Ask the right question

Like many systems, Net Promoter asks testers to rate aspects of a product on a scale from 1 (not likely) to 10 (very likely), as do many systems. Net Promoter differs in two important ways.

First, it asks a specific question: "How likely is it that you would recommend our product to a friend?"

Second, Net Promoter doesn't treat 1–10 as a straight scale. Instead it regards 0–6 as unhappy users, 7–8 as satisfied, and 9–10 as loyal enthusiasts. These three groups are surprisingly accurate. We recently studied surveys we've been sending for client projects, webinars, newsletters and more for over a year.

For example, we use SurveyMonkey to collect information. Some paraphrased responses (emphasis ours) to the "likely to recommend" question include:

• 7 — "[The presentation] covered some things that weren't in the documentation or went a bit outside the box, which was cool."
• 8 — "Awesome. There was a good variety of information on the subject. The material was clearly explained and the presentation was very professional."
• 7 — "The moderator did a pretty good job. The presenters were all experienced subject matter experts."
• 10 — "The presentation was spot on. I learned so much!"
• 5 — "It felt really rushed to get through that part, and then spend an hour doing essentially a 101 class. The first hour was OK, but the last hour was a waste."
• 7 — "The first session was very good, second session went too quickly. It could have used more explanation behind it."
• 6 — "I would have liked a quick overview of the various types of products offered."
• 9 — "You gave me a lot of confidence to generate prototypes for my application."
• 4 — "The audio kept dropping out. I couldn't hear."

2. Score your results

Let's take a closer look:

• 10 — "I learned so much!"
• 9 — "…a lot of confidence…"
• 8 — "Awesome." "…clearly explained… "…very professional."
• 7 — "…which was cool."
• 7 — "…pretty good job."
• 7 — "…was very good…" "It could have used…"
• 6 — "I would have liked…"
• 5 — "…the last hour was a waste."
• 4 — "I couldn't hear."

The top scores — 8, 9 and 10 — are promoters, or people most likely to spread the word. Scores 6 and 7 were lukewarm. Scores 5 and below were not satisfied. Specific complaints are important, but when looking at the big picture we award one point for each positive review, and subtract one point for each negative review. Neutral reviews neither add nor subtract.

• :)  "I learned so much!"
• :)  "…a lot of confidence…"
• :)  "Awesome." "…clearly explained… "…very professional."
• :|  "…which was cool."
• :|  "…pretty good job."
• :|  "…was very good…" "It could have used…"
• :|  "I would have liked…"
• :(  "…the last hour was a waste."
• :(  "I couldn't hear."

3. Get the big picture

The math is simple. In all, 30% of the responders above were promoters, indicated by green smilies above, and 20% were detractors. giving the class a net score of +10. Not bad, but not really good. Great would be if we could get that +10 higher over time by encouraging practices our promoters applaud, and shoring up what made our detractors grumble. Now that we know what to improve, we can be confident in our next iteration.

Automating the process

We use a bit of code* to help alleviate the manual labor in compiling dozens, if not hundreds, of results:

promoters_count = 0
detractors_count = 0
overall_count = 0

for each:
  if value > 8
    promoters_count += 1
  else if value < 7
    detractors_count += 1
  end if
end for each

overall_count += promoters_count
overall_count -= detractors_count


*This is the general logic, not a particular language. Check out our Ruby code in this handy gist (requires a Github account).

Have your say

A final thought: We’d love to hear what you think of these lessons in your inbox. Please take a minute to rate us — you guessed it — from 1–10. How likely are you to recommend these weekly lessons to someone else?

---