Secondary Data Sources
Slide 1
Before conducting a marketing research study, it’s best to consult secondary sources to both inform that study or perhaps to provide the data that would have been duplicated by that study, but at far less cost.
Slide 2
Fortunately, there are many resource people, such as librarians, to help with secondary data searches. You’re probably familiar with conducting internet searches using sources such as Google or Yahoo or perhaps a different search engine altogether. You’re also probably familiar with using library databases and searching those as well. There are many non-computerized sources of secondary information and there may be computerized sources with which you’re unfamiliar or unknowledgeable about accessing. Librarians and library personnel are effective in helping you to identify and successfully access that type of information.
Slide 3
What do marketing researchers mean when they speak about secondary data? Secondary data is data gathered and recorded by someone else prior to and for a purpose other than the current project. Such data is often historical, already assembled, and needs no access to study participants.
Slide 4
Databases of secondary data come in four flavors: bibliographic, numeric, directory, and full text. For bibliographic sources, think in terms of bibliographies or listings; although not self-contained sources, they provide listings of possible sources for further query. Numeric sources contain numbers; for example, census data creates lots of numbers that can be summarized meaningfully in tables and graphs. A directory could be a listing of people by profession, or newspapers in the U.S., or realtors in Las Cruces. Finally, for full text, think in terms of online databases like Business Source Premiere. Although some of the cites referenced in Business Source Premiere are limited to articles abstracts, titles, authors, and keywords, much of what’s available is the full text of the original articles. Full text databases exist for newspapers as well and many of those databases have been moved online.
Slide 5
Unlike primary data, which is collected for a specific and immediate research need, secondary data already exists. What are the kinds of secondary data? Researchers divide them up into two basic types: internal and external secondary data. I’ll discuss each of these and their relative merits briefly.
Slide 6
Internal secondary data is data that has already been collected by a company for some other purpose. Typically, that other purpose is associated with accounting, or tracking sales, or tracking back orders, or even tracking customer complaints. That information, regardless of its original reason for collection, can be used for other marketing decision making purposes.
Slide 7
Here, briefly, are the relative advantages and disadvantages of internal secondary data. Under advantages, I’ve listed two things. First, as this data has been collected by a company, it’s been organized in a way that’s consistent with the business that company conducts. As a result, that data should be in suitable geographic and business/product breakdown form to support marketing-related decisions. Second, the data ought to be relatively fresh; for example, sales data should be available from the previous months as opposed to annual sales data from the previous year. Hence, the data is likely to be fresh and in a form that lends itself to marketing-related decisions.
There are three disadvantages of such data. First, companies tend to collect such massive amounts of internal secondary data that handling and analyzing it become difficult. In other words, information overload is a possibility. Second, the inputs to a company’s internal secondary data often are tied to compensation. For example, a salesperson who meets his or her quota within the first three days of a month may try to smooth the entry of those orders into the ordering system; as a result, that salesperson may withhold booking those orders for a week or ten days, which might confound analyses about newly introduced products. Third, it’s company data which likely has been collected for accounting and financial purposes. As a result, it’s in accounting format, and accounting data may or may not lend itself to marketing decision making.
Slide 8
This slide shows a list of some external sources for secondary data, like government agencies and syndicated research services. For the latter, think in terms of Arbitron or Nielsen. Trade and professional associations often collect useful data. I once helped to collect data for a trade association called the Retail Floor Covering Institute, which was located in Chicago’s Merchandise Mart. My mandate was to survey members of the trade association and to create an annual report that all members could use as a baseline for comparing their performance during the previous year. In addition to customized, computer-generated reports sent to every self-identifying respondent, the trade association created an annual report for public consumption.
Professional associations, like the American Marketing Association, collect a wide range of information that it makes publicly available, some even through its website. There are custom research firms that publicize their activities (in newspapers and the like) as well as publish reports in professional journals and other outlets. Books and periodicals are sometimes an excellent source for secondary data. Directories can provide excellent listings of possible information sources. There are vendors who create and sell reports—often at high prices—to interested people and companies. Finally, there’s the internet, which is the source for secondary data I’m sure you know and use.
Slide 9
These next two slides address commercial sources of secondary data. For example, market research companies like A.C. Nielsen provide syndicated data, which is market-level data collected on an ongoing basis about competitors’ sales volume, brand share, pricing, promotion, and the like. (I’ll discuss this in more detail later.) Also, there’s demographic and census updates. Many organizations supply census updates in an easy-to-use or customized form, which allows them to extract additional rents from people who choose to buy this publicly available information from them rather than the appropriate government agency.
