NPS Blog

Webinar Q&A: What does a Net Promoter survey look like?

Fred and Rob discussed the upcoming book in a July 20 webinar. In this and subsequent posts, they answer additional audience questions from the webinar question-and-answer session.

Q: What does a Net Promoter survey look like, what types of questions are used and how are they measured?

A basic top down NPS survey typically includes:
1. Question(s) to qualify the respondent as a customer and any quotas
2. Likelihood-to-recommend question
3. Open-ended request for further explanation
4. Additional question(s) on other aspects of the product, service or brand
5. Optionally: information about the customer’s purchasing, usage and referral behavior that enable analysis of differences in financial outcomes
6. Any final questions that can help with segmentation (e.g., demo, buying behavior, etc.)

The specifics should be adapted to a particular company, sector. These questionnaires may range from 10-30 questions, depending on the situation.

Bottom-up Net Promoter surveys are much, much shorter. In general, they include:
1. Likelihood-to-recommend question
2. Open-ended request for further explanation
3. Permission to follow-up

7 Responses to Webinar Q&A: What does a Net Promoter survey look like?

  1. Hi. we are currently implementing CFL in our organization and facing the challenge of 0-10 scale understanding by our customers in bottom up NPS survey. Our market (south east asia) has literacy as a hurdle and to most customers giving feedback on a scale is a big comprehension issue. The comprehension increases when we change the scale fro 0-10 to (Yes will recommend, may be recommend, no will not recommend) – which makes the score more acceptable to agents, who in 0-10 question the scale a lot based on customer’s lack of comprehenshion.
    As we implement CFL, I strongly feel about the value addition in using feedback to generate actions, but am not sure on scale and do we lose any value if we change in bottom up from 0-10 to Yes recommend, may be, No recommend ?

  2. Rehan:

    You clearly face in interesting challenge. If your customers can’t read for themselves, you may find several challenges in the way of ANY type of feedback.

    The guiding principles around your choice of scale for your survey should include:
    1) The extent to which the scale enhances your customers’ ability to provide high quality, honest feedback to individuals and the organization

    2) The extent to which the scale enhances your organization’s ability to distinguish high value from low value, and to use that to prioritize actions and investments to improve the customer experience. Collapsing the 11-point scale down to 5 or (as you may be suggesting, 3 points) generally obscures the economic relationships.

    We have found that the 0-10 scale is able to handle the vast majority of situations in line with points 1 and 2. In our experience, we have found it rare that a simpler scale like the one you propose really adds value. The primary problem stems from those customers who simply refuse to use the top of any scale, even when they are thrilled with a company or its products. This results in a “sloppy” identification of promoters versus passives. It doesn’t resonate for the organization, and it doesn’t match up with the economic outcomes.

    Therefore, tread with caution before collapsing the scale down to just three elements.

  3. How can i as a manager of team of field engineers advise my team to get better NPS scores?
    What are the basic, winning ideas and behaviors

  4. Philip:

    A couple of simple ways to get started:

    First, get your team together to share their own best thoughts about what sorts of behaviors earn the enthusiastic loyalty of customers, and what sorts of behaviors don’t.

    Second, open up a dialogue with a few of the most enthisiastic promoters among your team’s customers. What would they say to a friend about why they like doing business with you? What did your field engineers do to contribute to that? What would these customers say distinguishes the service they got from your engineers and made it so much better than they expected? Keep this simple: just call up or visit with a few of these customers and ask a few simple questions.

    In the early stages, it’s typical to get significant value from simply identifying some common “good” behaviors your team can do more consistently, and “bad” behaviors they can do less often. These can lead to more consistently earning the trust and loyalty of customers. Your scores will follow.

    One warning: don’t focus the team on improving the scores. Instead, focus them on EARNING the loyalty of their customers. Use the scores only as a means of learning whether it’s working. This might seem a subtle difference, but it’s important. Earning loyalty takes time and patience. And it pays off big.

  5. Hi: Usually for the bottom-up NPS survey based on service delivered from call center, we asked the question like this- “based on you last interaction how likely is it that you would recommend (brand name) to your friends or colleague?”. Is it okay if we ask the question like this- “Based on your last interaction how likely is it that you would recommend our/ brand name customer service to your friends and colleagues?”. So, instead of asking whether the customer would recommend the brand if we ask whether he/she would recommend that particular service/product/any other area of experience what they have experienced recently. Is it okay? Can you give any reference?

  6. Shams:

    We have seen several approaches to touchpoint-triggered bottom-up Net Promoter feedback. The classic approach, developed over 10 years ago, leaves the question brief and direct, without modifications: “How likely would you be to recommend [brand or company] to a [friend, colleague or family member]?”

    In most businesses, we have found that the addition of modifiers such as, “Based on your most recent interaction…” or “Thinking only of your most recent interaction…” are unnecessary. Customers are naturally influenced heavily by their most recent interaction. So the additional words in the question merely make the question more complex and harder to read.

    Nevertheless, there are some businesses in which likelihood to recommend a brand or product (or store location) can diverge in the short term from the impact the most recent interaction had on that likelihood to recommend. This first came to our attention in hotels, where a frequent guest might be very likely to recommend a particular property in a hotel chain based on a long history of positive experiences, the convenience of the property, the room quality, etc. But the most recent stay may have been miserable for some reason. Overall likelihood to recommend might mask an emerging issue that more specific feedback might help the employees or management of the hotel address.

    For this reason, some of our clients, especially those in businesses with high frequency of interaction but long-term relationships, are now asking both overall likelihood to recommend the brand, as well as “To what extent did this most recent interaction increase or decrease your likelihood to recommend?” This is generally scored on a +5 to -5 or a +3 to -3 scale, and is followed by an open-ended “why” question.

    Separating the overall from the impact of the most recent interaction seems to provide significantly more information to the organization, and it appears to be more consistent with the way customers think about their relationship with the company.

    If you try it, please let us know how it works for you.

  7. Do you think it is important that the customers know that on a scale of 0-10, the score between 0-6 means least likely. I am asking this because if we ask the question as “Based on your experience…..on a scale of 0 to 10, with 10 being most likely and 0 being least likely” and if a customer either doesnt have much experience or isnt on either end of the spectrum, he might choose to give a 5 as an average/middle score. But this score may make him a detractor whereas he really might not be.