MARKETING HIGHER EDUCATION A periodic electronic Newsletter to help you market your school, community college, college, or university. Vol. XII, no. 10, October, 1998 -------------------------------------------------- TABLE OF CONTENTS Market Research: Using information to market your institution. by Robert "Bob" S. Topor, Topor Consulting Group International Seven Deadly Sins of College Web Sites by Professor Elizabeth "Liz" Pollard, Smoke Signals Enterprises -------------------------------------------------- Market Research Using information to market your institution. by Robert "Bob" S. Topor What is market research? It is any kind of research that gives you information about the current state of your particular market and offers guidance in improving your position in that market. In most cases, it will involve opinion research, employing scientific sampling and interviews or written questionnaires. Educational institutions most often use market research as a way to: (1) increase the effectiveness of their fund-raising efforts, or (2) attract more (or better) students, or (3) improve their image in the competitive marketplace. Even image studies or projects that are overtly intended to solve some immediate problem usually have money or students as their final goal. State and public institutions may direct their fund-raising efforts - and related market research - at legislators, rather than private sources. Whatever the goal, market research can help you reach it - providing you follow certain rules. The first is to define the goal very clearly at the outset. A clear definition will help you keep the project on track from beginning to end. It will assure that the appropriate groups are sampled; that the right questions are asked; and, finally, that results are analyzed and interpreted in the most productive way. Careful planning on your part can also help to keep costs to a minimum. Know what you want, then resist impulses to add "just a little more." Add-ons lead to cost overruns - usually big ones. Almost everyone adds on in the course of a study and then is shocked by how far the final cost soared above the original estimate. Should You Do It Yourself? There are two basic ways to conduct market research. You can do it yourself, or you can hire someone to do it for you. Doing it yourself has several advantages. After all, who knows more than you about your institution's problems and needs? You can call on in-house expertise: marketing professors; statisticians; your own computer center; the graphics, copy, and printing departments in your PR and publications offices; secretarial and clerical staff; and so on. In-house research should be significantly less expensive. On balance, though, all of these are offset by some major disadvantages. An in-house study often lacks credibility. People both on and off your campus tend to be skeptical about the objectivity of a survey created and analyzed by staff members who may have a vested interest in the outcome. The in-house team is not likely to view the institution with the detachment you could expect from outsiders. Not being able to see the forest for the trees, the insiders may write questionnaires that do not address the real issues, or, even worse, that serve mainly internal political purposes. Also, inexperience may lead an in-house team to waste much time and effort debating such things as format and wording of questions. An outside professional firm brings credibility, expertise, experience, and objectivity to the project. But it does so at a price - a high price in some cases. A compromise arrangement, combining outside expertise and inside cost-cutting services, is possible. The outsiders can be used for such sensitive jobs as sample selection, questionnaire development, and analysis and reporting. The institution provides computer services, printing and mailing, secretarial and clerical services, etc. This can provide professional expertise and credibility at a reduced cost. Explore Secondary Sources Once you have defined the problem and decided that a market research study is indicated, you must find the most effective way to go about your study. Before embarking on expensive primary survey research, though, you should explore all secondary sources for the required information. This kind of research often saves time and money. At the very least, it can help you write a more productive questionnaire. Secondary information sources fall into two categories: external sources and sources internal to your organization. You may find valuable data that has been collected on an earlier survey, or information that was collected routinely from students or donors. You should explore all internal sources; they will be by far the least expensive. If internal sources prove unproductive, begin checking external secondary sources. Academic, trade, and professional journals are logical starting places. The federal government is also a rich source of information: the Department of Education, the American Council on Education (ACE), the Council for Advancement and Support of Education (CASE), the Department of Commerce (Business Conditions Digest, Survey of Current Business), the Department of Labor, and the Census Bureau are key ones. Another useful source is the Dialog data base system, which is a computer-based reference library of scientific, trade, general interest, and governmental publications searchable by keyword. The results of many surveys conducted by or for the government are in the public domain. The Findex index of market research cross references most of this