Demand forecasting has been the bane of suppliers, manufacturers and retailers for years. Considering the costly and detrimental consequences that poor forecasting can have on a business, it seems more companies would give this part of their supply chains more attention. But many don't. For some, it is guesswork at best.

Demand forecasting is complicated and a number of challenges come into play:

• Demand frequency can be erratic, intermittent, lumpy or seasonal-or a mixture of all of these. Rates and patterns shift with buying trends and product lifecycles.

• External forces such as planned demand, promotions and holidays can have significant effects.

• Technology updates typically process at night, blocking inventory optimization and merchandise planning opportunities.

• Operational issues, such as large SKU volumes, make it more difficult to analyze forecast updates, seasonal profiles, trends and demand frequencies.

Fortunately, psychic powers aren't a prerequisite in forecasting demand patterns or overcoming these challenges. There are solutions and tools to choose from, but it's not easy to find one solution that addresses the full range of demand patterns many businesses experience. While some solutions require large amounts of expert intervention, most buyers and planners aren't certified forecasts analysts. They need a comprehensive, effective solution that's also easy to use.

Single-Method Forecasting Produces More Risks Than Results

Don't select a forecasting method that manages just one component-it creates more problems than it solves. For example, single method forecasting might handle seasonal issues, but it's too risky because demand patterns vary significantly based on SKU types and lifecycles. SKUs with short lifecycles of two to six weeks, such as video games or DVDs, outpace standard forecasting algorithms. By the time it catches up with the demand pattern, that SKU's life may be over. Conversely, fashion apparel SKUs can be notoriously slow sellers, requiring alternative forecasting techniques.

Different approaches are needed to forecast demand for different product types-such as new products, which require very responsive forecasting and replenishment approaches. Even in a business with some of the most endurable products, SKU types are highly diverse. The toy industry, for example, not only has steady sellers such as traditional board games, but also new, fad-sensitive, short lifecycle toys that require new forecasting approaches.

Best Pick Doesn't Bode Well for Accuracy or Performance

It's clear that single demand forecasting solutions have become antiquated as supply chains grow in size and complexity. Most demand forecasting and planning, or inventory management vendors, provide standard textbook methodologies. But these are built on assumptions that are user-defined, forcing companies to use a "best pick" approach-which usually means "best guess."

While this approach allows analysts to choose the forecasting method that fits their company best, there are too many shortcomings. Most companies have to staff forecasting experts to carefully watch over the solution and monitor demand signals to swap and tune methods. That's not ideal.

Customers with large SKU populations are forced to segment how they forecast demand, causing major efficiency problems, and more importantly, accuracy issues. Other problems start surfacing as well, such as lack of scalability and sub-par performance levels. Systems may need to be reconfigured or initialized again if the SKU dynamics change. In another words, this limited solution won't flex very well with inventory or demand changes.

Instead of settling for a single method or best-pick approach, look for a more comprehensive solution that continuously handles transitions and mixtures of demand pattern changes. Ideally, it should self-adjust and not require a Ph.D. to operate.

Choose a Unified Forecasting Method to Better Predict the Future

Despite its limited flexibility, best-pick demand forecasting was the most practical option for buyers and planners-until recently. Now a better choice is on the table that takes best-pick one step further. The Unified Forecasting Method (UFM) combines multiple, well-proven processes in a solution that comprehensively handles demand patterns in a continuous manner. And it works regardless of temporary or permanent shifts in demand behavior. Built-in intelligence significantly reduces and nearly eliminates constant system monitoring.

UFM accomplishes this by separating demand signals and dividing them into two spectrums: demand level and demand frequency. It tracks, evolves  and combines components within these two spectrums to predict future demand patterns. It does the heavy statistical lifting. Better yet, it provides more accurate demand forecasting for buyers and planners who won't require a degree in statistical sciences to successfully implement UFM.

Some items trend, and some have what is called intermittency-they have a sale and then several consecutive periods of no sales. UFM addresses the entire demand spectrum. From extreme intermittency to high-volume movement, it accounts for seasonality, trends and everything in between.

The Forecast Calls for Increased Revenue and Service, with Fewer Lost Sales

UFM takes a holistic approach and dynamically adapts the forecast method components to demand signals, helping retailers and wholesalers improve inventory accuracy and enhance inventory optimization. This, in turn, has resulted in greater revenue and fewer lost sales-not to mention better service. Two examples of UFM at work are:

Retail:

For auto parts dealers, staying in business means having those hard-to-find items in stock for customers when they need them. The business problem for the automotive parts industry-and retail in general, is a large volume of the SKU population with extreme intermittent demand signals. The typical approach is to throw more inventory at the problem, but that's not the right answer. Traditional forecasting simply can't handle intermittency-revenue will suffer every time.

The Unified Forecasting Method enabled an auto parts retailer to avoid out-of-stocks while mitigating the risk of having more inventory than required, with successful results that included:

• decreased in inventory for non-intermittent SKUs by 9 percent
• increased service attained for non-intermittent SKUs by 1 percent
• increased revenue for intermittent demand SKUs by 21 percent
• increased service for intermittent demand SKUs by 17 percent
• increased inventory for intermittent demand SKUs by 14 percent

Wholesale Distribution:

A very large distribution wholesale enterprise needed to further reduce its inventory investment while maintaining current levels of service. The wholesaler was already running an extremely fine-tuned version of inventory optimization and had left no stone unturned from a traditional forecasting point of view. UFM enabled the company to lower its overall inventory investment while maintaining service, which resulted in:

• decreased inventory by 4 percent 
• reduced safety stock reduction across its network by 7 percent
• increased service for intermittent SKUs while inventory actually decreased 

Profitable Buying Demands Accurate Demand Forecasting

Achieving that perfect balance between demand and inventory is a formidable proposition, but tackle it anyway. The ability to holistically address a broad range of demand signals makes it easier to manage forecasting environments. Search out new approaches to demand forecasting and seasonal profiling with solutions such as UFM. It provides inventory analysts, buyers and planners an edge that traditional forecasting tools simply can't deliver.

Adding a unified demand forecasting tool to existing inventory optimization and merchandise planning solutions enables companies to buy and plan more profitably. It's almost like having a crystal ball.

Source: Manhattan Associates