Businesses continue to grow in complexity and, at the same time, companies are facing increased competition. To survive and grow in such an environment, businesses need access to tools that allow them to understand their strengths, their weaknesses and to determine the optimal direction forward.
To do this, many are turning to data and analytic tools such as predictive and prescriptive analytics. According to Gartner's research, these tools should drive every process, inform all decisions and be at the center of the business.
This process of optimization entails several factors, including the need to identify those decisions that achieve the best return or performance in any given situation when taking constrains and limitations into account. This is what analysts are saying about optimization.
Optimization Software Trends
Gartner sees the market for prescriptive analytics growing at a CAGR of 22 percent while others like IDC and Reportlinker anticipate slightly higher growth. IDC expects that by 2020, at least half of all business analytic software will incorporate prescriptive analytics and that the spend on self-service analytical tools will grow significantly faster than tools reliant on IT and data scientist support.
According to a study on BI and Data Management in The Cloud conducted by Pierre Audoin Consultants and published on the group's website, business analytics are moving to the cloud. Currently, nearly half of all BI applications are already there and approximately 25 percent of all companies are fully committed to cloud computing. Gartner, in their 2017 Planning Guide for Data and Analytics support this view and see analytics going viral with analytics guiding all interactions, informing decisions and driving optimization processes.
Use of Prescriptive Analytics to Support Business Decision Making
According to Ventana Research, the three basic forms of analytics are descriptive, predictive and prescriptive. Descriptive analytics are commonly used to assess past performance. Predictive analytics determine what may happen and are used in forecasting, simulation and basic modelling. Prescriptive analytics, representing the highest level of analytics, are used to optimization business deicions by identifying the decisions most aligned with company KPIs while taking into account tradeoffs and constraints. According to Ventana, only one a few companies use predictive analytics in any meaningful way and even fewer use prescriptive analytics, so there's still a long way to go.
Forrester concurs with the concept that prescriptive analytics guide decision making and help identify the best course of action but often tends to lump AI, machine learning and mathematical optimization under the title of prescriptive analytics.
Role of Analytics in Optimization
The concept of optimization is well established but for decades remained the realm of the specialist. It was only with the advent of user-friendly analytical solutions such as prescriptive analytics that the real potential for business optimization could be realized. Granted, algorithms prepared by data scientist can perform the same function, but the tedious coding requirements largely preclude their use in fast-moving business scenarios.
Optimization calculations make use of a variety of tools including linear and mixed integer programming, constraints and heuristic algorithms. With these tools, it's possible to analyze complex business situations and prepare explicit decision-making models that can determine the outcome of several alternatives and arrive at the most favorable decision.
How Various Business Sectors Are Using Optimization Techniques
It's interesting to review just how different business sectors, such as energy, manufacturing, healthcare and government, are using optimization techniques to enhance their strategic, operational and tactical decision making. Here are some examples:
- Energy and Utilities: This sector is facing numerous challenges brought about by low and fluctuating energy prices, low margins and the need for long-term investments. HfS Research studied the impact of this on major utilities and noted how they are turning to digital technologies, big data and intelligent systems to manage and drive down costs. Others, like Thames Water in the UK have used optimization techniques to identify millions of pounds worth of savings.
- Manufacturers: Intense competition means manufacturers continually strive to meet customers' expectations by producing the right goods and reducing costs. Although ERP is used to manage day-to-day operations, manufacturers find they need more advanced analytical tools to guide their tactical decisions. An example is McKee Foods who uses prescriptive analytics to optimize operations, leading to increased profits, better decision making and business transformation.
- Healthcare: This sector is in a state of flux. Uncertainty regarding funding and the need to enhance patient care is driving rapid transformation. HfS Research reported on work done at the Virgen Del Rocio University Hospital that used advanced analytical techniques to develop a SmartICU solution to accurately predict acute hypotensive episodes amongst ICU patients. And, the Sir Mortimer B. Davis Jewish General Hospital used prescriptive analytics to model patient flow, which helped optimize the number of patients that could be cared for.
- Public Sector: Across the world, governments are struggling to contain costs while meeting service demands. There's a need to utilize scarce resources effectively. In Australia, the public sector is using analytics to optimize emergency response services, and in the U.S. it has been used by Homeland Security and the U.S. Border Patrol.
Limitations Inhibiting Optimization
Despite the progress achieved, there are still several issues inhibiting optimization capabilities. There is an acute shortage of skills and both IDC and Forrester believe this will persist. Gartner feels that risks associated with big data projects will increase and Forrester predicts that as many as half of all big data projects will stagnate.
The failure to take into account all possible scenarios, limitations and constraints could lead to incorrect predictions such as occurred with the 2016 presidential election. Other issues include the quality of data, how it's collected and the potential for ethics violations.
Overcoming Limitations and Moving Forward
There are several ways in which the impacts of these limitations can be mitigated. One solution to the shortage of skills is to deploy user-friendly optimization packages that reduce the requirements for data scientists. It also makes sense to align with vendors who have proven capabilities and support infrastructure so that organizations have access to requisite skills.
While there are no short cuts when it comes to modelling a business, it helps greatly if the software solutions allow for intuitive modelling and are easily configurable and scalable. Finally, before deployment, models need to be verified and calibrated using known situations to ensure their integrity. In this way, organizations can capitalize on the opportunities offered by optimization techniques.