Understanding Prescriptive vs. Predictive Analytics
It’s hard to overemphasize the importance of an FP&A team, their decision-making and tools to a business’ success. Which of prescriptive vs. predictive analytics should an FP&A team rely on to understand the financial reality of the organization, uncover the reasons for it, and optimize recommendations to ensure future success? Read on.
An Actionable Analytics Outline
Before contrasting predictive vs. prescriptive analytics, it’s important to understand both their role in an organization to start with, and also how they compare to other key forms of analytics. Why are analytics so vital to an FP&A team in the first place?
Analytics are a dominant method in how financial departments understand what’s going on in their organization, why it’s like that, and what should be done about it. Without this, it’s impossible to make valuable, data-driven recommendations to guide the company towards the decisions most likely to achieve its goals.
Gartner outlines four primary types of analytics in this context:
- Descriptive analytics
- Diagnostic analytics
- Prescriptive analytics
- Predictive analytics
Descriptive analytics and diagnostic analytics both focus on the past. They’re employed when you want to know what has already happened (descriptive analytics) and why it happened (diagnostic analytics). When you’re conducting a quarterly or annual review analyzing financial results, or evaluating budget versus actual variance, you’re working within this realm.
Predictive analytics and prescriptive analytics, by contrast, are focused on the future. Using historical and current data, predictive analytics makes predictions (hence the name) about what might happen next. Prescriptive analytics is used to make recommendations about how you can best achieve a desired outcome.
Predictive vs. Prescriptive Analytics
Predictive and prescriptive analytics have a lot in common. Both rely on machine learning models and data science, both require large amounts of data to be accurate, and both are sensitive to the “garbage in, garbage out” caution.
This means that for either form of analytics to be effective, it’s essential to ensure that their calculations are based on reliable, clean, balanced data. In an organizational context, it’s also important to make sure that data sets are not fragmented, skewed or partial due to data silos, inconsistencies, and similar issues.
Predictive and prescriptive analytics are also frequently used by FP&A teams as key elements in the process of forecasting and recommending action. Predictive analytics is what you’ll turn to for forecasting, while predictive analytics is what is needed for recommendations.
- Predictive analytics is a probabilistic calculation. Taking into account diverse and often complex factors regarding a range of data inputs, how trends may evolve, and different kinds of scenarios (e.g. conservative, standard, optimistic) the model will give a result reflecting how likely various possible outcomes or trends are, based on that information.
- Prescriptive analytics is action-focused. Rather than trying to predict the future, the aim is to work out how to get to the version of the future that you want. To be effective, it’s important to be asking the right questions, which means being firmly grounded in a business’ strategic aims.
Predictive vs. Prescriptive Analytics: Shared Foundations
FP&A professionals need to be clear about the difference between predictive and prescriptive analytics because the statistical and analytical methods used in each are distinct, and as explained their goals are different too; prediction versus actionable recommendation.
In both cases, however, it’s equally vital to have clarity about what the purpose of the analytical exercise is within the context of your organization. It is the job of the FP&A team to choose the right data points, parameters and variables to be considered as part of the calculations, and some of these vary depending on the nature of your industry, the type of company, and the business’ objectives and priorities.
It would once have been the case that a practical in-depth working knowledge of modeling and statistical methods would have been necessary to engage in predictive or prescriptive analytics. With cutting edge platforms using AI, however, this is no longer the case.
Platforms like Firmbase can empower an FP&A team to query forecast models, pinpoint key drivers, evaluate scenarios and make tailored plans through an intuitive, streamlined interface that’s so simple to use that it also helps increase planning engagement from business partners across the org who are not part of the finance team.
Firmbase also offers natural language formulas, so teams can type questions in and receive instant answers around model forecasts or key areas they’d like to analyze. It’s real-time and so intuitive that the platform can be used live during business discussions about what’s ahead, and what needs more focus within the organization.
Avoid Thinking About Predictive vs. Prescriptive Analytics
There used to be a strong tendency to think of a hierarchy within the types of analytics described above, in which prescriptive analytics was at the top. This was partly due to what was known as the Analytic Ascendency Model, and partly due to the inherent logic of needing to take each step as a foundation for the next.
It is true, of course, that prescriptive analytics in a sense builds on predictive analytics, taking its forecasts and then making recommendations based on the probable scenarios. However, the two are best thought of as complementary rather than an either/or decision, or even being in combat with one another.
For instance, during the Covid-19 pandemic, many businesses were scrambling to forecast what would be happening in the short and medium term in relation to their supply chains and the impact of that and labor fluctuations on their productivity. They were attempting to predict the likely business results. Without a reasonably accurate forecast, they would have been completely at sea without direction. Predictive analytics were essential.
On the other hand, having a forecast was not enough to help them weather the storm safely, much less take advantage of any opportunities it contained. This is where prescriptive analytics came into play. Given what was known, what changes should be made to orders, manufacturing equipment, headcount, supplier contracts and so on?
The combination of predictive and prescriptive analytics is what enables businesses to make the right, data-driven decisions.
Bring Predictive and Prescriptive Analytics to the Whole Business
To be truly effective, both predictive and prescriptive analytics need to draw from data sets that have both breadth and depth, and that reflect the complexity and nuance of a business. In a sense, in one way or another, almost every department will have part of the picture that is needed to feed into the analyses that fuel building, evaluating and adapting strategy.
FP&A teams are ideal leaders, owners or key movers in this strategic analysis and process. However, predictions are more likely to be accurate and recommendations are more likely to be successfully implemented if stakeholders from across the organization are involved as well.
With platforms like Firmbase, increasing the level of engagement is easier than it has ever been. Business partners fully take part using Firmbase’s friendly user interface.
Moreover, it’s possible to employ a consistent unified business language that’s reflected in forecasts and recommendations that highlight the things that move the needle and the business should focus on.
When you bring predictive and prescriptive analytics together, and weave them into the strategic evaluations of the entire business, that’s when you begin to see the real power of what they can do.