Slide 10
Commercial sources for secondary data can provide results from attitudinal and public opinion research. Recall the last election, where almost nightly Roper or some national network conducted an overnight telephone poll to assess potential voters’ opinions about the candidates or the issues. Other commercial suppliers can conduct and make research results available on consumption and purchase behavior, consumer responses to advertising, and magazine/ newspaper/audience loyalty/responsiveness. That said, most of the details regarding such studies are syndicated and proprietary.
Slide 11
Here’s a brief example about how an automobile leasing company used secondary data from Donnelly Demographics.
Slide 12
Although most of you are familiar with most of the secondary data sites listed on this slide, I thought it wise to provide because it might prove valuable to some of you.
Slide 13
Secondary data can be used for at least three different purposes: fact finding, model building, and database marketing.
Slide 14
If fact finding is the objective, then here are three possibilities. Marketers might want to know consumption patterns within a region, a town, a zip code, or a phone exchange. Patterns of consumption may provide marketers with valuable insights; for example, if introducing a product into a region, then superior warehouse site location and distribution decisions would rely on local consumption patterns. Another possibility is to assess trends. By looking at secondary data collected over time, especially data collected by the government, marketers may be able to identify longer-term trends and thus anticipate—rather than merely react to—changes in the environment. Finally, secondary data for fact finding could help with environmental scanning, not merely to track trends, but also to identify fundamental structural marketplace changes that have occurred over time.
Slide 15
Secondary data also can be used to build sophisticated quantitative marketing models, which are far beyond the scope of this course. Nonetheless, secondary data can be used to help to create the type of econometric models used to estimate market potential, to forecast long-run sales, or to select a new trade area or site for introducing a brand to a marketplace.
Slide 16
A third objective for secondary data is to create databases that could be used to track current customers or to develop lists of prospective customers. Such databases would include names, addresses, past purchases, responses to past marketing efforts; often, this data is collected from multiple sources. There may be some ethical issues regarding the aggregation of data from multiple sources and this certainly is an issue in Europe. I’ll discuss that more in my lecture on marketing research ethics.
Slide 17
Now I’ll briefly discuss the advantages and disadvantages of secondary data. Here’s a list of six possible advantages. (1) It tends to be far less expensive than primary data. Think about the large direct costs of fielding a study, collecting data, and analyzing that data in a meaningful fashion. Those costs are why I’ve argued in several lectures that marketing research is an asset; it has a cost and must be thought of in terms of cost-versus-benefit or net value. Secondary data is less expensive because the entities that collected it either did so for some other purpose or will sell it to many people or companies to amortize its direct costs over a sufficient number of entities to make it affordable to all. (2) It’s relatively faster to obtain. Field a research study means designing it or soliciting the parties who’ll be responsible for designing it, fielding it, and analyzing the data it yields—all which take a fair amount of time. (3) It’s already been created and is relatively straightforward to access. Secondary data is available far more rapidly than primary data. (4) It tends to be readily available. There are many ways to access such data, especially if it’s available via the internet or from government sources. (5) It can be used to aid in designing primary research. Secondary data (a) can inform the research, and (2) can indicate the questions that should be asked and the ways those questions should be formulated. (6) It can be used in conjunction with primary data to help understand the bigger picture. Primary data provides some information and secondary data provides complimentary information. When combined, they provide a fuller picture of the environment in question.
Slide 18
Although not all government data is free, much of it is free. A DVD with U.S. Census data will cost roughly $6,000 to $7,000. Government data often is available free, available online, available at the public library, and available at archival facilities like university libraries.
Slide 19
Although there are many advantages to secondary data, there are some disadvantages. (1) The data may not be consistent with needs; for example, people are grouped in an unsuitable way. Media reports group people in accord with key demographics, such as the 35-50 year-old group. That age classification may be too broad for some research purposes. It’s possible that the units of measurement for standard variables like age, income, occupation, and the like, also may be incompatible. (2) The data may be dated. The U.S. Government collects census data on an ongoing basis—rather than only once every 10 years—and in particular conducts a mid-census assessment to examine changes. Thus, even U.S. Census data may be as much as five years old, and that may be far too old for some research needs. Again, think how much Las Cruces has changed in the last five years and whether you’d want to trust five-year-old data to select a site for a new restaurant. (3) Finally, despite all the suggestions I’ll make in the subsequent several slides, it may be difficult to assess the credibility of secondary data. Despite knowing about data quality, and when to trust and when not to trust data, reports may be incomplete regarding the methodology used, and there may be issues that make it difficult to assess the trustworthiness of secondary data.
