The Importance of Secondary Data in Research

February 1, 2018


There are 210 designated market areas in the United States, but which are the five best ones for rolling out a new marketing campaign? Designated market areas are geographic regions of the country in which the Nielsen Company measures local television viewing and provides demographic data for marketers.

When deciding which areas to target, secondary data is a good place to begin. Using secondary data might not seem as exciting as collecting original survey data or scraping the internet to understand online behaviors, but it should form the core of any research. Secondary data lets you build on existing research, which leads to better results, and saves time and money.

The question of which regions to target for a marketing rollout was asked by a public university that has a large online student body and is considering expanding beyond its regional base. The university wanted us to consider every designated marketing area in the country, but doing so with any statistical precision would not have been cost-effective or necessary. Surveying 210 regions would have required more than 100,000 interviews and cost $1.5 million just to collect the data, most of which wouldn’t be used except to eliminate geographies.

By feeding secondary data into a custom-made algorithm instead, we were able to reduce the number of potential market areas from 210 to 25. The algorithm included:

Census data: A starting point for any demographic research. In this case, census data helped us build a basic demographic profile of each designated marketing area so we could understand the extent to which a target audience is concentrated in each one. We also considered population projections to see if a marketing area is growing or shrinking from a demographic perspective.

Publicly available Nielsen data: Nielsen publishes lots of free data about each marketing area, and we incorporated as much as we could.

Bureau of Labor Statistics data: To understand educational needs and how well the university’s offerings meet possible demand, we looked at unemployment rates and labor trends within the markets.

Syndicated-data sources: One advantage of working for a large research group is the access it provides to syndicated research, which we leverage for the benefit of our clients. Rather than conducting custom research for a specific client, market-research firms fund and conduct syndicated research and offer the data for sale on the open market. In this instance, we used syndicated data to find basic behavioral and psychographic information about each designated market area.

• Academic articles: Articles from academic journals can be dense and boring to read, but the tedium sometimes pays off. We found research in such publications that calculated the average distance people were willing to travel to attend a university. We used that information as a benchmark, and our algorithms favored designated market areas closer to the university.

Using these data, most of which were free, we developed an algorithm that ranked all 210 designated market areas. The richness of the data also let us develop a few alternative rankings for the client. In theory, we could have stopped our research there and given the university reasonable answers for the top-five market areas. But making smart investment decisions for a marketing campaign required numerous ways of collecting and analyzing original data on the top-25 market areas.

The next time you or a client have a research question, start by exploring what kinds of research and data are already publicly available. We are also amazed at how often clients already have data they need to make the right choices, but simply lack analysis. Spending a few minutes to figure out how to use data you already have — or can easily acquire at little or no cost — is the smartest way to start any research project.

Orin Puniello

Orin Puniello is the research director, predictive analytics for Ketchum Global Research & Analytics. Email:


Daphene Prasanth says:

Dear Mr. Orin, The above article for me personally was very informative. It was easy to interpret and interesting. My question on this is: If we choose secondary data for any analysis wouldn't that include biased information and create a risk for accuracy? Thanks, Daphene Student Public Relations and Event Management, Sault College.

March 30, 2018

Post a Comment

Editor’s Note: Please limit your comments to the specific post. We reserve the right to omit any response that is not related to the article or that may be considered objectionable.


To help us ensure that you are a real human, please type the total number of circles that appear in the following images in the box below.

(image of seven circles) + (image of four circles) + (image of nine circles) =



Digital Edition