information. Suppose, for example, that you wanted to target the most likely geographic areas for sources of new students. The Census Bureau makes available census demographic information consolidated by zip code. Your applicants for the past two or three years could be overlaid on each zip code. Your computer could then determine statistically the characteristics of the type of zip codes that were the best producers of applicants. Those and all "similar" zips should then be your most productive potential student sources. Role of Qualitative Research If the information you require is not available from secondary sources, then you'll need some type of primary survey research. Primary research can be subdivided into two major categories: quantitative and qualitative. Quantitative research yields numerical results of some type-counts, intensity, penetration, etc. Qualitative research yields information that is "softer" and is usually not numerical. Qualitative research is employed most often as the first phase of a projectable, quantitative research project. The purpose of qualitative research is to set the agenda for the subsequent quantitative research. The most widely employed qualitative technique is focus group research. Focus group research consists of a structured discussion led by a trained moderator in a conference room setting with five to 12 members of the target universe. The purpose of a focus group is to elicit from the target universe members their perceptions of the issues germane to the research question. Often, the organization generating the research project has preconceptions about the issues that the research should cover. The key issues within the marketplace, however, may be entirely different. Suppose, for example, that student applications at institution A showed a significant drop from the previous year and that during the same period the economy experienced a slowdown. A survey at this time would very likely tend to emphasize economic issues and price sensitivity. A few focus group sessions, however, might reveal that, while the economy was a concern, parents and students were looking for a more career-oriented curriculum than institution A offered. Carefully conducted focus group research will help ensure that the quantitative survey research covers all the key issues. Another important use of the focus group environment is to pretest a questionnaire for use in subsequent quantitative survey research. Walking through a questionnaire with five to 12 members of the universe that will receive the survey can prevent expensive errors. It will ensure that everyone understands the questions, that the questions are not ambiguous, and that jumps and skips from one question to another can be followed. And, because fixed-choice questions should be used as much as possible (and open-ended questions avoided), you can guarantee that all possible responses are included on the questions. You should avoid at all cost the tendency to read more into the results of qualitative research than is really there. Because focus group results are based on the opinions of very few people, they are not projectable to the target universe. Results of such research are perfectly adequate for defining the issues and pretesting the questionnaire, but only for these purposes. Rarely, if ever, are qualitative research findings reliable in and of themselves for making marketing decisions. A good resource for you is my book titled "The Complete Guide to Focus Group Marketing Research in Higher Education," available from my web site, under Educational Catalyst Publications. The web site is located at http://www.marketinged.com Quantitative Research Qualitative research usually leads into a quantitative research phase. Quantitative research can be divided into two categories: projectable research (called "probability research" from here on) and nonprojectable quantitative research. As the name implies, probability research results can be projected to the target universe with known precision. Because nonprojectable research is conducted with a convenience sample instead of a random "probability" sample of the target universe, you can't generalize from the numerical results. One type of nonprojectable quantitative research is carried on at a central site. "Mall intercepts" comprise the most common type of central site research. Market research companies are now in many regional shopping centers across the United States. The companies contract to conduct surveys among shoppers in the shopping center. Trained interviewers recruit and screen potential respondents in the public mall areas-not in the stores. Short surveys may be conducted on the mall floor. For longer interviews, the respondents are taken to an interview room. As opposed to focus group research, nonprojectable quantitative research can yield actionable marketing information. For example, suppose your admissions staff had ideas for two different brochures directed to the parents of potential students. With mall-intercept research, the most effective of the two brochures could be reliably determined with minimal expense. Of all the adults approached in the mall, only those with college-bound high school students would be interviewed. Qualified respondents would be asked to rate each brochure separately and then to pick their favorite. (Half would be shown brochure A first, half would be shown B first to control for what is called "ordering effects," a reaction based on the order in which people encounter things.) These results would be sufficiently valid for selecting the