4 minute read
Planning for Change: Forecasting and Aligning Budgets to Weekly Schedules
Rusty Secrist, Director of Product Management
This series of blog posts addresses the purpose of a good budget and some of the requirements for creating and managing that budget. We talked about how you have to understand what happened in the past in order to plan for the future. In addition, we talked through how you have to incorporate changes, such as new locations or programs, or more competition. The process should include reiterative forecasting and short-term planning throughout the lifecycle of the budget. As we’ve said, building a budget isn’t easy, but having the right tools and process in place to build and then effectively manage the budget is critical.
It may be helpful to clarify terms. Forecasting is a process of using the best available data plus known trends and events to anticipate what is likely to happen in the future. Most budgets start with a sales forecast. However, long-term planning, as for an annual budget, quarterly plan or period plan, is much more than an exercise in forecasting. The difference is that planning reflects strategies layered on top of the forecast to achieve set corporate objectives. Your ability to forecast accurately is fundamental to creating an achievable budget and to navigating to those larger corporate objectives. Thus “forecasting” and “reforecasting” have specific meanings that are only a part of the larger long-term planning process.
With that, we want to take a deeper dive into some the specific step of creating the forecast that is used in the budgeting process. A good budget has to have a great forecast. Otherwise, how do you expect to stay within the budget and subsequently expect your stakeholders to be held accountable to that budget? An accurate forecasting system is crucial to building the budget, but then equally as important to providing quick visibility into unpredicted trends during the year in progress. Having that visibility then allows you to react and adjust accordingly.
Artificial intelligence ecosystem: Do you know what it is?
So how do you build a good budget? I think many retailers are asking themselves that question a lot lately, especially given the impacts that a pandemic had on most businesses. Forecasting in today’s world must rely on some sort of artificial intelligence (AI) logic, whose goal is, simply put, to predict results based on incoming data. That sounds easy enough, but there is so much more to it as you dive into the details. At a high level, there are three components of an AI forecasting ecosystem: obtaining historical data, tagging historical events and calculating algorithm accuracy.
Garbage in, garbage out
As we work with more and more retailers, it is becoming very clear that most retailers have issues in getting and maintaining accurate and comprehensive historical data. We find that key data needed for long-term planning is often overlooked yet has a huge impact in budgets. A few examples are satellite register sales (like a sub shop or a prepared foods area) not getting included in some data, shrink not getting accounted for accurately, voids and refunds not getting recorded properly, etc. All of these are critical to understanding and must be included accurately in historical data that is used for a forecast. If not, any one of these could really throw off a forecasting system.
The concept of tagging has really become something that many retailers are just starting to do or wish they are better at. Tagging is simply identifying something that happened in a historical timeframe that needs to be flagged so that either it does not get included in a forecast, or that it is included in a forecast during a similarly tagged timeframe in the future. There are lots of types of tags, but a few examples include weather events, sporting events, sale events, remodels, and competitors opening/closing nearby stores. Any sort of automation that can be used to set those tags, for example integrating with a weather service, will significantly help in the process.
Machines are continuously learning
And finally, the most complex part of it all: using algorithms to forecast. Assuming that you have good historical data and that you’ve tagged historical events properly, you can now start using algorithms to forecast. An algorithm is simply a set of rules to be followed in a calculation. There are a lot of algorithms out there. Some are better than others. As you use algorithms to predict a future value, you can then compare those values to the actual values to determine how accurate the algorithms are for that data set. Running this type of calculation over and over for specific stores and weeks and data sets allows the system to determine which algorithm it should use for the forecast for that store week data set. A data set from one store may use an entirely different algorithm as compared to another store’s same data set. This is the power of an AI forecasting ecosystem: the system continues to learn and readjust as actual data is received.
Artificial intelligence in reforecasting
Once you have that annual budget produced using the AI algorithmic forecasting engine and all your layered changes to align to corporate objectives, that same forecasting engine can be used as the year progresses to monitor progress and realign as needed. When actual data starts coming in, each quarter, period or week can be reforecasted and compared to the original budget to understand where you stand against what you planned for at the beginning of the year. Understanding your performance against prior weeks gives you a clear view of how the rest of the year will turn out and whether and how you need to react and make adjustments.
In closing, your budget is only as good as the forecast that goes into it. Budgeting is far more than just forecasting, but accurate forecasting leveraging history and known events is an important step in the process. Leveraging AI technology is becoming a critical component for retailers as they navigate through producing a forecast. Being able to seamlessly transition from an annual budget to a period/quarter refresh to a weekly forecast and schedule is a vital process in any budgeting cycle.