30-Second Briefing

By Ed de la Fuente

A:  This is a great question because most market analysts and marketers wonder where these projections come from when they see a report that offers very little explanation of the methodology.  I’ll assume that you are asking either because you want to make some projections yourself, and are looking for first-rate methodology, or you are using various forecasts in planning the development of a new product and are trying to assess the variance in the forecasts you might expect.

First off, there are as many approaches to forecasting as there are analysts. Now multiply that by the different types of industries or products for which there may be a forecast. This leaves too many variations to discuss in detail in our “thirty seconds” here, but we can certainly look at some of the general methods used.

How Analysts Forecast Market Sizes

In my experience, these are the three most common forecasting methods:

1. Expert Consensus

There are several variations on this method, but the basic technique requires the combination of two or more industry “experts'" assumptions and projections.  These projections are discussed and revised until a consensus is reached on a final set of industry forecasts.  In one variation, known as the Delphi method, a group of industry experts is surveyed both for their insights on factors affecting consumption as well as the estimates on future demand.  The experts are industry players such as suppliers, manufacturers, and service providers. The result is a combination and analysis of this input.

2. Market Survey

Some analyst groups will estimate demand by surveying a sample of the expected customer base to determine their “purchase intent”.  This is often difficult for customers that may not be familiar with the new product under investigation.  Additionally, there is always a percentage of respondents who say they will “very likely” purchase but in reality don’t.  Typically, you need to reduce the “somewhat likely” and “very likely” to purchase by about 35%.

3. Technical Modeling

This includes using historical data and leading indicators to mathematically project demand. This technique is not often used for new products since there is little data available to create these forecasting models.  One technique that may work is identifying a “proxy”, a current or past product that addresses the same market, or provides a similar benefit to the product under investigation. The historical adoption curve can be used to model your new product’s demand. 

New Product Forecasting Basics

At the core of every forecast method is the need to identify the assumptions that will determine, 1) the target market size, and 2) the consumption rate of your product and service within the market size. The forecast is then a logical application of these assumptions.  Market size is the population of potential customers in the target segment and its sub-segments.  Consumption is the set of purchasers (and frequency of purchase for certain products) for the product, usually stated as a percentage of the market.  Consumption rate is also commonly referred to as “adoption rate”, “absorption rate”, or “penetration rate”.

A word of caution: When accepting these rates from analyst reports, make sure you know which population the consumption rate actually references.  For example, in the telecom industry it is easy to misinterpret the projections of a new service since the target market population may be referring to any of; a) all households in the service area, b) current base service customers, or c) only customers to which the new service is available.

Forecasting Tip

Since forecasting new product sales is basically an educated guess, it is best to develop your forecast at three levels: low, expected, and high.  As you identify the factors upon which to base your projections, consider how each variable is going to impact your forecast so you can develop projections for each range.  This will also give you an idea of how wide the actual variance might be.  The level of detail you use will depend on which stage of development you are in; i.e. less detail at concept, and more detail as you get further into development where you will be able to collect more market information.

Also be aware of the "garbage in, garbage out" syndrome. The more assumptions you use in your forecasts, the greater your margin of error. Soon you will have a meaningless forecast that is not defendable. 

Remember, even the best analysts still consider new product forecasting more of an art than a science. But if you start with accurate target market populations and apply relevant assumptions to determine valid consumption rates, you will create credible forecasts that can be used to successfully launch your products.

Done