5 ways to forecast for small and midsize businesses


Wouldn’t it be great if we could predict the future? As a kid, I liked to write sci-fi stories. I wrote stories about what the year 2000 would be like: flying cars, robots, and talking dogs. OK, I won’t admit what year I was actually writing these stories, but back then I had no way of predicting what the year 2000 would be like.

Fast forward to today. How valuable would it be to your business to predict future sales? If you knew exactly how many customers would walk through your door each day, would you change anything about how you run your business? To a manufacturing company, if they know exactly how many parts to make each day and exactly how many will sell each day, they are able to minimize their costs and maximize their profits. While we don’t know what tomorrow will bring, we can forecast what tomorrow will bring!

Consider these common forecasting tasks used by small and midsize businesses, not just large corporations:

  1. Generating time series forecasts based on recent data
    This is easily the most common request for forecasting that I see among SMBs. If I am a retailer, I am going to use my historical sales data to determine how I want to stock my stores for the next 12 months. This can even include a seasonality component to take into account features such as July 4th runs for barbecues and Thanksgiving runs for Turkey. Retailers will need to have more in stock to account for those upticks in seasonal demand.
  2. Generating hierarchical forecasts
    It is common among my customers that they need to forecast in a hierarchy. For example, the forecast for products sold in a grocery store by SKU are the most granular level. However, there is an aggregate up to product line, brand name, region, and the overall forecast.  The forecast may also have a store-level. This forecast has upwards of five levels: product, product line, store, region, and overall.
  3. Identifying and correcting problematic forecasts
    Because we can’t predict the future, some unpredicted event may occur that causes our forecast to require updating. An unexpected weather event, a drought that causes an increase in the price of food, or a decline in the stock market can all cause unpredicted changes in our forecasts. SAS Forecasting for Desktop automatically chooses the most appropriate forecasting model, optimizes model parameters, and generates the forecasts. This frees up your time to spend it where you need it the most- to focus your attention on problematic or high-value forecasts, and conduct “what if” analysis through the Scenario Analyzer tool.
  4. Adding manual overrides to the forecast based on business knowledge
    Adding in business rules or business logic to your forecasts is invaluable. SAS gives you the power to add in your business rules so that the forecast is able to account for information that drives your business- such as including a maximum value that can be put into stock, or information on a planned price-cut or special promotion.
  5. Publishing the forecasts to other systems
    A good forecast is meant to be a call to action. By having the ability to share your forecast with other systems, you are able to seamlessly work across your enterprise and take action. The retailer that needs to change the quantity of items they have in stock? If it’s the day before Thanksgiving, they are probably too late getting extra Turkeys shipped to their stores. All of the best analytics and forecasts won’t help you if you can’t quickly and easily share the information with your enterprise systems. Forecasts created within SAS can easily be shared with other applications so you are ahead of your competition.

Using SAS to bring these concepts into practice has helped many of our customers.  For example, the Northern Virginia Electric Cooperative (NOVEC) provides power to 144,000 customers. To keep electric costs down and to reliably serve customers, NOVEC needs to know how much power to buy, transmit, and deliver for its customers. SAS Forecasting provides NOVEC with a broad array of econometric and time series forecasting techniques, along with point-and-click interfaces that can grow with the utility.

Read more on how NOVEC is benefitting from SAS Forecasting, or read the white paper SAS Forecasting for Small to Midsize Businesses.


About Author

Wendy McHenry

Systems Engineer

Wendy McHenry is a Systems Engineer at SAS, and every day she gets to show her customers how SAS can help solve their problems. Her primary focus is on our SMB customers. Wendy has been a SAS user for over 17 years and joined SAS as an employee in the Fall of 2011. Her SAS focus areas include data management, business intelligence, and SAS Administration. In her spare time, she is a Girl Scout volunteer. Connect with Wendy on Twitter at: @wendymac98

Comments are closed.

Back to Top