at a detailed level. And quite often, what happens is we end up building models that support those needs. But those models are directly linked to the overall business plan, which is directly linked into the strategic plan. We end up with one plan,” said Coveney.
The third area is in the use of data and analytics. Leading organizations typically have a driver-based plan, which allows them to simply change a few of the drivers to create a complete P&L, balance sheet and cash flow forecasts for the organization. There are several benefits to this simplification. You only have to change a few numbers for this model to work, which simplifies maintenance and effort to update the model.
Also, a driver-based plan enables these organizations to run scenarios and sensitivities. What happens if something changes, for example, whether an exchange rate changes or whether supplies suddenly reduced by 10%? What does that mean for the business? The system can then work out the implications for the organization as a whole and from that, plans can be formulated to either mitigate or take advantage of the situation.
The fourth area relates directly to the third: leading companies have technology and systems that run in real time. “You can't do this in spreadsheets, you can't do this in simple consolidation tools. It requires a proper platform that is collaborative [so that] the whole organization is planning together,” Coveney said. That also becomes the springboard for the next round of advancements, including AI and ML capabilities.
“We're seeing a tremendous increase in the use of these technologies, and it's having a big impact on accuracy,” said Coveney.
The last two items return to human elements for FP&A. Fifth is to lean into the elements of being a business partner and trusted advisor. FP&A needs to challenge the organization, asking, "Do those plans really make sense? Is that forecast really accurate? Do the facts support where are going?” This is a delicate balance act: supporting the operating teams while being skeptical and asking difficult questions.
This leads to sixth item, the creation of new FP&A skills to accomplish all the preceding steps: To do this requires three roles to be filled: the architect, the storyteller and the influencer. The architect will apply their financial expertise to the data models and technology platforms to see what it means for the business. Then the storyteller will take that information and put it together in such a way that sorts out the key signals within the avalanche of noise surrounding the organization and take meaningful action. It is the sum total of these efforts that allow finance to have the technical, informational and persuasive clout to influence the organization.
The impact of leading state organizations is impressive: 84% of them base all or most of their decisions on data, 71% have a real-time view of business performance for the C-suite, and their forecasts are accurate 83% of the time, compared to an average of 50%.
How do you get to that leading state? “It happens one step at a time,” said Coveney.
Success requires addressing all six areas. Some companies will be at a mature level when it comes to leadership, but because analytics is letting them down, they're back at the basic state. Culture is a hard area to address and requires years of experience and persistent effort from leadership. Analytics is the easiest element to see, and while you can change that quickly, you need progress on all levels. The six elements should provide you a roadmap of where to work.
Case study: Microsoft
Our business environment is changing dramatically. Smart devices, smart office, autonomous vehicle…new environments are boosting in creating tons of data every day. Our challenges vary from surging data, legacy systems, inadequate tools, not able to cope with the growing business complexity, which will eventually lead to human errors, and even generate bigger problems like a governance risk and other threats. To overcome these challenges, Murakami presented how Microsoft has instituted four areas of focus.
- Financial analysis and reporting: using business intelligence tools to automate reporting, increase agility in the business decision-making process, and move away from Excel.
- Strategy and forecasting: developing AI and ML to forecast capacity and demand planning.
- Business of process automation: includes leveraging a chat bot for basic Q&As to improve agility and efficiency, automatically setting up contracts and tax reports.
- Risk management: One example of this is compliance of predictive analytics; they capture the unusual transaction automatically by AI, flag it and raise the alert.
Murakami stated that while accuracy is very important, what is also important to take note of is the amount of time saved. To create a quality forecast the “traditional way” took the team “roughly two to three weeks” and involved a significant number of people and many sync meetings to finalize the numbers and storylines. “It’s a very time-consuming process,” he said. “It also involves the judging up and down of the numbers, meaning it is biased. If you leverage and shift 100% to ML, we'll need just one or two people, and it can be created almost in real time and without any bias.”
“In Japan, we used to have like 60 people and spend a huge amount of time doing the forecast,” Murakami said. “But now we're able to free up 58 of those people and let them work on the more value-added tasks. This is happening worldwide in Microsoft.”
How do you turn the data into action in the quickest way possible? By cleaning up the data and creating a single source of truth. “It is critical that we work on cleansing the data, creating consistent definition taxonomies,” said Murakami.
“Machine learning enables us to learn from the past to predict the future,” he said. “The distance between data and action is much closer now.”
Turning the data into action and, more importantly, adding value to the business in a fast and relevant way has been a long journey for Microsoft. “And we're still struggling to get this improved further,” said Murakami. Making the change requires a strong leader to drive it and a process to run it.
Here is how it works: The ML models join the expertise of data scientists with the application of the finance team. Our data scientists built a ML model where we use the customer view data from our CRM and historical close rates to each of the different stages of the deal and pipeline. A ML model would apply the close rate to the current pipeline to be the forecast while applying trend projections and statistical (probabilistic) methods. There are about 20 to 30 models in the tool, and the final output of the forecast from the ML will triangulate among four or five of these logics. The AI works on top of ML to pick the most relevant logic for the country.
This modeling was done centrally at the headquarters in the Washington, DC (in the U.S). We learned that a finance professional and a data scientists need to closely work together to fine tune and improve the model. The finance person is not the professional in the statistics or the machinery model, and the data scientist is not professional in the business. So, the key important thing is to work together to improve the model. And once we get the final output, we can either transition the forecast into Excel or a BI tool to share with the business.
Just how big of a role of finance can really play in driving the business is an open-ended question. “I'm really excited about all the great changes coming for finance professionals,” said Murakami.
Did you miss this AFP FinNext Asia session? Watch the full recording.