most effective brochure. It may even be desirable to find the "best" brochure for different cities, e.g., brochure A may be better in Denver, while B may work better in Miami. Central site research has several advantages over other types of quantitative research: (1) it can be completed in a relatively short time; (2) it is relatively inexpensive (provided that the screening criteria are not too restrictive); (3) visual props can be easily employed, in contrast to mail and telephone surveys; (4) large amounts of information can be collected; and (5) complex questionnaires with involved skip patterns can be used because the respondent does not have to contend with following the questionnaire. As in qualitative research, one must avoid reading too much into nonprojectable quantitative research. While adequate for answering many controlled experimental questions - as in the above example - results may not reflect the views of the target universe on sensitive issues. Projectable or probability quantitative survey research yields results that can be projected to the target universe with a known precision. The three types of probability survey are mail, telephone, and personal (or in-home) surveys. Each type has particular advantages and disadvantages, usually involving project cost, turnaround time, potential biases, and data volume capacity. For any given project, the constraints of the situation usually mandate one of the three techniques. Before examining each technique in detail, let us briefly examine sampling and questionnaire design. Sampling The device that makes projectability possible is proper sampling. The purpose of sampling is to obtain information that is representative of a universe while examining only a portion of that universe. You should carefully target the universes to be studied. Don't sample the general public if your purpose would be better served by a survey of opinion leaders or of parents of students. Some other likely targets: alumni, faculty, students, prospective students, those who applied to your institution but attended another, school counselors, legislators, donors, potential employers and media "gatekeepers." Compared with taking a census, sampling has several advantages: (1) it is significantly less expensive; (2) it is faster; (3) it is more accurate (paradoxically); and (4) it allows time to collect more in-depth information. Sampling has two disadvantages. First, information is not collected on everyone. (Sampling could not be used to calculate electric bills.) Second, sampling "error" is introduced. Sampling error arises because not everyone is measured; it is the price paid for using a sample. However, the size of this sampling error, which depends upon the sample size, is known; it is referred to as the "precision.". If, for example, the total size of the group (universe) to be studied is 100 and 5% precision (plus or minus) is acceptable you need to sample 79 people. If only 10% is acceptable only 49 people are necessary. Likewise, if your universe is 1,000 people and 5% precision is acceptable you need to sample 278 people. For 10% precision, only 88 are necessary. For a universe of 10,000 people and an acceptable 5% plus or minus precision, 370 people should be sampled. For only 10% precision 95 people would do it. For 100,000 universe and plus or minus 5%, 384 people are necessary. For 10% precision only 96! In most cases, proper sampling means simple random sampling. That is, every member of the target universe has a known and equal chance of being included in the survey sample. In practice, systematic sampling is an adequate substitute for simple random sampling. Systematic sampling is also known as nth name sampling because the total universe (i.e., population) size is divided by the desired sample size to yield the increment, or n, to be used. For example, if you want a sample of 400 from a universe of 4,000, you will need to sample every l0th person. But don't start automatically with one and go to 10. From a computerized random number table (or by drawing numbers from a hat), select a random number between 1 and n inclusive, which is 10 in this case. Let's assume that number is six. From a list of the universe, select-or have your computer select-the person corresponding to this random number-the sixth person-as the first member of the sample. Subsequently, select sample members by taking the nth (l0th in this case) name from the first selected name. You must also remember that you should generally get at least a 50 percent response rate to your survey. That means you would have to double the sample size to determine the number of questionnaires that you would send. (Both factors are vital: the sample size and the response rate. For a universe of 10,000, you would need to send out 750 questionnaires, assuming a 50 percent response rate, to get a sample of 370. You couldn't send out 1,480 questionnaires and be satisfied with a 25 percent response rate, even though you would receive 370 completed responses.) If you need to make statements about subgroups of the sample with precision equal to that of the sample as a whole, the sample size for each subgroup must be determined separately. In effect, a separate survey is conducted for each subgroup. This is called stratified random sampling, When the survey results are being processed by computer, care must be taken to weight each subgroup back to its proportionate size in the universe at large. You will need some expert help from your computer people on this. If the response rate is too low, a significant potential for