- 7/23/2013 5:16:05 PM
No, I don't think they do, however, they are probably unaware of how deep one can actually go in deriving further insight. Many organizations simply just tract their promoter scores over time and they feel that it isn't actionable. Needless to say, I've seen it can be very actionable and we've shown several organizations how it is possible.
- 7/23/2013 4:51:34 PM
@rscollica, I agree. But, I have to ask, do you think marketers are doing enough today to dig deeper than the NPS? Do you enough of them at this point supplementing the NPS metric with deeper analytics?
- 7/23/2013 4:38:24 PM
Yes, indeed! Any metric that attempts to classify human understanding will arguably be incorrect at some point. However, from what I've seen, even when customers tell you their willingness to recommend rating and then have a verbatim text, the general trends in sentiment from the text and the ratings, but certainly not exact. So, point being, whatever metric is being used, understand the limitations, but use the analytics to the best and use this along with domain expertise to make the best possible decisions for the organization.
- 7/23/2013 4:31:35 PM
@rscollica -- thanks for joining the conversation. I think you're suggesting that NPS serves as a starting point but really should be supplemented by insight gleaned through text analytics. Did I get that right?
- 7/23/2013 3:04:37 PM
So as I read through these comments I'm struck by comments about if these scores are relevant or not. First, I must say that if any vendor says that they can predict actual customer sentiment more than 80% correct, then they are selling you snake oil with the smoke and mirrors included! Second, it is possible, however, to estimate the association of what they are writing about and their sentiment indicated in a promoter score. Since people are the generators of the survey results, and also the social comments in call notes, verbatim comments/chats, and the like; any metric to measure such will be as imperfect as the people who generated the information themselves. All that set aside, one can determine how many actual detractors there are and what they are saying can be quantified accurately enough in order to predict the scores - even if the scores aren't totally accurate. The scores themselves will not be totally accurate due to what others have just stated in this blog thread. The confidence in generating such estimates can in fact also be estimated (i.e. how much error or misrepresentation exists in the data?) and this too can be included in estimating how much impact an organization may have or not have by reducing the detractor topics/levels estimated in promoter score surveys. If you'd like to know more on a methodology on how this can take place: please visit: http://blogs.sas.com/content/text-mining/ There is a lot of resources and white papers which explain the methodology much further.
- by kq4ym, Data Doctor
- 7/22/2013 9:40:59 AM
Social media is winning out. Even discounting emotional responses to companies and products, it's fairly easy to guess which are valid comments. And social media is certainly more entertaining as well. And maybe that's what we're really looking for anyway.
- by Zimana, Blogger
- 7/22/2013 3:56:08 AM
The results of ForeSee's report reveals the importance of where our attention is shifting to. The idea of asking a dual level question prevents a bias in context, that positive commentary is the only value. Being alert to negative commentary is just as valueable, so the scoring and attention span should reflect a dual context.