Welcome to the Listening Post!

The Ultimo Question: Net Scores

Tweet

 

A question I kept getting asked last month from clients, prospects and coworkers was about net scores: when should they be used and what net score calculation works best for NPS?

A net score is simply a calculation of the difference between the top box and the bottom box. So if 40% of answers were in the top box (e.g., answering “completely satisfied” on a satisfaction scale) and 15% were in the bottom box (“not at all satisfied” or “completely dissatisfied”), then the net score would be 25. Net scores can range from -100 (if all answers are in the bottom box) to 100 (if all answers are in the top box) but are often written as -100% to 100%.

This calculation was of course popularized by the Fred Reichheld’s Net Promoter Score. In actual use, though, you will see many different takes on the NPS calculation. For instance, here are some of the calculations I’ve seen:

• Top 2 – Bottom 7, on a 0-10 scale. This is the classic NPS calculation as introduced in The Ultimate Question.
• Top 2 – Bottom 6, on a 1-10 scale. This is a popular variation, used by many who dislike the 0-10 scale.
• Top 3 – Bottom 5, on a 0-10 scale. One solution to low NPS scores is to define the top box more broadly and the bottom box more narrowly, which I’ve seen firms do simply to inflate their scores.
• Top 3 – Bottom 3, on a 0-10 scale. Schneider, Berent, Thomas and Krosnick (2008) found that this calculation, of all the possible permutations of net-score calculations for an 11-point scale, had the strongest relationship between willingness to recommend and actual net sales.
• Top 1 – Bottom 3, on a 1-5 scale. This is the standard 5-point NPS calculation.

All of these possible permutations make it difficult to compare self-reported numbers for NPS, yet the research also indicates that one net-score calculation is not best for every company in every industry. Accordingly, Vovici 6.1, released last week, now supports generic net-score calculations, allowing the analyst to specify the cut-off points for both the top box and the bottom box. Any of the above calculations can be included in Vovici reports.

Some critics of NPS complain that the net score loses information and that net scores that are the same can have distributions with widely different shapes, yet that is true of any summary statistic. Another complaint is of the volatility of net scores, yet when seeking to create a voice of the customer program that shows improvement in results, the volatility of net scores can lead to more dramatic improvements in results than when an organization is trying to simply increase the percentage of customers who are completely satisfied.

The traditional debate of which summary statistic to use – top-box scores vs. means – now sometimes includes net scores as well. Which summary statistic to use is often a political decision rather than an analytical decision. Each type of summary statistic omits some key information about the underlying data, which is why some statistics work better for some organizations than others. Rather than simply do what everyone else is doing, take the time to explore your data to find which approach is best for your organization, and co-create your VOC program with a trusted partner.