Avatar-Based Marketing Revisited: Market Truths methods

Three Part Interview with Market Truths’ Mary Ellen Gordon

In June 2006 Harvard Business Review Senior Editor Paul Hemp goosed the very curious business world with a think piece called “Avatar-Based Marketing.” His opening salvo: Companies spend large sums trying to segment, reach, and influence potential customers. They should think about targeting those customers’ online alter egos, as well.

Two years and many millions of virtual world user and VC dollars later, PGL revisits the subject with Mary Ellen Gordon Managing Director of the award-winning VW marketing research group Market Truths. Their Q1 (2007) survey on the VW Second Life made a big splash and help set precedent for researchers and marketers finding their way in this strange new land.

Quantitative and qualitative network analysis

Mary Ellen Gordon: We [MT] had come into this when Web analytics were developing and did not agree with how click throughs or even measuring time spent on a page were being valued. You can’t tell if someone is on a page for a long time because they are interest or because they can’t find the contact information they are looking for. I don’t mean to imply that traditional Web analytics have no value, but rather that I don’t think they’re sufficient on their own. They’re still useful for looking at trends, etc., but provide the most insight when combined with data collected directly from the user. So, to pick up on that example above, if you surveyed visitors to the page in question, you could find out whether those who stayed longest were more engaged or most confused or if time on the site had no correlation at all with engagement with the site, purchase intentions, etc.”

Also, what I am really thinking of is how Web surveys have evolved online. In the very early days, there were traditional research companies trying to do Web surveys without really understanding the technology (so not making use of the ability of the Web to generate dynamic content) and technology companies trying to do Web surveys without really understanding research (and so asking poorly worded questions, not providing any in depth analysis, etc.). Then a lot of DIY survey software came out either as off the shelf packages or as online applications, and that made the situation even worse because then everyone felt that they could do their own survey even if they didn’t really understand technology or research very well (enter surveys with spelling mistakes, messed up skip patterns, etc.). To facilitate that a bunch of online panel companies emerged (they send people through to do surveys in exchange for incentives). The upshot of all of that has been that people’s inboxes are flooded with poorly thought out surveys, and so response rates have gone down. The people on those online panels make up a huge proportion of online survey respondents even though they are only a small proportion of the population overall. There are obviously questions as to their representativeness of the population overall, and there is a lot of concern about cross-panel duplication (people being on multiple panels), professional respondents (people who take so many surveys they become atypical even if they were not in the first place), and people doing surveys just for the incentives without paying much attention to the actual questions or their responses (there are even automated programs for filling in surveys to facilitate this).

Our positioning has always been at the intersection of marketing, research, and technology and doing quality, in depth research. That made virtual worlds a logical target for us. By being first, we hoped we might be able to avoid the whole negative spiral that online surveys have experienced, as described above, as we could be part of shaping the environment and expectations about research. It was also a way for us to stand out more from the whole pack of companies offering online surveys.

Method and Process

PGL: What were the methods and procure of running this study? MT established a Second Life (SL) research panel comprised of SL residents.

G: The first stage it is self-selecting. We advertise in SL and associated media. You decide if you want to take part in panel. We want people who want to express themselves. We want the expression to be part of the motivation not just Linden dollars [panelists are compensated for their time in virtual world money]. They come to our office [in SL] and click on kiosk to be part of the panel. Just to expand on this, in addition to real life (RL) brands and customer reports geared toward RL businesses, we also do reports more aimed at SL only businesses and SL specific issues that we know SL residents often have strong opinions about.

PGL: What are the criteria for participation?

G: They need to be in SL for 30 days and have a verifiable account. The 30-day standard means that the participant is beyond the introductory stage of being in-world, and the verified account reduce chances of griefers [a player whose goal is only to harass others]. The participant clicks “yes” to be part of the panel, the fills out the preferences form, which is a multiple choice on Web questionnaire with a unique URL for each participant. By taking the questionnaire out of world we insure that the avatar and user are the same. The questionnaire looks like a Web survey. We ask the following:

RL age
Rl gender
Sl age––We capture this automatically from their universally unique identifier (UUID)
Sl gender
Frequency of us––The frequency question is about how often they want to be invited to participate in research.
What are the participant’s interests
What is participant’s RL country of residence
SL profession is not part of the profiling questionnaire, but it is something we have asked about as part of specific projects.
RL profession is not part of the profiling questionnaire, but it is something we have asked about as part of specific projects.

190 participants were part of the Q1 (2007) survey; 201 in Q3 (2007) survey.

The form of participation in the study was threefold:

1. Surveys
2. Focus groups in-world, text procedure 90% request
3. Interviews in-world, text procedure 90% request

Gordon says of her research experience in-world, “People are more open in SL focus groups because there is more anonymity.”