error is introduced. The error is called nonresponse bias. If 70 percent of the sample does not respond to the survey (i.e., a 30 percent response rate), there could be something about the 70 percent that makes it different from the 30 percent that responded. For example, suppose a survey on gun control was mailed to a random sample of the United States population. Suppose further that people opposed to gun control do not like to answer surveys. The results would indicate much stronger support for gun control than really exists in the population as a whole, because the non-respondents as a group were more opposed to control than those who responded. If you are familiar with the universe being surveyed and it is relatively homogeneous, a 40 to 45 percent response rate is adequate, provided that the demographics of the respondents match known demographics of the universe. If there is any uncertainty about the overall mix of your universe, you should aim for a response rate of 65 percent or more. You can achieve a response rate of this magnitude cost effectively by careful application of the techniques described by Dillman in Mail and Telephone Surveys: The Total Design Method. ********** Thank you for taking time to read this very important article. Marketing research is one of the most important things you can do to market your institution. I am available for answers to your questions. Contact me by e-mail (topor@marketinged.com) or give me a call at 650-962-1105. Happy researching! ********** Bob Topor * Marketing Evangelist Topor Consulting Group International E-mail: topor@marketinged.com ********** -------------------------------------------------- SPECIAL OFFER For Marketing Higher Education newsletter subscribers ONLY. Download the updated second edition of Bob's "classic" book, Marketing Higher Education - A Practical Guide, directly from his Web site at http://www.marketinged.com Special Deal: As a subscriber to this newsletter you have permission to make copies and distribute on your campus... a great aid for marketing committees! Make as many copies as you like (limited to your campus). You can get this book from the Home page on Bob's web site. It costs $40 (U.S.) It is easy to download to your computer, then you can reproduce it in your print shop or make photocopies. The first edition of this book has been one of CASE's best selling publications and has been used around the world. If you have questions call Bob at (650) 962-1105. NEW! "The Complete Guide to Focus Group Marketing Research in Higher Education" book is now available for downloading to your computer... Bob Topor's 55-page practical guidebook for running focus groups is now available for downloading. As subscriber to this electronic newsletter you have permission to purchase this book and make unlimited copies for use on your campus (copyright free). Regular cost is $32 in printed copy. You can purchase it for only $25. It is a great guide for how to do focus groups and has been Bob's best selling book ever! Don't miss this special offer! It's easy to order (secured credit card) and download... just log on to Bob's web site at http://www.marketinged.com and find it on his home page. Questions? Call Bob at (650) 962-1105 or e-mail him at topor@marketinged.com NEW! Download "Wasabi & Ginger" ($35.95 US) .... a book by Bob and his partner Dr. Moshe Engelberg .... for life development. It combines ideas from the business world with ones for personal satisfaction and success. You can use these ideas both in your business ventures as well as your personal life and success activities. Bon Appetit! This book is easy to download from Bob's web site: http://www.marketinged.com -------------------------------------------------- Seven Deadly Sins of College Web Sites by Elizabeth "Liz" Pollard In the course of our work reviewing Web sites for colleges and universities, Bob Topor and I have examined thousands of them. Some of the most common mistakes have become obvious to us, not only mistakes in design, but those in marketing principles as well. This article is a sampling of some of the worst problems we've found. Some are fairly simple errors and easy to correct, others may require some ingenuity and policy interpretations at your campus. All, however, will affect the marketing effectiveness of your site. We present these here with descriptions and explanations as a few caveats to keep in mind as you develop your college Web site! 1. Graphics -- Large graphic menus and image maps should have a text equivalent, presented either simultaneously or as an alternative. Especially in dial access mode, large graphics tend to load very slowly and may be completely skipped by some users! Many prospective students and donors may not have the latest computers or modems, and such graphics are discouraging to them! If graphics are the only source of hypertext links, the user has no way to navigate until they finish loading and will give up! One of the worst examples of this problem is the huge, slow opening graphic which is nothing but a logo or scenery, with no links present except one which leads the user to the REAL Home page! There is no other way to get there! There are techniques available which can reduce the size of many Web graphics and make them load faster. If you must open with such a a page, these should be used to minimize loading time. In addition, be sure there is an alternative way to get to the Home page so the user doesn't have to wait until the whole graphic loads in order to enter the site! 2. Menu items with no links -- Items presented within a menu structure are often left without links until a page is completed. When this happens, it confuses the user, who expects menu items to have links, and not entirely unreasonably so! Such false leads waste a user's time and discourage further exploration! It's better to leave out a menu item until later if this is the situation. Almost as bad are links which lead to the WRONG place or are so outdated that they produce an error message. They give the impression of sloppy maintenance! 