This once required the skills of highly-qualified data scientists. Fortunately, this isn't now the case thanks to intuitive, visual prescriptive analytics software that simplifies and accelerates model development. In their place are citizen data scientists: power users, managers and analysts whose primary responsibility is line management.
In this blog, we explain:
● What prescriptive analytics software does.
● The problems it solves.
● How to identify it.
Prescriptive analytics is a decision support system based on linear programming and optimization theory. Gartner's prescriptive analytics definition describes it as the use of mathematics and logic to recommend a course of action. The technique determines the best decision given available data.
While Gartner includes decision automation software, known as heuristics, in its prescriptive analytics definition, this isn't strictly correct. Heuristics follows a predefined set of rules to determine an answer which may or may not be the best. In prescriptive analytics optimization, analysts use a model simulating a real-world problem. They then apply machine learning and AI to determine the best decisions to achieve specific objectives.
Several factors indicate when prescriptive analytics optimization is recommended as opposed to other forms of decision-making, such as heuristics. These include:
Prescriptive analytics software is ideal for determining the best or optimal decision in a set of complex circumstances. These typically include numerous variables with multiple possible outcomes.
It's possible to define a set of overriding objectives. These could include metrics such as maximizing revenue, cost and profit or optimizing production.
Numerous constraints exist. A constraint is a bottleneck, limitation or regulatory barrier that must be respected. It may be something simple like an HR policy or a complex trade agreement. Factoring in constraints makes the difference between feasible and infeasible decisions.
The problem is complex with multiple variables. To understand the impact of variables, consider a plant manufacturing five products on three lines for six customers. This limited example has 90 possible combinations and demonstrates the difficulty of establishing an optimal solution.
Decision-making is often a combination of judgement, experience and gut feel supported by limited data. There must be a better way.
There is. With prescriptive analytics software, you construct a mathematical model of the problem. You populate this model with structured and unstructured data (big data). Then, using optimization capabilities, you run multiple simulations to discover which of many possible decision solutions is the best.
Here are some applications that are well-suited to prescriptive analytics analysis.
Over time, organizations develop formal and informal practices based on what was logical at the time. These include buying practices, production methods and procedures that don't stand scrutiny when subject to analytical analysis.
Operational budgeting, capex planning or S&OP generally follow sequential planning processes spread over several weeks. Resource constraints often mean revisions are squeezed in without regard for how they affect core assumptions. Using prescriptive analytics, planners can evaluate the feasibility of these changes taking constraints and limitations into account.
In highly dynamic situations such as process control and commodity trading, conditions and pricing change constantly. Using prescriptive analytics to determine what to do helps operators understand trade-offs, make better decisions and increase transparency.
Sales and marketing executives use marginal costing to aid sales decisions. Unfortunately, incorrect assumptions that margins are fixed lead to poor sales decisions. In reality, margins are impacted by multiple factors, such as volume, cost of production and raw material pricing. Savvy executives use prescriptive analytics software to answer what-if questions to determine the actual margin if they adjust prices to win an order.
Spreadsheet planning is a major factor in many organizations. While spreadsheets are flexible and easy to use, they're also static, difficult to maintain and prone to errors. The Wall Street Journal recently caused a stir when it questioned whether Excel is still suited to modern-day financial planning. Using prescriptive analytics software for planning overcomes these problems and results in better plans.
Do these examples ring a bell? If so, you're probably ready for prescriptive analytics software. One way to verify this is to run a pilot project. This allows you to evaluate whether prescriptive analytics software offers better insights and helps make informed decisions.
You don't need to recruit or hire a data scientist. Instead, try an intuitive drag-and-drop prescriptive analytics software platform such as River Logic's Enterprise Optimizer. Its visual tools simplify model preparation, and the transparent interface allows users to easily change parameters and update models. It's not that difficult to get started with prescriptive analytics software.