Taking the good with the not-so-good
Editor's note: Pete Cape is global knowledge director at SSI, a Shelton, Conn., research company.
Normally in research, the time pressure is on the researcher; this could be to develop the questionnaire more quickly or complete the fieldwork one day sooner. On the telephone the respondent faces fewer time pressures, save those they impose on themselves in agreeing to take the survey when they perhaps ought to be doing something else. Once in the survey itself all the control of time is in the hands of the interviewer. They set the pace, ask the next question when they are ready and are the only ones in the conversation who truly know how much longer it is going to take.
But not in business-to-business (B2B) research. Here, the time pressure is all on the respondent. We contact them at their place of work – so not only do we make demands on personal time, we also ask for company time in which to fill out the survey. Any experienced B2B interviewer can attest to the time it takes to make that first appointment to even speak to the respondent and the persuasive effort required to book the time for the interview itself. Given the pressure to do real work, it is not surprising perhaps to find that our surveys are not the No. 1 priority on the B2B respondent’s to-do list for today.
A way out
Online interviewing offers a way out of this. The online survey can be stopped and started in the respondent’s own time, completed to their rhythm. They have all the time they need to contemplate the questions and think about their answers. Take this one step further: They have no need to complete the survey during working hours at all. Once at home, the doctor doesn’t stop being a doctor, the CEO is still a CEO, etc. Another step: Once the survey is being done in consumer time, then the topic and nature of the questions can be more consumer-styled. All those questions about ads seen or brand imagery and association suddenly don’t seem so pointless when you are sitting on your sofa rather than at your desk.
If online offers the solution, why is it not the universal solution? The short answer is sampling or rather the lack of ready access to the B2B samples that researchers need.
Of course there are many consumer access panels and if B2B people become “consumers” after 5 p.m., then why can’t we simply use consumer access panels as the source?
For many target audiences we can. Around 46 percent of the U.S. population is in employment, 66 percent of those aged between 16 and 64. Unfortunately, the vast majority are of only occasional interest to B2B researchers and clients. Teachers, nurses, shop workers and secretaries, builders and clerks, preachers and refuse collectors: All do valuable work but their opinions are not often sought out. The exclusive nature of the most sought-after B2B respondents – IT professionals or C-suite executives, for example – can render the consumer access panel useless for the researcher’s needs.
Another factor weighs against the consumer access panel: the nature of B2B sampling. More often researchers are interested not only in the views of top executives but top executives from the nation’s largest companies. A consumer access panel would have to be vast to have any chance of including any of our target audience, assuming that a top executive would be even remotely interested in, or have the time for, being a member of a consumer access panel. Traditional telephone approaches to B2B sampling guarantee you are at least knocking on the right door and you know the right person exists behind it, even if it is hard to get someone to answer.
When our target audience is more mundane, the consumer access panel offers huge advantages over traditional B2B sampling procedures. Traditionally a researcher would sample companies and ask for the target person within the company. Imagine the target is secretaries. Some (most, perhaps) companies have more than one secretary. On the call being answered by the switchboard, how does the switchboard operator decide which secretary should be selected? Alternatively, how does the interviewer decide which one to ask for? In either instance the practicalities impart a systematic bias on the sample.
Another bias also results. Your final sample will be biased towards secretaries who work in smaller companies. To understand this you need only think of the probabilities of selection for the study. Assume any single-company telephone number has a one in 100 chance of selection. A secretary working for a company where s/he is the only secretary has a one in 100 chance of selection. A secretary working in a company where they are one of 100 secretaries has only a one in 10,000 chance of being selected. This is in essence the same problem that consumer researchers struggle with in telephone sampling and that is solved by using Kish grids. Imagine having to ask the switchboard operator for the name of every secretary in the company before making a random selection!
In traditional methods then, the sampling of companies is the first stage and the absence of random procedures at the second stage can result in a sample that is not representative. Using access panels the target person is taken out of their company and put in the sample frame along with all other people who do the same job. The selection procedure is single-stage with equal probabilities. Assuming the propensity to join an access panel is not related to employment choices, then the sample is representative. The same would be true for all job types were it not for the fact that they are scarce and there is a bias away from joining access panels.
In order to provide online access to specialty occupations, it is necessary to build specialist B2B panels. This is not without its problems.
The first problem is that members of the panel may not have enough surveys to complete. Many B2B audiences are irregularly researched, so any standing panel would simply melt away over time. While of course it is possible to replenish the panel and/or maintain it with dummy work, this all adds to the cost for using a panel and increases the risk on the part of the panel company who builds it speculatively. This problem has a domino effect on sample-size feasibility. Panel companies are unlikely to have standing panels large enough to undertake very big samples of very particular job titles.
The second problem relates to people moving jobs. Person A, the IT director at Corporation X is of great interest and value but only because he represents Corporation X. As soon as he switches to a new and exciting role at Startup Y, he is no longer of any interest. This turnover leads to update and replenishment costs.
The need to replenish and maintain also brings opportunities to enhance the profile of the panelist. Panel companies can find out more about their panelists, moving beyond job title to role functions perhaps, bringing a deeper and/or broader understanding of the person and allowing for more targeted access.
Finally, we have the unpalatable side of all access panels: the presence of fraudsters. Small in number, in total they are disproportionately attracted to the relatively highly-rewarded B2B studies. The effort involved in catching these people, should they get past our barriers, can be substantial.
What are the solutions?
We’ve outlined lots of problems; so what are the solutions? How can B2B be done effectively online? How can we utilize the consumer access panel resources we have? How can we reach those hard-to-reach groups?
Let’s start with the hardest part first – getting access to the elusive C-suite executives. We could follow the traditional route and use telephone recruitment. But that is difficult, expensive and doesn’t solve the underuse/maintenance problem. Possibly it is better to think about a shortcut: Maybe we need to access B2B respondents through third parties that hold relationships with them. These may include publishers (for example, trade press readership), loyalty programs (business travel, corporate credit card, etc.) or other sources. These sources may not be as pure from a sampling point of view but their use is normal in the B2B world. The preexisting relationship does the maintenance part; the market research is just an occasional extra.
This approach has the advantage in that it circumvents the fraudster while getting closer to our target audiences. To deal with the fraudster issue we may also need to independently verify the person. By calling their place of work, where switchboards are almost certain to be answered, it is relatively easy to establish that they are who they say they are. This has the secondary benefit of providing a cross-reference to the company that is so important for sampling.
Talk to the right people
Combinations of approaches have always been the practical realities in B2B research and with online this doesn’t change. It is usually more important to talk to the right people than it is to worry about consistency of sample sourcing. Hybrid approaches, using samples from consumer panels and specialist panels, perhaps even incorporating phone-to-Web (or Web-to-phone, for that matter) can all, and should all, be part of the B2B researcher’s armory.
Finally, we need to think about the value of the people we have and ensure we manage the relationship actively and sensibly. Senior IT professionals, for example, should be interviewed about topics that concern them or about hand-picked general topics that are perhaps more interesting than the average. Care and attention must be taken in every part of the market research process; these respondents are valuable and must feel valued for their opinions. This goes way beyond monetary rewards (such people don’t need the $0.50 we might offer) and touches on human motivation: the need to feel valued, to be part of a community, to know that they are good at what they do and that they freely chose to do it.
By paying attention to the details we open up the possibility for B2B researchers to enjoy what their consumer colleagues have had for the past 10 years: the best and fastest data collection methodology in our toolbox.