Highlights of main Forecasting methods for New Products

In a changing environment, forecasting the demand for new products and services is key and very important in light of the investments put behind a launch. In addition, with strong economic competition around the world, evolving customers expectations, the emergence of new technologies and innovations, estimating and forecasting the potential market is increasingly challenging and risky as shown by the estimated failure of circa 70-80 percent of new product launches in the consumer-packaged goods (CPG).

Forecasting the demand of new products is often done using judgmental methods – surveys of buyer’s intentions, Delphi method, market test, etc … – or statistical modelling through a combination of time series and /or multivariate techniques.      

However, thanks to the preponderance of data stemming from the information explosion in all activities, forecasting through Machine Leaning has now gained more importance than ever. ML could provide key solutions to many issues linked to statistical modelling methods and could be assessed as an alternative to statistical models for time series forecasting.

Forecasts are Key !

New Products and Services are the lifeblood of any firms. Without them, the firm will face its decline stage and either dies.

Key reasons for developing new products: change in market and in technology, new customer expectations, increasing competition, diversification of risk, growth and development ….

According to Harvard Business School Professor Clayton Christensen, each year more than 30,000 new consumer products are launched and 80% of them fail.

What causes 80% of all new products to fail is mainly “a lack of preparation: Companies are so focused on designing and manufacturing new products that they postpone the hard work of getting ready to market them until too late in the game.”(Joan Schnieder and Julie Hall in  Harvard Business Review, https://hbr.org/2011/04/why-most-product-launches-fail).

One important issue of the failure is a lack of forecast.

New product forecasting is challenging compared to forecasting demand of existing products: the first has not or very limited historical data available since the second has enough data to run forecast.

However, because many industries are facing shorter product life cycles, new product forecasting gains importance (Van Steenbergen, 2020). Besides the challenges regarding the lack of historical data, there is limited analysis time and there exists a general uncertainty related to consumer acceptance and competitive reactions (Van Steenbergen, 2020) 

Forecasts are essential for the operations as they will help to:

  • Predict and plan for demand throughout the lifecycle period
  • Reduce the risks inherent to any launch of new products
  • Make wise investments and assess potential risks
  • Adapt manufacturing, planning and sales Processes

 

Because these decisions are guided by forecasting, a proper sales forecasting approach is key to prevent complications during or right after the product launch. Poor forecasts can result in stock-outs or overstock situations, which have a direct impact on the company’s profitability and may also decrease customer satisfaction and market share (Van Steenbergen, 2020).

While classical techniques (Judgmental, Market Research, Cause – Effect) have been used for decades, Artificial Intelligence, Machine Learning and Deep Learning have gained considerable prominence over the last decade.

Categories of New Products

New-to-the-World

New-to-the-World products are only the true new product category within the whole “New Products” family. These products are completely new because they lead to a full new market. They represent around 10 to 15% of the new product category. They are also the most challenging of all. Examples: GSM, smartphones, microprocessors, ….. We don’t have any data  

New-to-the-Firm Products

Also known as New Product Lines, they take a firm into a category new to it. The products are not new to the world, but are new to the company. 

We don’t have internal data — external data may or may not be available.

Additions to existing Product Lines

They represent extensions of already existing products, designed to flesh out the product line as offered to the company’s current markets. Example: Airbus 318 and Airbus 319 derived from Airbus 320.

Data  of existing products are available.

Improvements and Revisions to Existing Products

Adding new product features and/or improving the existing ones would increase benefits realized by existing customers. 

Data  of existing products are available.