Editor's note: Frank Pleticha is lead consultant at answers2action LLC, a Brooklyn Center, Minn., research company. He can be reached at frankpleticha@gmail.com. This article appeared in the December 9, 2013, edition of Quirk's e-newsletter.

The garbage in, garbage out principle has never been so true as when it's applied to the design of marketing research. Too often, researchers begin a project with a list of survey questions or a technique in mind. The outcome is typically a long, cumbersome survey instrument, leading to data with little actionability. Worse, the researchers have positioned themselves as techies and not as true, value-added consultants to the marketing decision maker.
 
Here's how to avoid that failure scenario. This five-step research design process has worked successfully for me over the course of a couple hundred research projects.
 
First, ground yourself in the type of marketing decision the company needs to make. Strategic or tactical? Are you exploring new territory or trying to explain something you're already seeing in the marketplace?
 
Secondly, ask yourself: What's at stake here? What's the cost of a wrong decision? If you think of research as a risk-management tool, how large does the research-driven "insurance policy" need to be? If the cost of a wrong marketing decision is only $10,000, you wouldn't want to propose a $50,000 marketing research study to guard against this (relatively) small mistake.  Conversely, if the cost of a wrong decision runs in the millions, then a $50,000 - or even a $100,000 - research study might be appropriate.
 
Third, given what you've learned from steps one and two, ask the decision makers to identify what they really need to know. What data will inform the marketing decision and what is just nice-to-know?
 
Fourth, how, specifically, are you going to use these insights once obtained? I've found it's helpful at this early stage to invent a hypothetical answer and run it past the decision maker as a type of sanity check. For example, if you learn that 62 percent of this market segment doesn't shop online, how will knowing that impact your business decision? If you can't link a potential finding to some type of action, consider dropping that question from the survey.
 
Fifth and finally, what, if anything, do you already know about these issues from secondary sources? Perhaps 25 percent of the answers already exist and you don't need to develop survey questions in a primary research project to reconfirm what's already known.

Valuable partners

It's only after I have answers to these five questions that I start to think about the research technique and the data collection instrument. Using this five-step research design process places the cart behind the horse and helps to ensure that you're seen as a valuable contributing partner to senior-level decision makers.