3. Contact information. -- If you want users to contact you for further information, to apply for admission, or to make a donation, they must be told how to do so, and the information must be simple to find or they will go elsewhere. Ideally, contact information should be available on every page, but there should at least be clear pointers to a page where it may all be found. At the very least, the Home page should display clearly the name of your institution and the city and state of its location. Don't assume that everyone knows automatically where you're located! In some cases, there is no contact information anywhere at the site. In others, one must go deep into the menu structure to locate such data. Prospective students expect to be able to contact the institution for further information, catalogs, applications, and similar materials without spending hours looking for contact information! Make it simple for them! 4. Email addresses. -- Email contact information is especially conducive to rapid communications, as it is convenient and inexpensive. This makes it especially appealing to prospects who are already on the Web. Use it to your best advantage! Provide email links from your site to critical offices or individuals! If you present an email directory of faculty and/or staff, remember that most users will have no personal names to start with and provide an index or search capability by department or position. Interactive forms which email automatically with the click of a mouse are nice, but since not all end user equipment and software is created equal, it's helpful to provide an email address on the form in case the user can't email directly from his browser. 5. Programming blunders -- Examples: A "Home" button that goes nowhere or is difficult to identify; A "Back" link that goes Home instead; links that have been allowed to go out of date. These make logical navigation and investigation almost impossible, and they leave the impression that the developer (and perhaps his institution as well!) doesn't know what he is doing! Links should be tried and tested over and over again before the page is mounted on the Web to make sure things work as intended. Likewise, as things are changed and updated, links should be tried again at regular intervals to make sure their intended targets haven't changed or disappeared! As noted above, it leaves a sour taste and an impression of sloppy maintenance to find yourself elsewhere than expected or presented with an error message! The moral of this story is that a professional to program the site or review it, as well as to maintain it, may be well worth the money which would otherwise be lost in admissions or donations which escape you! 6. Text Links. -- Use of links within text can be very useful to guide the user from page to page on a particular topic. However, if they aren't obviously intended for such use, no one will ever click them! Short, one word links or links without sufficient cues may be missed entirely in a long section of text. Make sure enough text is used (i.e., two or three words, not one, perhaps even capitalized) to be perceived as important words. It just isn't nice to mislead people or give them no guidance at all! The user has to search for links to where he wants to go and will frequently give up in disgust! 7. Confusing Navigation Clues. -- A good example is the proliferation of added-on pages without a unifying theme, logo, or appearance to guide the user, and sometimes without clear connections to the Home page. Each segment of the site should be connected clearly with navigational buttons, preferably placed uniformly on each page, and should have a unifying theme (graphic look, logo, or similar organization) to make the relationship clear! A "button bar" placed prominently at the top of a page can accomplish the purpose nicely, but it's helpful to repeat the choices at the bottom of a page to minimize the need to scroll. At each point in a logical presentation, there should be a clear way to go "back" without using the "Back" button on one's browser. In addition, "Back" within the context of your presentation may not be where the browser "Back" button would take one! In other words, don't leave the user guessing where he should go next when you can guide him subtly and enhance your presentation in doing so! ********** Liz Pollard * Smoke Signals Enterprises Web site: http://www.marketinged.com E-mail: lpollard@smokesig.com ********* NOTICE: As a subscriber to this electronic newsletter, you have permission to reproduce and use this article on your campus. All others please note ©1998, Topor Consulting Group International. Comments about, or requests to reprint should be directed to Bob Topor at topor@marketinged.com. ********** -------------------------------------------------- Original posting: 10/29/98 Marketing Higher Education Newsletter is published by Topor Consulting Group International (http://www.marketinged.com). Newsletter posted by WEBB Internet Marketing & Consulting (http://www.firstchapter.com). copyright 1998 Topor Consulting Group International