This article on planning analytics is the first in a series of guest posts written by Dan Vesset, Group Vice President of the Analytics and Information Management market research and advisory practice at IDC.
In recent years, all the headlines about big data, business intelligence (BI), analytics, performance management, data lakes, and AI have diverted our attention from the ultimate reason for these solutions — decision support. Implementing all of these tools is meant to help humans make better decisions, to augment the ability of a person to decide with the assistance of information. So, it’s important that we return our attention to the decision-making process that’s the lifeblood of smart business operations.
These decisions, ranging from high-level strategic to operational to customer-facing tactical, are made by executives, managers, and operational employees, respectively. In all cases, decision-making really is a process. And decision support is not simply information delivery or reporting or creating a dashboard or a spreadsheet. It is a series of steps that enable a person to make the best possible decision based on the best information available at that moment.
Different types of decisions have different characteristics such as:
- Decision Risk. What is the risk involved if the wrong decision is made? Is the decision easily correctable, or is it a major, “bet-the-company” decision?
- Time Window. How much time do we have to make a decision or resolve an issue? How firm is the decision deadline?
- Variability. To what extent is the issue routine vs. ad hoc? Is this a frequently made decision or an exception?
These and related characteristics determine the types of technical and process capabilities in which an organization needs to invest. Decision support capabilities can be segmented into five related categories, each of which is deployed to answer different types of questions:
- Planning analytics: What is our plan?
- Descriptive analytics: What happened?
- Diagnostic analytics: Why did it happen?
- Predictive analytics: What will happen next?
- Prescriptive analytics: What should be done about it?
In this series of blog posts, we’ll address each of these analytics capabilities. This first post will focus on planning analytics.
Each of these steps is aligned with a key question from decision makers. Each of the questions and steps of the decision-making process builds upon the answers to the previous questions. To answers these questions business subject matter experts, analysts, and data scientists need to work together to employ the post appropriate analytic techniques for the wide variety of data now available for enterprises. It’s important to recognize that data scientists have a role to play at each of the steps, not only at predictive and prescriptive steps. Data scientists are in very high demand — so much so that they are not always available for all business functions. That is why machine learning and other analytic techniques of artificial intelligence are increasingly embedded into business analytics software, where they automate repeatable tasks within each of the decision-making steps. This automation is made possible by training machine learning algorithms on data from ongoing monitoring of decision making steps and processes. This type of task automation allows the software to augment people and operationalize data science within business functions.
Planning analytics: What is our plan?
It all starts with a plan. Whether it’s the overall corporate plan or one of many lower-level plans, this step is usually associated with enterprise performance management, which includes financial planning, budgeting, and forecasting. Planning is in a unique position compared with the other analytics categories, because it relies on outputs of all the other steps. It requires an understanding of past performance, identification of deviations from the norm (plan vs. actual), evaluation of possible scenarios, prediction of likely outcomes, and assessment of risks and constraints.
Far too many enterprises rely too heavily on spreadsheets for planning activities, despite the complexity and significance that planning has in guiding business strategy and its execution. Spreadsheets per se are not the problem; they remain the original “killer app” of business software, ubiquitous and easy to use. The real problem is that the disconnected, siloed, and ungoverned use of spreadsheets does not foster efficient planning processes, and spreadsheets cannot scale effectively in large organizations.
In a recent IDC survey of 300 business users, 88% of respondents said that they use spreadsheets for performing what-if analysis. Line of business users fill in spreadsheet-based plans and email them back and forth with finance managers to arrive at the final plan. Managing multiple versions of spreadsheets in such a manual process that it makes it virtually impossible to maintain governance and reach an agreed upon “single version of the truth.” In addition, because the planning process is so time-consuming, plans are usually only revisited at the end of the period to determine performance variance. That does not provide an opportunity to course-correct during the period.
In the same study, we found that 49% of respondents use manual copy-and-paste methods to enter their data into spreadsheets. This process is rife with human errors and inefficiency. Some organizations string together spreadsheets in impressive feats of data manipulation. For example, in a recent interview with a biotech company, we found that 72-tabbed linked spreadsheets were used to create budgets, and that the budget-to-actual update process took four months to complete. With a new planning application, the company reduced that period to just two weeks. In another example, a manufacturing company used the predictive capabilities in their planning application to estimate actuals several days prior to period close, a process that previously took a team of 40 analysts a full week.
While we all acknowledge the value of spreadsheets, enterprise-grade planning requires dedicated, connected tools that help collect, prepare, and analyze data; adjust planning models and deliver them to downstream processes or applications; and provide insight to upstream decision makers. The functions that these tools provide broadly fall into four categories:
- Collect and prepare data: Finance teams collect data from a variety of sources to build budgets and forecasts. Data is consolidated and prepared to provide an accurate account of the current state of operations. Planning applications that are integrated with transactional applications have access to data on demand.
- Analyze data: Based on historical data, financial planners build a variety of models, including variance models. Driver-based plans and rolling forecasts enable this process to adapt quickly to changing business conditions. Line of business owners often also build their own budgets, plans, and forecasts for their individual departments, which are then combined into one overall plan.
- Deliver analytics: Reports or dashboards are published and delivered to executives, business users, and other stakeholders based on the analysis of the data and governance policies. Decision makers can view visualizations, including charts, graphs, and maps, and adjust their decisions to improve business outcomes. Planning tools that are integrated with transactional systems can drill back to the systems-of-record for detailed analysis.
- Adjust models: Planners adjust their models based on variance reports that can help improve the accuracy of forecasts. This planning cycle is usually conducted periodically (monthly, quarterly, biannually, annually). In organizations with more advanced capabilities, continuous planning is deployed and plans are adjusted on demand.
Enterprises are most successful when planning is continuous, active, and collaborative — and when business users work in tandem with financial analysts and planners throughout the entire analytics cycle to improve enterprise operations and decision making. These needs cannot be met by manual processes reliant on spreadsheets, which is why planning analytics solutions are a key part of ensuring that timely and accurate decisions are made with the best and most up-to-date data.
Read the second post in this series, “Descriptive analytics 101: What happened?”
For IBM’s view on the Analytics Cycle, check out our smartpaper, “How Can You Trust Your Data Without the Big Picture?”