In my previous post “10 basics things to know about what is NPS”, #2 was about “How is NPS calculated.

NPS is calculated asking the customer the “ultimate question”:  “Based on your experience with X, how likely are you to recommend Y to Z?” What I didn’t elaborate on yet, it’s that there are dozens of options about how to express X, Y and Z and each of them will:

  • Impact your NPS score
  • Give you a different information about customer perception

This is fine as long as you:

  • Remember that what matters the most it’s not your score, but your score evolution once you have identified your baseline
  • Are aware that the answer you get is based on the question you ask, so it needs to be interpreted in the right context
  • Make sure you identify first what is that you want to measure and then decide how to formulate the question
  • Stick with it: once you have decided the question that works for you, don’t keep changing it because this will make any comparison to past trend useless and any understanding of customer perception tough
  • Know that the ultimate biggest value of the NPS survey is in the free text comments: the pure Voice of Customers


So what are the key threethings you need to know to word your NPS question effectively? Let’s go back to X, Y, and Z and look at each one.


X. “Based on your experience with X”

This is about which experience do you want to measure. It distinguishes two categories of NPS surveys:

a.      Transactional this is usually a one-off interaction with a channel or touchpoint through which the company interacts with customers. It’s used in surveys asked shortly after the transaction or interaction. Most common X in transactional surveys are:

  • Recent sales visit from your account manager (i.e. In B2B environment)
  • Recent order with E-shop (i.e. online shop)
  • Recent visit at Shop (i.e. a retail store)
  • Customer Service
  • Support (technical, software, etc.)

b.      Relational this usually measures a prolonged experience with a company product. It’s usually asked once or twice a year, in some cases more often (depending on  the intensity of interaction and value of the product for customer business), but usually not more than once a quarter. Most common X in relational surveys are:

  • Product XX
  • Company XX


Y. “How likely are you to recommend Y”

This is probably the most critical part of the question as it determines what is that you are ultimately measuring and what is that you ultimately want your customer to recommend: a given touchpoint, a given product or a given brand.

Let me start by saying that customers don’t think in terms of touchpoints.

They think in terms of experiences and fulfillment of their needs.

Did product A/company B or experience C meet and exceeded their needs so much that they would proactively go around to recommend it?

And usually, the last experience is the most impactful. This is the reason why you want to measure NPS also after critical interactions on given touchpoints. But this doesn’t mean that you need to measure “likelihood to recommend touchpoint D” .

We can differentiate three types of questions:

  • NPS brand question: “How likely are you to recommend <<brand name (company B)>>” which works well on certain relational surveys and on transactional touchpoints like CS/SS or Sales
  • NPS product/experience: “How likely are you to recommend <<product A>> or <<experience C>>” which works well for products, e-shops & what I call “value-added-touchpoints.”
  • NPS hybrid: “How likely are you to recommend <<us>>.” This usually works well for start-ups or businesses with little differentiation.

When to use which of these questions? Here some guidelines based on my experience.

On E-shops and Product NPS.

Here it makes sense to be specific about asking “How likely are you to recommend E-shop Acme/ Product X” regardless of touchpoints.

I have seen cases of NPS brand question on products with NPS – 40%, where even the customers giving a 9 or 10, mentioned in their comments all sort of things that they were not happy with. These customers are “brand promoters” (so they still trust the brand regardless of bad experience with one specific product) but they are not “product promoters”, While they might stay loyal to the brand, they will not recommend the product and might still refrain from a subscription product offered by the brand. In these situations, very typical for a brand with dozens or hundreds of products, you want to measure “product NPS” and not “brand NPS”.

On touchpoints

Here I tend to differentiate:

  • added value touch points” –> I would consider in this category: Training, a special self-support site, home repair visits, … and anything that generates value to customers and potential revenue and/or cost savings to the company
  • non-added value touch points” –> I would consider in this category: “standard” customer service (CS) or technical support (TS). And with “standard” I mean anything that is there to solve a customer issue reactively without special added value to the customer. Though I wish that more and more company would provide more than “excellent customer service” as opposed to “standard.”

For the first type, a TP NPS question can make sense (though is not necessary), while for the second one I’m intended to completely discourage a “TP question.” Main reasons being that:

  • From a customer point of view, if a TP is not adding value to me, I’d rather not call them unless needed (“best service is no need for service” concept). If I’m forced to call it because of some issue (like often the case with TS and CS), so don’t want to recommend it to friends & colleagues. It is different in case your CS is offering extra value added to your customers. During my holiday last August, for example, I happened to forget my beloved Nikon D5100 (and two expensive lenses) at a place booked via Airbnb, and I realised after we had already driven 300 Kms further. Calling Airbnb CS, they arranged the pickup and delivery to my next location. In cases like this, consumers are intended not only to recommend a brand because of that experience but the customer service itself. In fact, it was a friend of mine that reacted to my Facebook announcement that I forgot my camera, that suggested I should call Airbnb CS and, as result of their help, I did make a Facebook post highly recommending  Airbnb CS.
  • From a business perspective, CS/TS are usually non-revenue generation TP, so you don’t want your customers to recommend them unless you are Zappos. It is different when that TP (i.e. a home repair visit) provides a great benefit to the customer and a great saving opportunity to you.

As we learned the hard way during my time at Philips, changes from brand to a product or a hybrid question are very substantial changes and can cause quite strong variations to the score which don’t reflect a change in customer perception.

Of course, as you start the NPS journey, it will be quite likely that at times you will take a road and then realize that is not the best for you. This can be the case with this part of the question, so it will most likely happen that you need to change it at least once during your NPS journey.

My advise is: learn from the first 6 months/1 year of the journey, stop to think and review what works and what doesn’t and plan well for the changes that need to take place. Then do them at the start of the following calendar year (or if too far, at the beginning of the following quarter) and stick with them as long as possible.

Z. Who to recommend to.

In 95% of the surveys I have seen, Z is “to friends or colleagues,” so this is considered a no-brainer.

In my experience, I saw that this is often not the case and that this part of the question is underestimated.

The key point to think about here is “who can your customer recommend your product to and who do you want them to recommend it too”?

When I was at Philips, we used to measure the NPS of our Employee E-shop. Keeping the Z as “friends of colleagues” generated lots of confusion because being the shop only accessible to employees, it made no sense to recommend it to friends. So we used Z as “colleagues

In my NPS surveys towards my clients I usually choose as Z “within your business network”.

So here my key recommendation is: pick what makes the most sense for what you are trying to measure and to improve and stick with it.


What is your experience with X, Y and Z? Any other consideration you would add to the above?

Let us know in the comments below.


NB> Our friends at Satmetrix want us to remind you that Net Promoter, NPS, and Net Promoter Score are trademarks of Satmetrix Systems, Inc., Bain & Company, and Fred Reichheld.

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.


Setting up an NPS program?

Speed up the adoption of NPS in your organisation by taking the NPS Onboarding Masterclass, a two-day full immersion training that combines high quality content, peer exchange and blended learning. This in-company training is designed by Rosaria Cirillo Certified Customer Experience Professional (CCXP) & Authorised Resource Trainer (ART) by CXPA, and will provide your organisation with in depth learning, exercises and case studies related to the seven critical elements pivotal to the success of this program. Interested?

Pin It on Pinterest

Share This