follow us on: RVCF Facebook RVCF Twitter RVCF LinkedIn RVCF Google+

Upcoming Events

Supplier Open Forum
Conference Call

Retailer Open Forum
Conference Call

Supplier Open Forum
Conference Call

Retailer Open Forum
Conference Call

Supplier Open Forum
Conference Call

Retailer Open Forum
Conference Call

Supplier Open Forum
Conference Call

Retailer Open Forum
Conference Call

Annual Fall Conference

RVCF Live LINK: Listen Now

Why Planning and Sales Forecasting are Bad for Replenishment
Stuart Dunkin, Data Profits Inc.

Learn the differences in forecasting and planning or pay the price How often does key management review a forecast accuracy report? The irony is the same people that don't review the forecast accuracy report are the first to talk about out-of-stocks, overstocks, and open to buy issues. The same people will have this "I can't believe this has happened" look; arm waving, yelling, finger-pointing, and similar activities will often occur in meetings (yes, I've worked in retail). Your forecast is the critical element to retail success. Forecasting is used to plan, budget, and replenish. Replenishment includes lead-time days and order cycle days or weeks of supply that with a forecast convert from days to a number of units needed in the supply chain. The big question is this: Why is forecasting accuracy left off of many KPI screens, but the results of poor forecasting are front and center in your company as the hot, must-fix agenda?
To answer the question, let's review some forecasting basics to help determine where help is needed. We'll review the differences between forecasting and planning as well as learn the differences between sales forecasting and demand forecasting. Then, we'll merge the pieces to understand the benefits that impact your planning and replenishment results.

Differences between Planning and Forecasting

The first step to better inventory results is a total understanding of what forecasting is and, equally importantly, what it is not. Product Forecasting is a management process; it is not a software service. The distinction is important because, as mentioned, forecasting is critical to retail operations. Products must be purchased at the best price, locations are needed to store and display the products, logistics services are used to move products, and merchandise services are utilized to present the products to the customer. Each of these critical operations is impacted by the forecast and the forecast's accuracy.

Many retailers believe the most important decisions about forecasting revolve around software selection or development. Retailers often adopt a basic belief that "If I have good software (if I have expensive software), then I have a good forecast." Many studies have demonstrated that the

Demand-driven retail requires bottom up supply chain

most expensive software solutions do not lead to the best forecast accuracy. The reason the high-end systems failed is linked to a failure to monitor and manage the forecasting process after implementation.

Another way retailers often confuse what forecasting is and is not is by failing to understand the relationship between forecasting, planning, and goals.

  • A demand forecast should be a "bottom up" analytical prediction of what future sales will occur.
  • A plan should be a "top down" management decision of what the company will do.
  • A goal should be the target everyone in the company works to attain or exceed.

The primary goal of forecasting is to guide management in creating sales plans and planning future activity. The purpose of planning is to drive operational and strategic initiatives in a manner to maximize return on investment and increase profits. The primary function of a goal is to motivate the entire company to meet or exceed expectations. Forecasting and planning are closely linked, with forecasting being used to impact planning. You do start your planning with the forecast and build on top of the forecast, right?

Forecasting should strive for accuracy; unfortunately, many of you mix company goals and plans with a forecast. Mixing forecasting with plans and goals always leads to "game playing." The buyers and planners have an agenda based on their quota and bonus programs. The forecasting team has an agenda and a bonus based on forecast accuracy. Forecasting and forecast management are analytical functions that should be based on forecast accuracy and demand, not goals or plans.

Companies worry that managing two to three sets of numbers (forecast, plan, and goal) will be problematic. The answer for most companies is to change the forecast or manage the plan only and then ask why they are out-of-stock, overstocked, or not on plan. Forecasts, plans, and goals are distinct because the behaviors they are meant to influence can be in conflict.

Differences between Sales Forecasting and Demand Forecasting for Replenishment

Replenishment requires accurate demand forecasting

Sales Forecasting and Demand Forecasting deliver different results while using potentially the same algorithms. Sales Forecasting and Demand Forecasting are two totally different methodologies. When forecast accuracy is measured, "sales forecasting" accuracy averages less than 85%. Demand Forecasting can deliver forecast accuracy better than 90%. It is ironic that so many people use the term demand forecasting and fail to understand the meaning of the term. Again, if we define what demand forecasting is and is not, then we can start to improve process and results.

Contrary to what some software companies tell you, demand and sales forecasting methodologies are totally different but can use the exact same algorithms.

The algorithms do NOT define if the forecasting methodology is "sales" or "demand" forecasting. Think about this just a minute: Wouldn't you expect sales forecasting and demand forecasting to use the same algorithms? Another common misconception is: If software forecast at a product/location level, this is demand forecasting. Both sales and demand forecasting methods can deliver product/location level forecasting, the difference is rooted in the sales data being sent to the forecast algorithm.

