The Impact of Forecasting Improvement on Return on Shareholder Value
John T. Mentzer; Journal of Business Forecasting

Improving Salesforce Forecasting
Mark A. Moon and John T. Mentzer; Journal of Business Forecasting

Benchmarking Sales Forecasting Management
John T. Mentzer, Carol C. Bienstock, and Kenneth B. Kahn; Business Horizons

Conference: The Dynamic Supply Chain--Strategies for Implementing Sustainable Competitive Advantage,
Nov 17-18, NY
Hear Glen Margolis, Founder & EVP Steelwedge Software, Carol Ptak, Former APICS CEO and Author, Others
Learn more

Forecasting and Consensus Planning in a Rapidly Changing Environment
June 15, 2005, 1pm Eastern, 10am Pacific
Principal Speaker -
Christine Pfefferle, Director of Global Demand & Order Management, Tellabs  
Click here to view recording

Extending SAP for Global, Integrated S&OP ,
May 25, 2005
Stu Reekie, Air Products
Click here to view recording

By Kenneth Kahn, Ph.D. University of Tennessee

Marketing, operations and finance professionals deal with a myriad of considerations and issues when introducing new products. For example, there are various kinds of new products and many available techniques for forecasting these new products. Defining the environment and determining the right methodology are critical to generating accurate and meaningful forecasts.

Let's begin by looking at the new product landscape. The Product Development Management Association (PDMA) has identified seven different types of new products--cost reductions, product improvements, line extensions, new uses, new markets, new category entries (new-to-the-company), and new-to-the-world products.

New Product-Market Matrix

Because each of these is unique in terms of newness, forecasters should not use the same technique for every type of new product introduction. Rather, forecasters should keenly think about which forecasting approach best suits the unique issues posed by that new product respectively. In most cases, companies have a good handle on their product technology issues so much of the uncertainty in new product forecasting is driven by market conditions and market dynamics. This requires careful consideration of market assumptions and their impact on the forecast.

Another issue that makes New Product Forecasting complex is use of the phase review or Stage-Gate TM new product development (NPD) process. This type of NPD process specifies product development activities and the approval criteria necessary in order to progress to subsequent NPD stages:

• Strategic Planning (opportunity Identification)
• Concept Generation (ideation)
• Pre-Technical Evaluation (concept evaluation and business case analysis)
• Technical Development (detailed design)
• Commercialization (marketing plan finalization and launch)
• Life Cycle Management (post launch management)

Different stages require different kinds of forecasts, where revenue and profit forecasts would be emphasized in early stages transitioning to unit forecasts in later stages. Naturally, the forecast form desired will influence the forecasting technique one should use.

New Product Development Process

New product forecasting techniques fall into two main categories: (1) Judgmental Techniques (Subjective Assessment and Customer Market Research) and (2) Quantitative Techniques (Time Series Forecasting Models and Regression)

New Product Forecasting Techniques

In the Judgmental category, the Delphi Method represents one popular technique. The Delphi Method is an anonymous or confidential consensus building approach, where a forecast coordinator asks key stakeholders for their forecasts and rationales for these forecasts. The overriding goal is to achieve a consensus forecast. Because a series of forecasting iterations is necessary, the Delphi Method takes time. There is also the risk of "anchoring" where a stakeholder is unwilling to compromise on a forecast.

Another popular judgmental technique is Assumptions-Based Modeling. This technique creates a set of assumptions to break down and model the marketplace. Typically, an assumptions-based model starts with a total market potential value and proportions the market down to a penetrated target market forecast based on given assumptions. Marketers tend to prefer these models because they are valuable for discerning sales potential and useful for assessing assumptions via what-if scenarios.

Among the quantitative techniques, Looks-Like Analysis or Analogous Forecasting is commonplace. Looks-Like Analysis is an approach that maps the sales profile of a prior (current) product onto a forthcoming product. If the prior product's launch period is included (the first 12 months of sales), the technique can be combined with an assumptions-based model to predict first year sales at a monthly versus yearly level.

Diffusion Modeling is a sophisticated technique that uses an S-shaped curve to forecast sales over time. To apply a diffusion model we need to understand the saturation point, the mass media effect, and the word of mouth effect. Mass Media communicates the innovation of the product, i.e., new attributes or functionality; and Word of Mouth is the consumer's vote on the new product. For further information on diffusion modeling one can reference Vijay Mahajan or Jeffrey Morrison.

An emerging quantitative technique is Curvilinear Analysis. This technique takes into account the base demand curve modified by the layering of curves for pertinent market factors, e.g., pipeline fill, promotion impact, and pricing changes. When the curves are combined, a final new product forecast is generated.

Chart #4 - New Product Forecasting Methods Segmentation

Considering the various types of new products and the above mentioned techniques, the following segmentation is suggested. Because of the availability of market data, quantitative techniques can be employed to forecast current market products like product improvements and line extensions. New market products lack data so it is more appropriate to employ qualitative techniques like assumptions-based modeling and the Delphi Method.

As a company develops a new product forecast, be mindful that new product forecasting should be consider a unique process in its own right and that consideration for its fit into the Sales & Operations Planning Process is crucial. Moreover, the process of new product forecasting should be distinguished from regular demand forecasting given the different information needs and the different personnel involved.

Even with keen consideration of these issues, the reality of new product forecasting is that you're going to be wrong! Usually one will be more accurate forecasting product improvements and line extensions, and less accurate forecasting as a company diversifies into new-to-the-company and new-to-the-world products. As a guide, results of a 2002 published benchmarking study on new product forecasting are offered below (refer to Kahn 2002).

New Product Forecasting Accuracy

Six points to keep in mind when you are doing new product forecasting:

•  There is no silver bullet when it comes to new product forecasting. There is no one technique. Think in terms of a toolbox approach.

•  Forecasting success is a result of data, experience, a cooperative relationship with marketing, cross-functional communication, business knowledge, and being connected to the customer.

•  Participation of analytical business functions like market research and sales forecasting correlate to a more accurate, satisfying forecast. If you do your homework, you tend to be more accurate.

•  New product forecasting requires reconciliation of multiple techniques based on judgment and analytical skills. However, the greater the number of techniques used does not simply increase new product forecast accuracy nor satisfaction with the new product forecasting process.

•  Ranges are important. Don't focus on point or specific number forecasting.

•  Tracking is critical. If you're doing assumptions based forecasting you really need to track your assumptions upon launch so you can understand what is affecting performance.

Dr. Kenneth B. Kahn is the leading expert in the field of New Product Forecasting and is the co-Director of the University of Tennessee 's Sales Forecasting Management Forum. He has published in a variety of journals, including the Journal of Product Innovation, R&D Management, Journal of Business Research, Journal of Forecasting, and, Journal of Business Forecasting, Marketing Management. He is author of the book Product Planning Essentials.