Slide 20
Ultimately, it’s up to you to assess the credibility of any secondary data you might use in making a decision. Here’s a list of six questions that you should ask yourself about data before you decide to trust it. (1) What was the purpose of the study? It matters greatly whether the study was for public policy purposes or for some commercial venture. (2) Who collected the data? Was it the government? Was it a highly reputable commercial firm? Was it a fly-by-night research organization? The quality of the data is related to the abilities and the credibility of the people who collected it. (3) What data were collected, and do they truly pertain to research needs? (4) Are the data timely? In other words, when was it collected? For example, in Las Cruces, which is changing seemingly almost weekly, timeliness is critical. Five-year-old U.S. Census data may be useless and outdated for identifying a location for a new restaurant. Knowing when data were collected helps to assess its adequacy and timeliness. (5) How were the data collected? To some extent, the methodology is related to quality, and quality is related to its credibility. (6) Is the data consistent with other data? Convergence would be comforting. If you saw a government report, a few commercial reports, and perhaps a newspaper-sponsored report, and they all seemed to converge on the same basic assessment, then you’d feel far more comfortable trusting that collection of converging sources than you’d be if one source indicated one thing and another source indicated an entirely different thing.
Slide 21
For those of you who prefer a flow diagram to serve as a framework for assessing the quality of secondary data, here’s one that has stood the test of time. First ask “Did the data help to answer questions set out in the problem definitions? (I’ll assume yes for all these answers). If yes, then did the data apply to the population of interest? If yes, then did the data apply to the time period of interest? If yes, can the units and classifications presently apply? (By apply, I mean to field research needs). If yes, if possible, go to the original source of the data. If yes, then was the cost of the data acquisition worth it? There’s always a cost-benefit issue with all data collection, even if it is secondary. If yes, is there a possibility of bias? If there’s no possibility of bias, then the data is worth using, but if there’s a possibility of bias, then the issue is “Can the accuracy of the data collection be verified?” If yes, then use it without hesitation, but if no, then conduct some sort of risk assessment. This is a reasonable framework for summarizing the points or the set of six questions offered in the previous slide.
Slide 22
Regarding an assessment of information on the internet, posted on different websites, this checklist of possible questions would help to assess the credibility of information on those sites. These questions are divided into five categories: purpose, authority, scope, audience, and format. This is a reasonable set of questions for making an appropriate assessment.
Slide 23
I would be remiss if I didn’t’ alert you to possibilities of problems with secondary data collected outside the U.S. In addition to all the limitations I’ve mentioned about secondary data in general, there are additional pitfalls in using global data of this type. First, many times that data is unavailable. Although we’d like to believe that the governments of all countries collect appropriate data or at least comparable data to what our government collects, some governments don’t collect such data because it’s cost prohibitive. As a result, it’s unavailable. Hence, certain types of data available in the U.S. or in other western countries to make marketing decisions may be unavailable in non-western countries. Global secondary data also may be less accurate due to less oversight in tabulation and poorer data collection methods. It also may be far more dated than what’s available in the U.S. There’s national census data collected every ten years and intermediate efforts to collect and update census data. In other countries, the equivalent of census data may be far less timely.
Slide 24
As a final warning about secondary data, be careful when that data is presented as either percentages or as index numbers. Percentages can be based on large or small populations, and large or small samples. With very large or small samples/populations, seemingly meaningful percentage differences may not be statistically significant, let alone managerially meaningful. Be careful that percentages, if judged to be meaningfully different, are based on truly meaningful samples or populations.
As for index numbers, they’re often computed relative to a base year or a base period, and those bases can shift over time at the researchers’ or data reporters’ convenience. Although this dates me somewhat, I recall when the Consumer Price Index was based on 1967 prices. Later, I recall the U.S. Government, after several years of Jimmy Carter and inflation rates approaching 20%, shifted the base year to 1978. I suspect the base year is now some time in the 1980s—during the Reagan administration—when inflation shifted to a much lower annual rate. Be careful with index numbers, especially government index numbers, are comparing numbers from identical base years. Otherwise, those numbers must be converted appropriately.
Slide 25
To recap, this lecture addressed five issues about secondary data. First, it defined secondary data and differentiated it from primary data. Second, it summarized the different types of secondary data more broadly and differentiated internal from external secondary data. Third, it briefly indicated how secondary data could be used by marketing researchers. Fourth, it briefly explored the relative advantages and disadvantages of secondary data. Finally, it provided a few broad guidelines for evaluating whether or not secondary data is appropriate.