Sales Forecasting will aggregate all types of the sales data to feed into the forecast engine. The aggregate sales types include regular, promotion, event, lost, and closeout sales. The forecast gets skewed because the aggregation includes closeout sales, promotions and lost sales. Worse, many forecasting apps, some with "demand forecasting" in the name, are actually using a sales forecasting methodology. This is a top down methodology and works well for management but is not a good methodology for retailers with promotions, events, and closeouts.

Demand Forecasting is when the separate sales types (regular, lost, promotion, event, and closeout) are forecasted independent of the other sales data types. The result is a demand forecast for regular sales, often called "base forecast," and a forecast for each of the other key demand types. A demand forecast can be created for promotions and events because the promotion or event sales alone are passed into the forecast engine.

In practice, many of you are manually trying to separate sales into types, adjusting the numbers into sales by type before you update a forecast. In the past, few software packages were capable of handling the math to separate the sales by type and forecast by each sales type. They were/are million dollar solutions. Today, many legacy software companies with existing "sales" forecast methodologies are opting not to rewrite their software due to the expense of updating so many pieces of their total package. There are new software packages that install fast and are significantly lower in price.

In promotional execution the benefits of demand forecasting can be easily understood, down to the product/location level.

Example: This April, a product, normally selling 1,000 units per week or 4,000 units monthly, was on promotion in week three. The promotion resulted in selling 500 additional units, a total of 4,500 units for the month. Without separating out the additional 500 units of promotional sales, a sales forecasting engine will forecast based on the 1,500 units sold that week. The demand forecasting engine will have two forecast based on 1,000 units regular and 500 units promotion. The sales forecast system would increase the forecast, the aggregate sales of 4,500 supports that increase. The demand forecast system would leave the forecast alone because the regular or base demand of 4,000 was met.

The promotional lift is immediately seen and understood in the demand forecast system both during the promotion and later when end of month forecasting and open-to-buy updates occur. A sales forecasting system requires multiple edits: first, to add a trend to raise the forecast to support the promotion and secondly, a trend to lower the forecast, post promotion. It is not reasonable to assume a human or team of people can manage all the product locations accurately to filter sales for promotion events, particularly when thousands of items are on promotion throughout the year. Without automation and a good software solution, the regular, lost, and promotion sales execution suffers with less accurate forecasting resulting in lower service levels and profitability.

Benefits of Real Demand Forecasting and Less Game Play

Benefits of managing the inventory at a product/location level with accurate demand forecast and understanding the role of plans begins with the forecast. If accurate demand forecasting is applied to planning and inventory management, customer service levels increase, inventory investment is decreased and plans and budget will be more realistic without the need for game play. You can compare plan to forecast and determine your direction. Both forecast and plan can be impacted by weather, new products, and competition. The goal is to identify what's impacting the forecast or plan and make course corrections rapidly to forecasting or edits to your plan. You can quickly move ahead of your completion and see new sales grow this year.

Differences in Forecasting and Planning for Replenishment

Join us each month for our short series on forecasting and replenishment. Today we helped define some differences and set some expectations concerning forecasting. Articles in our continuing series:

  1. How do systems calculate a forecast - the algorithms, indexes, accuracy, and expectations?
  2. When is product order point - Safety Stock, Lead-Time, Plan, Just in Time
  3. How often should you reorder - Inventory Optimization, Weeks Supply
  4. What should be on the PO - Top down based orders (plan), Bottom up based orders (Service Level ~ Consumer centric)?

Copyright © Data Profits, Inc. 2013 All Rights Reserved.

Stuart Dunkin, CEO of Data Profits Inc, is a strategic visionary who leads by calling on his broad range of experience as a former retail executive, consultant for top retailers, and work for E3/JDA software. Data Profits delivers forecasting, replenishment and collaboration supply chain tools to our customers. Data Profits interviewed over 100 retailers who outlined - the disjointed relationship between supply chain data, people and business goals. This influenced our software designs. Data Profits Software connects people, meets service goals, increases profits and provides a highly visible supply chain across a retail network. The user configurable software installs in less than 30 days at more than 50% less than legacy solutions.

Stuart attended Auburn University for a Bachelor of Science degree in Business Administration. He continued his education at Emory University, pursuing a joint Masters of Divinity and Masters of Business Administration. Stuart writes almost weekly for a popular retail supply chain blog found here. Call Data Profits at (770) 574-4100 or visit their website and subscribe to their blog at and start to "Tighten the Links in Your Chain™."