Demand planning / Demand management: Winning the battle for improved forecast accuracy

“Generals are always prepared to fight the last war,” Winston S. Churchill once said.

When it comes to forecast accuracy, is your battle strategy still based on forecast accuracy challenges and KPIs that worked in pre-COVID times? In these new post-COVID times, our line of sight has changed – and in some cases, has rapidly diminished, as irregularities in product supply are now commonplace in many industries.

We previously looked at two methods for reducing inventory while increasing fill rate: addressing excess and obsolete inventory, and precisely calculating safety stock using forecast error as an input. Here, we address connecting front-end business operations’ demand and supply planning to elevate and measure forecast accuracy, and continuously improve it.

Availability of front-end data is a common denominator in forecast accuracy. The forecast planning cycle often targets SKU level data in pursuit of granular level planning. However, if you do not have the necessary level of data, what is the best way to proceed?  One alternative is to mine an aggregated level of demand data which, we assert, can produce a very accurate forecast.

Begin by making two decisions:  1) The right level of forecast to drive optimal results across your Supply Chain, and 2) the level of data available from your systems.

If you are struggling to determine the appropriate level to create demand forecasts, consider that while the details of grouping of products into product families are specific to each company’s situation, the considerations used to group products are consistent.

Focus should be given to the clustering of individual items or offerings in one or more ways: 1) the way in which items sell through various channels, 2) by considering the methods by which products are produced, and 3) in consideration of the complexity of each product offering. Line-item extensions are almost always better planned at brand-level, whereas industrial products are often better planned with an eye towards raw materials and the conversion/production/manufacturing of those raw materials into a set of similar products.
 

Demand forecasting

After you answer these questions, this better aligned aggregation of demand will generate an increased level of forecast accuracy which can then be used to support the calculation of safety stock using this more accurate forecast error.

Once this increased confidence in your forecast accuracy is in place, it is time to leverage a sales & operations planning (S&OP) or sales, inventory, & operations planning (SIOP) process. These processes help to guide organizations to a balanced, well-planned order fulfilment strategy (supply) to meet this forecasted demand. With this improved forecast accuracy, your business benefits from higher fill rates and higher customer experience levels across the enterprise.

What makes S&OP/SIOP a powerful process in your forecast accuracy toolkit is its cross-functional consensus discipline. While sales/value creation organizations often have a wealth of market and customer demand information, they can also bring a strong bias to the forecast. To counter this bias, and other forecast accuracy obstacles, S&OP/SIOP demand planning provides a function for compartmentalizing data into constrained and unconstrained demand. The associated analytics create other delineating trends such as predicted vs actual contract close dates, number of changes to a customer’s requested delivery date and similar data, which can help drive forecast accuracy.

Proactive monitoring of KPIs

There is no lack of forecast accuracy KPIs available to calculate the effectiveness of your forecast. Searching “forecast accuracy measures” online will net a wide variety of industry standard KPIs. Some are better geared towards high-level market demand sizing, while others target consumer demand at a shelf-level SKU detail. For most CPG companies, creating and measuring forecast accuracy is best done at a product family level.

C-Level executives are often frustrated when sales forecasts don’t materialize into expected top-line revenue for the business. Forecasts are never 100% accurate; however, an organization’s goal should be to get into the high 90%. Feedback on why the forecast had a +/- variance to actual results should be fed back into the S&OP/SIOP demand planning process for continuous improvement and identifying corrective actions, driving sustainability of an increasingly more accurate forecast.

S&OP/SIOP is the foundation for disciplined demand forecasting because it orchestrates a 360-degree consensus, connecting front-end functions—-new product ideation, sales, product management, marketing, and engineering with end-to-end supply chain processes. By leveraging a few simple but important metrics and determining the optimal structure and aggregation of your demand, your company can reap the benefits of better forecast accuracy within a few months.

Dorothea Grimes-Farrow is a senior consultant at SGS-Maine Pointe, a global supply chain and operations consultancy, with extensive experience in product development, sales and operations planning, and supply chain operations. Dorothea is an advisor to organizations who want to enhance or adopt new and agile business models for sustainable accelerated time-to-market, cash release and cost optimization. Contact Dorothea at [email protected].

Michael Conley is a consultant at SGS-Maine Pointe, with experience in cpg, food & beverage, grocery, retail, electronics, chemicals and industrial manufacturing. His end-to-end supply chain expertise crosses forecasting, sales & operations planning, logistics & distribution, manufacturing & operations, as well as procurement. Michael can be reached at [email protected].

SC
MR