Internet research - recruitment

Internet research always seems so easy and cost effective. There are no interviewers to pay, no printing costs and everything can be controlled from your desk.

However, getting hold of people whilst managing respondent's rights to privacy and without spamming is a very important consideration. And what of the quality of the survey when the respondent arrives at your site? Would you like your top customers blasted with an email inviting them to a tatty Frontpage built site?

SurveyGarden is our dedicated on-line research facility for hosting our own and third party questionnaires.


Recruitment seems simple - get a list of emails an mail everyone. Unfortunately, spamming as this is known is not going to be well received by those receiving the email. In addition, if you are looking for a particular selection of people, it can be difficult to know if details on the email list truly represent the people you are interested in. Secondly, you need to consider how people will react coming to your survey - what impression does it leave of your company? On the Internet people drop out quickly of badly designed surveys. Care must be taken over survey design as well as contact.

Like other forms of research, Internet research relies on a sampling method, but unlike other forms of sampling, Internet response is more difficult to control and is sometimes referred disparagingly as convenience sampling since it depends on who is around and willing to take part at the time.

The main approaches for Internet recruitment are therefore pop-up surveys, which are very good for direct research about the web site being visited as people are recruited on the spot. But pop-up blockers that come with many browsers mean that many potential respondents will not see the invitation. For this reason Javascript and other types of survey pop-up have been devised, but this makes pop-ups more difficult to use across browsers. Designing cross-browser surveys should be part of your survey design requirements.

Increasingly, the difficulty with pop-ups means that individuals are being contacted by emails and asked take parts in web-based surveys where it is easier to control and verify the content of the responses given. Under market research codes of conduct, the emails used have to have been opt-in lists to avoid accusations of spamming. Most sample is provided via online panels such as GMI or Toluna. These companies hold databanks of several million people who have opted in to take part in research. There used to be questions about representativeness of these panels, but these have issues have become hidden somewhat through use.

For specialist markets, an alternative solution is to use an off-line list, for instance your own database, or to build a list of willing participants on the Web (for instance via a Pop-up survey), or through telephone contact. Using your own lists is often an excellent solution for business-to-business research where traditional forms of contact are largely restricted to telephone surveys. But if you do have to resort to a telephone call to invite someone to take part, often it would be more effective just to complete the interview on the phone.

One word of warning on using your own lists, as it can be extremely difficult to keep the list up-to-date. In business markets for instance contact turnover can be as much as 25% per year. This means that unless you are regularly managing and using the lists you have they can quickly run out of date. And if you are buying lists from third parties, you do have to be careful with the age of the data.

An alternative which can be very effective for special interest or community groups is a snow-ball approach. In a snowball a small number of active people are invited and asked to pass details of the survey on to their contacts. We have used this very effectively in the sports arena, using a network of participants and volunteers to pass the survey details around to their friends and other participants, eventually reaching a sample size of around 800.

An additional factor which needs to be considered is that you may have to incentivise people to participate. Free prize draws are very common (even to the point of being off-putting now), and panels often use a points system allocating points to those that take part in the research which can be exchanged for products later.

Naturally, once recruited, you need to collect the information. Survey design should be an important consideration as there will be a drop out rate from those starting the survey as they go through the questionnaire. In addition the quality of the design itself will greatly reflect on the commissioning business..


Several companies look at community groups or social media as potential sample source, looking to place questions within discussion groups or to use the community groups as a place to recruit potential respondents. Having had the advantage of a dedicated users newsgroup when working as a research buyer there are some pluses to this approach.

However, whilst there are plenty of places on-line which will be discussing your company and topics of interest to your business, the people posting on these areas are typically very unrepresentative of customers as a whole. They will more typically be expert users and often have very specific issues and requirements that are not suitable for a wider audience.

Consequently, an alternative is to create your own community group for research which can then be tested and seeded providing feedback in real time when required and enabling the community to develop it's own ideas, materials and questionnaires to help inform the business.


For help and advice on carrying out on-line research projects contact info@dobney.com or take a look at SurveyGarden.