Garbage In/Garbage Out: Your Forecast Output is Only as Good as Your Inputs

Clients of any marketing research company can aid their research partner by providing the most accurate information to be used for forecasting inputs.


Proper, realistic forecasting that appropriately estimates true adoption and revenue has been the holy grail of marketing departments since Henry Ford started mass producing vehicles for the common man.

One of the key values of forecasting is that there are actual numbers that can be referenced. No anecdotes, no open ended feedback – just raw numbers that makes my OCD left brain happy.

On the other hand, forecasts rely on data that you, the market research consumer and co-collaborator, must bring to the table. Some of this data is market intelligence you have (or should have if you don’t), and some of this data is based on how your company does business. In both cases, if this information is not based on actual data or your assumptions are incorrect, the value of your forecasting model can be called into question – and may lead you down a less-than-desirable path.


Forecasting Is Not a Solo Activity

So what does this mean to you? Simply put, your research partner is likely not the expert in the intricacies of your company, your product or service that you are. While good marketing research firms should have solid industry backgrounds and a firm grasp on your product and benefits, their knowledge of your company and products is limited. YOU are the expert. You know your customers, competitors, roadblocks and opportunities best. And, most importantly, you know your own company.

When designing your forecasting model with your research partner, be prepared to answer some key questions:

  • Who are you selling to?

  • What is the total opportunity?

  • What impact will my company’s operations have?


Know Your Target Before You Start

Who is your customer? Is there something special about them that makes them more or less interested in your product? Are they price sensitive? Are they tech-avoidant? Failure to properly know your customer before you forecast might yield a forecast that underperforms due to poor targeting, not necessarily a poor product.


Know the Current Market

In order to understand how much revenue you might realize, you need to understand how much there is available. Improper assumptions concerning the actual size of your potential market will yield a forecast without proper context. In other words, you can’t squeeze $6 million out of $4 million market. Those who develop your forecasts must have an accurate view of the addressable market within your reach. For example, while there may be 130,000 dentists in the US, you likely will not have the ability to reach all of them, or even a majority of them.


Know Your Sales and Manufacturing Capabilities

No matter how good your sales team is, odds are they can’t get to all of your customers in the first month. Is your manufacturing team ready to produce enough widgets to satisfy your forecasted initial demand right away? Most often the case is, manufacturing will likely be able to handle demand, but the sales and marketing efforts will need time to make meaningful connections to all current and potential customers within your sales reach.


The Best Forecasts Start with Good Information

Forecasts for new products are in large part based on self-reported likelihood of purchase and other questions, collected anonymously in a relative vacuum. There will always be a certain amount of statistical variability in the estimates, which can be accounted for through a Monte Carlo simulation. If you start with good information, however, you decrease this variability and increase surety of the forecast.


Your research company must have your best information and effort to provide an accurate forecast.

In addition to purchase intention related inquiry, there are several other components to a good forecast, such as delivering concept test stimuli that effectively display the new product and its value, questions that properly address the buying cycle and barriers to intended actions, as well as other calibration-type questions. Your research partner should bring those to the table, customized to your unique market and product. They should also ask you questions concerning your target audience, the total market, and how your sales and manufacturing handle product launches. If they don’t ask these questions, you may wish to consider a service provider with more experience.

You, the forecast consumer, is indispensable to the forecasting process. Through either your own input, or through the input of other key stakeholders within your company, be prepared to offer the additional market and operational data the researcher needs to generate your forecast. The quality of your forecast is in many ways directly related to the quality of the information you provide.