Net Promoter: Buckshot Marketing?

I’ve been meaning to write about net promoter score for a while but it’s one of those big topics that probably needs to be tackled in sections. As I’ve mentioned before, simplicity can often be the key to widespread adoption of a new metric and you can certainly see why NPS appeals to marketers on that level. The only problem is that I really don’t think that every WOM program boils down to “Would you recommend our product/service to a friend?” since WOM interactions are slightly more complex than simple recommendations.
First of all, from a purely statistical standpoint, I highly recommend reading the FREEDyourMind piece on NPS. Among the very valid points that Larry Freed makes are that the margin of error is far too high (he claims an NPS of 24 could be anywhere between 14-34), the scale isn’t clearly defined (he claims 6′s are rarely actual detractors) and that there is no connection between NPS and growth. For even more detail along these lines, I also recommend checking out Bob Thompson’s very detailed piece at Customer Think.
While I am tempted to give Bob and Larry the benefit of the doubt on the statistical relevance, the point I’m most interested in exploring is the connection between NPS and growth. There are certain categories, like automotive and travel/hospitality, that are driven by moving consumers from “passive” to “promoter.” I believe that it’s less a measure of success to get that snapshot of NPS in a random sampling of the target demographic than it is to measure a specific set over time.
It can probably be done in similar manner to brand awareness. Since we basically share a floor with Dynamic Logic, who I consider to be the hands-down best at measuring brand awareness online, I often debate the methodology for this kind of measurement with my friends over there. From a marketing perspective, I would only be interested in seeing an increase in NPS in the same set of surveyed individuals over the period of a marketing interaction. For banner impressions, I believe that the Dynamic Logic method of measuring people who have seen an ad against those who have not makes plenty of sense but I don’t think the same is true for someone who has submitted a video for Doritos Super Bowl promotion and waited around for a month to see who won or someone who has visited H&R Block’s tax advice outpost in Second Life and been incentivized to walk into a retail outlet.
I also think that the question doesn’t speak to the true value of WOM. I’m less interested in if the surveyed body would recommend the product than how they would recommend it. Sure it’s great if a large percentage of people will recommend that you buy a SlingBox but if these people can’t explain to you the key benefits of placeshifting then they’re not making valuable recommendations for Sling Media. I’d much rather ask a follow-up question that offers the product’s key messages with a couple of common “dummy” messages to see if the promoters will actually drive growth of the product or service.
Not only would this help connect NPS with growth but it would also help measure message adoption over the course of an engagement, which possibly also helps to answer the “how do we measure engagement?” question.
Unfortunately, I’m going to miss Walter Carl’s session on this at the WOMMA Word of Mouth Research Symposium tomorrow, which I’m sure will be very interesting (and hopefully well documented on someone’s blog). I’m offering a free drink to anyone who sends me a rundown, and I hear free drinks are very hard to come by in Las Vegas.

