The Basement

I'm late to the party but I wanted to make some comments on the excellent research paper recently published by Northeastern University's resident WOM(M) expert, Dr. Walter Carl, entitled To Tell Or Not To Tell? -- Research Report on Disclosure & Organized Word-of-Mouth Marketing Programs.

The big point made in the paper, of course, is that there is a strong business case to encourage disclosure in organized (stimulated) word-of-mouth marketing programs. This has been reported and discussed everywhere, including being presented at WOMMA WOMBAT. I encourage everyone who's interested in stimulated WOMM to read the paper and not just the blogosphere coverage, but if you really want to get the most out of it, read Dave Balter & John Butman's book Grapevine first. The reason is because Dr. Carl used the BzzAgent folks for the research project and I think that an understanding of the BzzAgent model lends tremendous context to Dr. Carl's work.

[DISCLOSURE: I'm a BzzAgent but I'm a lousy one as I haven't actually filed any BzzReports. Sorry Dave & Jono.]

One point that hasn't been heavily discussed (although is clearly evident in the report) is that the methodology used (due to BzzAgent's business practice) under sampled WOM episodes in which the "agent" and "conversational partner" (CP) had a relationship characterized as being either "acquaintances" (5.2%) or "strangers" (4.3%).

Not surprisingly, CPs categorized as "acquaintances" or "strangers" were also the least likely to be aware of the agent's affiliation with the WOMM organization (only 36% and 22% awareness respectively). Before anyone freaks out, BzzAgent asks their agent to disclose their affiliation during any WOM episode that is part of an active campaign. That said, it's hard to share information with a stranger and then tell them that you're affiliated with a WOMM firm. Regardless of how sincere and honest you are about the recommendation, without the trust of a more familiar relationship the disclosure makes you look like a shill. Subsequently, the less familiar the relationship, the less likely the CP is aware of the affiliation. This makes sense.

Of the WOM episodes Dr. Carl examined, 44% were involved a CP that was unaware of the agents affiliation. But since the likelihood of awareness of the affiliation is increased with familiarity of the relationship, WOM episodes where the CP was unaware of the affiliation most likely occur between acquaintances or strangers. Since, "acquaintance" or "stranger" relationship naturally convey a lower level of trust than more familiar relationships, it is unclear (at least to me) if the negative response to discovering affiliation later is due to the lack of disclosure alone or also dependent on the the unfamiliar relationship.

In other words, if my best friend recommends a product to me, I trust the recommendation. Disclosure of the affiliation (when or if) is irrelevant because of the high level of trust. (Although because of the relationship I'm also more likely to already know of the affiliation, but I still assert that it doesn't matter.)

If on the other hand, I recommend something to a stranger, the message is not trusted - regardless of disclosure. If I do disclose (before, during or after the episode) the stranger's trust is further decreased. From the perspective of the WOMM firm, an agent that shares something with a strangers is not particularly valuable because of the low level of trust. Furthermore, it presents a higher risk of negative backlash when the CP discovers the affiliation. So why would a WOMM firm want their agents to talk to strangers?

So while Dr. Carl's recommendation is to encourage disclosure, it seems to me that equally good advice is for agents not to talk with strangers.

In BzzAgent's case this happens naturally. BzzAgent doesn't prompt agents to talk with strangers. Instead, BzzAgent realizes that the value of WOM lies in the trust of personal relationships. As such, BzzAgent expects and "counts" WOM episodes with the most familiar relationships - romantic partners, close friends, family, etc. The higher the trust, the higher the pass-along. In fact, I suspect that the overall CP negative feeling towards agent affiliation could actually be reduced by asking agents to avoid WOM episodes for campaigns with CPs that are considered "acquaintances" or "strangers".