Nic Matham of KZero, a UK consulting company, published some stats from the L’Oreal campaign in Second Life (more background here). Now first let me say that I really *liked* this campaign. KZero and L’Oreal did the kind of project that I’ve wanted to do for years (Reuters, iVillage and Bantam Dell came closest in different ways). L’Oreal focused on virtual goods, which generate better word of mouth activity in SL, and they leveraged the success of existing SL entrepreneurs by partnering with them as a channel. In doing so, they scaled their campaign across the SL grid, rather than fighting against single-location scalability issues.
Nic writes that over 34,000 virtual goods (which were looks/skins for avatars) were picked up during the 3 months of the project. Given that the total active avatar population in Second Life is 544,290, their market penetration share was somewhere between 1.6% and 6.2% (depending on if each avatar took just one skin, or all 4 skins).
Now I’m going to step back for a second. As another data point for a “virtual goods” project in SL, we can look at Nissan’s effort in Second Life. The Nissan presence is still around but it has not been actively updated or maintained for some time. Since launch in Oct 2006, 89,000 virtual cars have been picked up by residents. The bulk happened during the initial Sentra and Altima kickoffs, but there has been a steady amount of transactions throughout. When I wrote about the project’s long tail last September, 68,000 cars had been picked up. That means that on average over 3,000 cars have been picked up each month since then, on a location not promoted or active. I’ve long been convinced those numbers could have been higher if we had been able to take more of a distributed approach across the grid, rather than an island-focused approach (I lost those arguments with the agency). L’Oreal’s campaign lasted 3 months. Nissan has kept their presence in SL up for about 18 months. The average transactions per month are different between the two, but not mind-blowingly so, and are arguably quite low. By the grace of its timing, Nissan had a lot more going for it than virtual goods — it got massive media and blog coverage.
I bring all this up again because I remain puzzled by the expected return to the brand holder beyond experimentation and learning, and I’m also asking whether SL is really the best place for this kind of learning compared to, say, Gaia. There are those who argue that SL is still a good place for marketing (as opposed to market research, simulation, training, collaboration and other corporate uses), and I am open to learning why, hopefully without getting my head seared off in a flame.
I do believe that L’Oreal was a well-designed project, and one that fit the SL community. I give both KZero and L’Oreal serious kudos.
I also believe in the power of virtual goods for branding. If you talk to Matt Bostwick of virtual MTV about the results they have seen from sponsor integrations in vMTV, and the impact they have measured on consumer purchasing decisions, your eyebrows will go up. The statistics are really good.
Mark Kingdon, Linden Lab’s new CEO, seems to agree in his comments regarding corporate marketing in SL: in essence he is saying “right location, wrong time.” I mostly agree — the exception being that 2006 and very early 2007 was actually the right time because of the PR bonanza which served to bring SL many new customers as well as deliver a return to the brands involved. People might criticize the format of the early SL experiments, but I do think those early adopter brands got a huge amount for their money.
PS. and in case someone is wondering why I seem to be mean about SL one day, then nice the next, then mean again, the answer is simple — I love the folks at Linden Lab and I love the dream of Second Life, but I am just trying to examine this space honestly and call things as I see them.
PPS. if you want to look at an amazing virtual goods model, look at Stardoll.