Video streaming opens up a new realm of possibilities for Internet research and this article addresses some of the project management and security issues which face market researchers wishing to implement this technology.
Peter Winters 
21 Nov, 2006
Using video within Internet research surveys; project management and security considerations
New possibilities
Imagine that you had a new concept that you needed some market research feedback on, but you were concerned that:
- you could not communicate the excitement and emotional involvement of your product story to your target respondents, and
- that your target respondents would not be able to communicate their real thoughts and feelings about the product back to you
And communicating the emotional content of your concept or brand is important! In recent years, there has been a growing understanding of the importance of emotion in helping us understand the world and make decisions. As Greenspan and Shanker contend “to create a meaningful symbol, the image must be invested with emotion”1. Following this line of reasoning, in pharmaceuticals, a drug becomes meaningful if it can convey emotion to doctors – such as caring for patients or generating a sense of purpose in what they do.
Pharmaceutical brand management should be about developing drugs which includes such emotional meaning. This suggests that quantitative research should be better at measuring the emotional meaning of pharmaceutical branding initiatives. For some research projects, we should move beyond just a rational description of “Product X”. We should also allow respondents better ways of communicating back their reactions to us.
In this respect, Internet research is emerging as a powerful tool for creating situations which respondents can experience and respond to. This methodology allows researchers to communicate a “stimulus story” for respondents in a highly-controlled, value-free environment. The availability of reliable video technology, embedded into Internet research surveys, has further increased the power of researchers to create an impactful “stimulus story” to be tested in research. This “stimulus story” could be advertising concepts, marketing messages, new pack designs or some other new concept which can be created in an electronically, virtual form.
Also, with good questionnaire design, respondents can be liberated to express how they really think and feel about what has been presented to them, in a way which is really not possible in traditional, interviewer-led research. Indeed, it could open-up a way for respondents to communicate some of the reactions to a “stimulus story” that they did not even know they had - their unconscious emotions.
An important skill for researchers is to design questions which allow respondents to articulate these responses. These might include adjective batteries, archetypes which act as metaphors for a brand, pictures of people, brand partners and the sensitive design of rating scales. Several years ago I was persuaded by G. Albaum2 that rating scales which combine direction (e.g. positive or negative predisposition towards something) and strength are more sensitive if asked in a two-stage format.
It is different with interviewer-led methodologies; respondents cannot help but be influenced in what they say when speaking to someone else. A useful analogy is with ethnographic video diaries that have gained prominence in qualitative pharmaceutical research recently. As examples, there were two papers at the EphMRA 2006 conference about these techniques (in the Technology Section) and Brand Health International also has experience in this area. One thing that has struck me is how clearly people are able to express their experiences and pain to an anonymous video camera in a way which they surely do not do in their consultations with doctors.
Project management of Video Streaming in Internet Research
It takes time to ensure that an Internet research survey, using video streaming technology, works perfectly! In my experience, making sure that everything is working correctly seems more similar to the process of software testing than checking a market research questionnaire.
To begin, you need to develop the video(s) – and ensure that it/they communicate exactly the “stimulus story” that is required. All the colours, text, sizes, any sounds, material information, mood and so on, have to be designed in a way that is correct for the concept.
Then the file needs to be adapted in size to suit video streaming. The size of the initial source file is not really an issue, but the file to be shown to respondents may well need to be reduced in size to make it manageable for the survey. Currently, the survey video size might typically be around 5 Mbytes, though this will likely vary according to how long it is. There is a trade-off here between quality of video and download time, and testing will be needed to reach the best solution. If a security system is enabled which does not allow cache storage on respondents’ computers, then download times could be longer.
The Internet survey can be set-up to provide different versions of the videos according to whether respondents have a “slow”, “medium” or “high” speed connection. Each of these versions needs to be checked! For example, it could be the case that the video would be correct at high and low speeds, but have a mistake at the medium speed.
During questionnaire completion, the videos will usually load over a few seconds. Whilst this is not a long time, it is always a good idea to let the respondents know what is going on. A message that reads “Please be patient whilst the video loads” at this time can help ensure your respondents keep to the task.
The survey design software should allow great control over the respondent experience – and this all takes time to design. For example, you may wish to disable the “Proceed” button whilst a video is playing to ensure that respondents do not miss anything. If you are showing multiple test concepts, you should anticipate an order effect and manage that in your experimental design.
Technically, using video streaming on the Internet is a challenging area and you will likely need specialist software, such as provided by Clipstream, to reach a high proportion of respondents. Even with such software it is a good idea to anticipate that some respondents will not be able to view the videos - and so procedures should be put in place to identify these people and edit them out of the analysed sample. On a recent study of over 2,000 doctors and patients in Brazil, Germany, Spain, UK & US that we conducted, we had an overall failure rate of around 2%. These respondents were identified and excluded from the analysed sample. We were glad to see that those that failed were spread quite evenly throughout the total sample.
Security
Security becomes more of a concern as videos enable more sophisticated content to be tested via Internet research. There are established procedures to safeguard security such as running surveys on an “invitation-only” basis amongst bone fide doctors, and requiring respondents to respect the confidentiality of the questionnaire and stimulus material.
In addition, there are now a number of software features that can enhance the security of video streamed content. Clipstream can set the video to remain on their servers; and even if the video is cached on the respondents’ computers, it can still be secured to Clipstream’s server such that the video is not playable by any other means than through the survey link. The screen-capture function can be disabled in Java for Internet Explorer (respondents can be told to complete the questionnaire using this browser). It is also possible to disable the print function in CSS. New features are being developed, such as a watermarking capability which gives the opportunity to thwart photos of the screen since the watermark can be set to uniquely identify the viewer. If security is required of still images, such as with JPG files, these can be transformed into video format to take advantage of these security features.
These software procedures do not make content tested via Internet research 100% secure – for example, respondents could take a photo of the screen with a digital camera. However, they do greatly enhance the confidentiality of survey test materials and I believe that they will quickly become the default standard for Internet research product testing.
You can view an ethnographic film put together by Brand Health International which includes a number of security features, including the disabling of the screen-capture function. Please contact me at peterwinters@brand-health.com if you have any comments on it, particularly if you have any trouble viewing it.
1 The First Idea: How Symbols, Language and Intelligence evolved from our Primate Ancestors to Modern Humans, (2004), p.25 Stanley I. Greenspan & Stuart G. Shanker
2 Albaum, G. 1997. The Likert scale revisited: an alternate version. Journal of the Market Research Society, 39(2), 331-348.
Your response:
Login to respond to this article: