Net Promoter System

September 1, 2019

NPS (27) – How not to deal with a Detractor (Nike and me) – 27th article in a series on the Net Promoter System®

Maurice FitzGerald / Blog, Net Promoter System, NPS / Customer Experience, marketing, Net Promoter System, Nike, NPS, strategy /

Welcome to the 27th article in my series on the Net Promoter Score and System. You can read all of the prior articles on our blog page here.

Nike and Lululemon are probably the two most respected and recommended sports clothing brands as I write this. What I want to relate is what goes wrong when a brand tries to support a non-core product. The product in question is the running application for iOS and Android called the Nike+ Run Club. Please put yourself back a little in time. I originally wrote this almost three years ago and have decided not to edit the present tense I used at the time.

From problem to FAQ to Facebook

I have been using the Nike+ running app for a couple of years on my iPhone and have logged over 1,000 kilometers (600 miles). The Run Club is the latest version, which I currently have on my iPhone 6. I always use kilometers as my units. On the second run with the new version, the displayed number was a distance in miles, but the units displayed were kilometers. My average pace showed as 8 minutes per kilometer, rather than my usual 5. I checked the distance I had run online, and confirmed the nature of the error. First stop, the Nike web page devoted to their applications. Nothing found in their Frequently Asked Questions. A note there directed me to their Facebook page as one support option, saying it works from 9 a.m. to 6 p.m. Pacific time. So, I went there to share my experience and ask for help.

They have apparently not heard of their own app

The Facebook experience was quite surprising. I used their search function to find out whether anyone else had had the same problem. Apparently not. I decided to explain my own. Describing it was not a problem. Using their pull-downs to categorize it was a problem. First, the app was not listed. Since an old version was there, I selected that (the Nike+ Running App), as shown in the diagram below.

Then I was asked to supply my OS version. As I write this, iOS 10 has been released. Apparently, Nike had not yet heard of iOS 9.

Naturally, the app versions available did not include mine.
– 24 hours later… nothing.
– 48 hours later… nothing.
– 72 hours later… nothing.
– 96 hours later… nothing.

So therefore…

So I deleted the Nike+ Run Club app and installed the Endomondo Sports Tracker app instead. Endomondo is owned by sports clothing manufacturer Under Armour. Your Detractors will leave you if they complain and you do nothing to respond. All my running clothing is from Nike. Will I start to look more favorably on Under Armour gear? Only time will tell.

Next time

The next article will be the first in a set about the robust operational and analytical infrastructure you need to have a successful Net Promoter System implementation. It is all about the team and the tools you need.

As is often the case, the above is a slightly-edited version of a chapter in one of our books; in this case Net Promoter – Implement the System All of our books are available in paperback and Kindle formats from Amazon stores worldwide, and from your better book retailers.

August 29, 2019

NPS (26) – Activating Promoters (so they really do recommend you) – 26th article in a series on the Net Promoter System®

Maurice FitzGerald / Blog, Net Promoter System, NPS / Customer Experience, marketing, Net Promoter System, NPS, strategy, Tesla /

Welcome to the 26th article in my series on the Net Promoter Score and System. This time the subject is money! We are talking about the Net Promoter System. The key word here is ‘Promoter’. Read on… You can read all of the prior articles on our blog page here.

So, you have identified a lot of Promoters, Passives and Detractors. As humans, we like the sense of accomplishment that comes with putting out fires. It is tempting to spend all our time on Detractors. After all, we understand their problems. We know what to do. As for Promoters, well, it is far less clear what we should do. Here are some examples and ideas.

Tesla’s Promoter program

Tesla is in the fortunate position that a lot of people love their Model S cars. They have taken the principle of activating Promoters to a new high. Existing customers are invited to enter the company’s referral program and receive a referral code. If one or more of their friends or colleagues use the referral code when buying their own Tesla, the Promoter receives ever-increasing rewards. In early 2017, the rewards were as follows:

Every referral means an additional ticket in a periodic prize draw for a Ludicrous P90D Model X.
Two referrals mean a Tesla sports bag.
Three brings a Tesla jacket.
New Tesla wheel rims come with five referrals.
Promoters get invited to the next major Tesla event after ten referrals.

The top Promoter in 2015 had 188 referrals and won the car. There is a client-side incentive to accept referrals too. If you buy through a referral, you get a $1,000 price reduction. Tesla prices are not negotiable in the showroom, and a purchaser can simply enter the referral code when ordering online, or in the showroom.

Using an existing platform, or creating your own

Rewarding loyalty from existing customers is relatively easy. My wallet contains everything from Starbucks and Café Nero loyalty cards with sequences of punch holes, to airline loyalty cards. Putting a system in place to encourage your fans to refer their friends can be more complex. There are several online platforms for the work, and I have no particular preference among them. Starbucks, Virgin America, Staples, Lands’ End, Microsoft and other companies have used Extole (www.extole.com) for their online referral programs, allowing Promoters to use Twitter, Facebook, email and text messages to share their experiences and link to web pages and social media. Both the referrer and referees are eligible for rewards.

At the simplest level, if you operate at retail level and use a physical loyalty card that tracks ten coffee purchases (for example), one approach would be to give a 50%-off refer-a-friend coupon to each customer who completes a loyalty card. While it is true that the coupon will not necessarily be used for a new customer, the same is true of Extole and other systems for consumer businesses that do not have exhaustive records of all customers.

Promoters may not have purchasing power

Where your customers are companies, end users of a product or service often have no authority to order from you. They have some influence over the person who can order from you, and you need to stimulate them to exercise that influence. But it may be a mistake to believe that individuals cannot help. Things have been changing radically in high-tech and it is easy to see how the trend is spreading elsewhere. In the old (but recent) days, anyone buying software in a medium to large company had to go through the IT department and they in turn went to procurement. The reason was that you needed new hardware to run the new software, and the IT and procurement people controlled hardware purchases.

Let’s suppose you have happy end users of software in a company. You would like them to consider your new product. Ensure it is available in a free version that limits (for example) the number of users. This is how messaging application Slack got started. Individual groups of users that did not have to pay anything spread their enthusiasm to departments and whole companies. Since it is Cloud-based software, no new hardware is required to run Slack. This means that departments can order it online, on their own. At a greater scale, this is also how Salesforce.com grew and beat Siebel. Sales organizations could decide to use Salesforce.com even without the agreement of the IT department, as no new hardware was needed. Of course, whoever was buying still needed a budget.

Conclusion

Promoters are not the only people who should be ‘activated’ to help you. After all, most people who would like to recommend your company have probably not answered your survey. However, Promoters are a great starting point. You know you can approach them with confidence, since they have already said they are likely to recommend you. Start with Promoters, then talk to others who seem to like your company a lot.

Next time

The next article will be about where and how you should concentrate your implementation energy, based on personal experience.

August 23, 2019

NPS (25) – The Outer Loop and the Huddle – 25th article in a series on the Net Promoter System®

Maurice FitzGerald / Blog, Net Promoter System, NPS / Customer Experience, marketing, Net Promoter System, NPS, strategy, Virgin Atlantic /

Welcome to the 25th article in my series on the Net Promoter Score and System. The subjects this time are the Outer Loop and the Huddle. You can read all of the articles here.

The Outer Loop is where generic, systematic improvement opportunities get addressed. There are two types of Outer Loop activities:

The first are those that can be addressed within a single team or function. For example, if your French customers consistently tell you they do not want service center people to use their first names when speaking to them, that can be addressed within the service center.
The second category of improvement work is that which requires collaboration across different businesses and functions. If your customers tell you that you need to be able to deliver products within 24 hours, when your current minimum is a week, different parts of the company will need to work together to make the improvement.

Prioritizing Outer Loop work

The essence of any business strategy is the concentration of your scarce resources on just a few things that really matter and help you to beat your competitors. When one initiative is complete, start a new one. For the cross-functional work, I recommend having three to five initiatives, no more. It will make the work relatively easy to communicate and to resource, and the team will see regular progress. The (separate) mono-functional teams should have the same limit. A business or function does not necessarily have to have an improvement initiative. You should not waste resources on things that customers don’t care about.

Useful sources of data for Outer Loop work

Here are the sources of data that are most useful for Outer Loop work:

If you have it, the competitive NPS benchmark data should be the most important source. While it might seem subtle, the purpose of your work is not continuous improvement of everything; rather it is the improvement of a few things that make the biggest difference competitively, whether to catch up, or to provide differentiation.
Relationship feedback data is generally the second-most important source to use. If your target customer base is a small number of large companies, it may be all you need.
Opportunities that the Inner Loop process has not been able to address are in third place. The reason they are not at the top of the list is that they need to be prioritized in a broader context. While it may seem odd, there will be things a lot of customers would like you to improve but where there is no real benefit in doing so, compared to other potential actions.
Operational data is necessary to establish the current state of things customers complain about. For example, your customers may complain about long delivery times for products, and your operational data may confirm that the most popular product was out of stock for two weeks, but it is available now. You may decide to discuss how that situation can be avoided in the future or may just decide that it was a one-off event. Personally, I totally failed to fix one of these situations at HP, despite trying hard for years. Our manufacturing plants for servers shut down for a week at the end of each calendar year. The order management systems continued to provide customers with estimated delivery dates that did not take the shutdown into account. I tried to get this changed for four years in a row, without success. I believe the root cause of my failure was a corporate function that was being measured by these committed dates, rather than actual shipment dates. And don’t get me started on a metric where the ‘shipment date’ was the day something was put into a shipping container, rather than the day the shipping container actually left.

Managing improvements — the Huddle

The primary vehicle for the selection and ongoing management of improvement initiatives is the Huddle. The main mono-function Huddles should be weekly or more often and can be very brief. The majority of their discussions and work need to be on things team members can do on their own. You can complement these by longer meetings on a less-frequent basis. Cross-functional Huddles are harder to coordinate and should be planned as monthly events. The same cross-functional team can have a longer session twice a year to work on the overall customer experience strategy.

Typical agenda for mono-functional Huddle

Like any other meeting, there needs to be an agenda and people need to prepare for it in advance. Here is a reasonable starting point:

9:00 Welcome — John
9:05 Check status of items on Huddle action tracker — Paul
9:10 New feedback from transactional surveys — George
9:20 Status update on Order TAT initiative — Ringo
9:25 Any Other Business and Close

Fifteen minutes can be plenty for a huddle too.

A great method for prioritizing work

Jim Dixon of Bain has published a great methodology for prioritizing improvement work. It finally addresses the challenge of making customer journey mapping into something truly useful. Journey Mapping is a popular customer experience technique. Its main disadvantage is that it considers every customer touchpoint as equal. It also fails to distinguish between hygiene factors and things that give you a competitive advantage. Jim’s work solves both problems.

Hygiene factors

A hygiene factor is something like taking a shower: nobody cares if you have one, but everyone will notice if you don’t have one for a long time. Think about a real-world example: you don’t think about data security when using online banking. You would not recommend your bank solely because they did not leak all your personal information and passwords. However, you would certainly be very unhappy if they did.

Jim Dixon takes this into account by considering each touchpoint in terms of its relative ability to delight or to anger customers. The source for this information should be customer comments on the open questions in your feedback system. If Detractors say that long waiting times are the top reason for them being a Detractor, and Promoters never mention waiting times, waiting times only have the potential to anger, and are a hygiene factor.

To an extent, this logic is built into the best of the customer journey systems. What Jim adds is competitive benchmarking. This gives you a way of visualizing and taking account of areas where your competitors have an advantage, as proven by your research. The diagram below is from Jim Dixon’s work and shows to combine it all into a single chart.

Use in a Huddle

If an Outer Loop team were considering the graph to prioritize improvement work, the two circles with diagonal lines would take top place. This example does not have any green circles, which would be areas where you have a competitive advantage you may want to push even further.

You would probably avoid taking any action on items in the bottom-left box, as your customers don’t care about them. Existing ‘bottom-left’ improvement work can and should be wound up to provide resources to work on the new priority areas. This process should be used to establish the initial ‘long list’. A further step is needed to get to the final short list.

Prioritizing initiatives

The image below is a simple diagram that will help you to prioritize improvement initiatives based on impact and ease of implementation. In the diagram below, you would get to work on the ones in the top-right box right away. This example is from a survey of Swiss business school students.

(The image at the top of the article is copyright Bain & Company. Used with permission.)

Next time

The next article will provide suggestions about activating Promoters. These are the people who have said they definitely would recommend your company. How can you help them to actually do so?

August 19, 2019

NPS (24) – The Inner Loop – 24th article in a series on the Net Promoter System®

Maurice FitzGerald / Blog, Net Promoter System, NPS / Customer Experience, marketing, Net Promoter System, NPS, strategy, Virgin Atlantic /

Welcome to the 24th article in my series on the Net Promoter Score and System. The subject this time is the Inner Loop; the process for quickly addressing the concerns of individual customers. The ‘class issues’ that affect whole categories of situations and customers will be covered next time, in my article on the Outer Loop. You can read all of the articles here.

Your reliable trusted research process produces feedback that individual employees and teams need to understand and react to. Let’s consider the simplest and most common use of the Inner Loop: transactional surveys. A service agent in your technical support team has just finished a call with a customer and closed the incident in your CRM system. The system automatically sends a feedback request to the customer. (The service agent has of course told the customer to expect it.) Surprise, the customer says their problem was not solved, but they were able to find a different way to solve it on their own. First, this feedback needs to go directly back to the agent concerned. Personally, I feel it should go back without human intervention if possible. I don’t see a reason that the agent’s manager would need to know, depending on how the agent reacts of course. While I have picked an unsolved problem as the example, the agent should get all positive, negative and neutral feedback directly.

Follow up with the customer

Ideally the agent or his / her peers or manager should follow up with the customer to thank them for the feedback, no matter what it is. The call should happen within 24 hours, preferably much more quickly, before the customer forgets about what they have written. Unfortunately, most companies say they are not able to afford the ideal process and only call back the Detractors, at best. (I can’t count the number of Detractor responses I have given to Air France and Sabena surveys without ever being contacted in any way. Maybe that contributed to Sabena’s financial issues and its ultimate disappearance from the market.) If you have made the (bad) decision that you can only afford to call the worst Detractors, start with just those who gave you a zero or 1.

There is no formal script for such calls. You need to apologize, then have a discussion. Apologizing works, and you must do it on behalf of the company. “I am sorry we screwed up.” not “I am sorry they screwed up.” Whatever your cost limitations, you must at least thank every customer who took your survey for doing so. Even if you use a standard email for this, it should come from the head of the appropriate department and give examples of improvements that have come from prior customer feedback. The department head’s email and phone number should be included, with an invitation to call or write. That won’t cost much and at least your angriest customers get a way of sharing their anger directly with you, rather than just spreading it across the Web.

No call-back script? Really?

It is obvious when people are speaking to you from a script. It comes across as insincere. You need a formal or informal expert system for the unscripted call-back dialog to be consistent and provide value. The best informal system is coaching from a supervisor, supplemented by discussions with teammates doing the same type of work. Teammates can help during live calls if you have an instant messaging system. “Let me just check whether any of my colleagues have seen this.” Then a short message to the WhatsApp or Slack group for the team, for example. I have also seen this work in a small software development team where the person taking the call would simply (and audibly to the person on the line) talk to a more expert person near them to resolve the situation.

Don’t be afraid to over-recover Detractors

When I joined the DEC Software Product Group in Geneva, we had to travel regularly to Boston. There were no non-stop flights, and we tried a variety of airlines. The team favorite quickly became the relatively new Virgin Atlantic. To get to Boston, we transited via London’s Gatwick airport. We all loved the airline’s ‘Upper Class’ and were rabid Promoters, often asking other colleagues why they would go any other way. I am sure we contributed lots of money to the Virgin bottom line.

After about a year of this, one of my colleagues, Peter van de Moosdijk, had a relatively serious incident. His flight from Boston to London was delayed for three hours due to a technical issue. Virgin realized that it was the middle of the night back in Europe and handed out a form for each passenger to fill in, saying whom they wanted Virgin to contact in the morning and what they should say. Peter filled out his form and boarded the delayed flight happy that his wife would know that he was going to miss the connection in London and would take a later flight.

He phoned home on arriving in Gatwick, and spoke to his panicking wife, who had been to Geneva airport to pick him up, and had heard nothing. The following day, Peter wrote a (paper) letter to Virgin’s CEO, Richard Branson, recounting what had happened and saying that this was not what he had grown to expect. Three things then followed. I particularly like the first:

A few days later, the doorbell rang while Peter was at work and his wife was at home. Virgin had sent her a huge bouquet of flowers and a card with an apology. I feel this demonstrated a rare level of insight about who actually had the problem. Not their customer.
The following day, a letter arrived, apparently signed by Richard Branson (or someone who could copy his signature). He said that what had happened was unacceptable, and that he had sent a copy of Peter’s letter to every Virgin office in the world to ensure it would never happen again.
At the time, Virgin was the only European airline with a frequent flyer program. Peter’s account was credited with enough miles for his family to go on vacation in California the following summer.

Now that’s over-recovery. Rather than talking about the negatives, Peter, his colleagues and friends all wound up being Promoters, telling this story over and over again. Our manager even attended a customer experience event hosted by British Airways. He asked whether any of the BA people had flown with Virgin and they had not. Then he told the story. Silence in the room.

Calling Promoters back

If a customer has told you they would definitely recommend you and that the latest transaction has made this even more likely, you need to have a ready way of activating them; meaning letting them actively promote you. In many businesses, you could simply thank the customer and say you would really appreciate it if the customer could write a product review on a specific website. You may also like to invite them to join your user group or give them a free ticket to a company event. Transactional feedback comes mainly from end users of products and services. In B2B situations, they will not normally be the ones making the purchase decisions. End users have a more direct impact in consumer businesses.

Why talk to Passives?

In businesses where the cost of change is high, Passives are the category that are most likely to leave you. This is the case for B2B companies that sell things on the basis of an annual contract, or who have dominant market shares in some area. While that may seem surprising, there is a certain logic to it. Passives normally represent about 40% of your customers. You tend not to talk to them much, if at all. The Detractors get lots of attention and often wind up happy. The Promoters are people we like to talk to because they love us.

The Passives? The typical attitude of most companies is “Well, they are not complaining, so I am sure everything is going just fine. No need to talk to them.” If you have any competitors, this is a disaster. You have to assume that your competitors are already talking to your Passive customers and being particularly nice to them. If you seem to have forgotten about them, they will leave you. Of course you can try to stand that on its head, work out which of your competitors’ customers are their passives, and start to talk nicely to them.

If you don’t have the resources to talk to everyone

How should you prioritize follow-up if you do not have enough people to call everyone?

Your top priority should be the people who have provided long responses to the verbatim questions. They care. They have taken a lot of time to provide thoughtful input. You should take at least as much time to call them back and discuss their input. Anyone who provides lots of suggestions is a potential Promoter, regardless of current status.
The second priority is your Detractors.

Of course, if you don’t have the resources to call anyone back, you really should not be surveying.

Email and instant messaging for collaboration

Many teams use email to discuss and resolve complex customer situations where people outside service center operations have to get involved. It is not a great solution. Email makes it difficult for people who have just joined the conversation to catch up on what has happened up to then. The use of CCs in email chains may mean that no single individual has the full picture. Sophisticated instant messaging systems like Slack may help if you want to be able to view and search the entire conversations easily. It is tricky to find a good balance, as people who only need to contribute something to a small part of the discussion should not have to be notified of every additional message.

(The image at the top of the article is copyright Bain & Company. Used with permission.)

Next time

The next article will be about the Outer Loop.

August 12, 2019

NPS (22) – How to almost do customer research (AppDynamics) – 22nd article in the series on NPS

Maurice FitzGerald / Blog, Customer Satisfaction, Net Promoter System, NPS / Chief Customer Officer, Customer Experience, Customer Satisfaction, CX, Net Promoter System, NPS, strategy /

Welcome to the 22nd post in my extensive series on NPS and the Net Promoter System. In the last two articles I discussed the major mistakes made by British Airways, Audi and others. This time I cover a company that almost gets it right. Unfortunately their results seem to be a sort of ‘vanity metric’ without a robust source.

A company with industry-leading NPS performance and a customer-centric culture can ruin their message by exaggerating, providing misleading information, and ‘comparing apples with oranges’. When you are performing well, there is no need for such communication. What follows is an example of an over-enthusiastic marketing department using an internal NPS result to (presumably not intentionally) misinform customers. It provides a wonderful example of messaging that gets ahead of the actual facts.

The press release

AppDynamics is a company that provides software for managing computer systems. They have recently been acquired by Cisco. Unfortunately, they have become masters of misleading communication about NPS scores. They have repeatedly published silly press releases that discredit their excellent performance and have the potential to discredit the Net Promoter System. In an early example in September 2015, they led with this headline:

AppDynamics Continues With Category-Leading Net Promoter Score of 85 for the First Half of Fiscal Year 2016

More than 500 customers weigh in to confirm positive experience with Application Intelligence Platform

They go on to say:

AppDynamics, a leading application intelligence company, today announced that it achieved a Net Promoter Score, or NPS® Score, of 85 for the first six months of its fiscal year beginning February 1, 2015. AppDynamics has maintained an NPS score over 80 for each successive six month period over the last three fiscal years, demonstrating its leadership among enterprise software companies, which on average earn NPS scores of mid-twenties according to Satmetrix.

And follow that by:

According to a benchmark study from the Temkin Group, scores for the overall software category in 2014 range from the low teens to the mid-40s. Companies with NPS scores of 60+ are generally considered top performers; USAA (insurance and financial services) and JetBlue are on this elite list, as is Apple. Enterprise software companies have typically had average scores below 20.

So, what is wrong with this?

Here are the main defects:

They do not provide any information about the source of the score of 85, though it is implied that it comes from an internal initiative called the Customer 360 program. No other company appears to be covered by that program, so it is not a competitive benchmark.
They mention the respected Temkin benchmarks without mentioning that Temkin do not include AppDynamics in their data.
Strangely, they were already well above average for the software industry with scores in the low 40s at the time, according to a double-blind benchmark studies by IPSOS and Qualtrics that I used at HP. The same studies also show IBM, CA, HP and BMC between 40 and 50 for Business Service Management software. The average score for IT management software overall is 38, with 8,800 customer responses. 175 people responded for AppDynamics. VMware was leading at 59.
While software industry benchmark scores have improved markedly since September 2015, 85 was not a credible benchmark score at that time. No company on the planet had ever had a score of 85 in any double-blind benchmark survey in any industry by the end of 2015. I have personally validated the methodology Qualtrics uses with Bain and it corresponds to their views on how double-blind benchmarking should be done. I shared the benchmark scores with various competitors before I retired from HP in early 2016. We discussed the AppDynamics press releases too. Some competitors had commissioned their own benchmark surveys, and found similar results to what I mention for AppDynamics. Nobody found scores higher than 45 at the time.

The software industry has been improving, and here in 2019 it has become more credible that such high scores are possible. I have no doubt that the Customer 360 program is making a difference. My former colleagues tell me that AppDynamics has produced the highest benchmark scores seen so far. They and their competitors have all improved. It will be interesting to see how this trends.

Temkin benchmarks

For the sake of completeness, the illustration below shows the Temkin NPS technology industry benchmarks for 2016, the time of the press release shown above. 62 vendors are covered, almost all of them software vendors. Even without the intermediate points on the table, you will quickly see that the AppDynamics statement that “Enterprise software companies have typically had average scores below 20” is simply not true. Note that most Temkin reports are free of charge, ever since Qualtrics acquired Temkin Group in 2018. Of course the AppDynamics statement could be based on yet another unattributed source.

Conclusion

When they had great performance, the sort of press release, website content and sales material AppDynamics used in 2015 did it a disservice. Don’t publish scores from your own company surveys and represent them as objective benchmark comparisons. Your competitors know the numbers are wrong and many can easily provide your customers with correct data. The correction may call into question other claims you have made about your products or services. If an NPS number appears to be too good to be true, it is!

If you persuade your senior management that the comparison is fair and real, you also put your future work at risk. After all, your mission seems to have been accomplished. The work should now be put into business-as-usual mode and the investment scaled back. As it is, AppDynamics was on a path to market leadership, though perhaps not at quite the speed their communications people believed in late 2015. Perhaps this is why Cisco acquired them in early 2017. Finally, note that in the spirit of fairness, I wrote to the person whose name was on the press release to make these points, and he did not reply.

Next time

The next article will be the first of two on the heart of the Net Promoter System: the feedback, learning, and improvement processes. The process has three elements: the Inner Loop, the Outer Loop, and the Huddle.

August 10, 2019

NPS (21) – How not to do customer research (Audi and a supermarket) – 21st article in the series on NPS

Maurice FitzGerald / Blog, Customer Satisfaction, Net Promoter System, NPS / Audi, Chief Customer Officer, Customer Experience, Customer Satisfaction, CX, Net Promoter System, NPS, strategy /

Welcome to the 21st post in my extensive series on NPS and the Net Promoter System. Following the numerous reactions to what I posted about the British Airways research process last week, here are my views on the way car dealers survey customers in general and about Audi’s process in particular. I also cover a supermarket.

I suppose I should start by emphasizing that Audi is far from being the only car manufacturer doing what I am about to describe. It seems to be a standard practice for car brands.

Buying a car

Following her purchase of a new Audi, my daughter received a survey request from her salesperson. He provided the link, and added a covering note saying, “Please contact me if you cannot give me a 10.” I was surprised to see evidence of this horrible practice reaching European car dealers. It is the rule, rather than the exception in the USA. The General Motors head of customer experience explained it in a pubic seminar. Sales people typically need to average over 95% satisfaction in the GM dealer surveys to be able to participate in their bonus scheme. One low score can require ten perfect scores to even it out.

David Mingle, Executive Director, North American Customer Experience for GM said that he knows this and is not going to do anything about it. He said he personally pays no attention to the scores, only to the written comments. He says customers tend to be honest in their written comments. The sales people are only measured on the scores and don’t care what the customer writes about the experience. This is clear evidence of a research process that provides no value. If individual dealerships and sales people have no motivation to read and act on the comments, why should a customer answer the survey at all?

Consequences

While this may seem to be an interesting anecdote, though backed by facts, it has major consequences. These consequences are evident in the USA and will come to Europe too. In the USA, car prices are considered to be highly negotiable. There are plenty of websites that give advice about how much a General Motors franchisee has to pay for a particular car, providing the customer with the lowest possible price. I invite you to do your own searches. You will easily find suggestions to use the survey score as a price negotiation point. It is attractive in that the cost to the sales person of lowering the price is zero, while the benefit of a 10 may be high, and costs you nothing. If car companies internalized this, I believe they would change their bonus systems. Indeed, if you receive messages saying “If you can’t give me a 10…” during a negotiation process, you should probably negotiate a little more, no matter what it is you are buying.

Not the only defect in their survey

The Audi survey is somewhat strange. I feel they do not understand the difference between something that matters and something that makes a competitive difference. The illustration below shows the way their US dealer survey from a couple of years ago describes what a 10 means:

It is not appropriate to label intermediate numbers in rating questions, just the extremes. Personally, I find the difference between “Outstanding” and “Truly Exceptional” hard to understand, let alone worthy of a three-point difference. I also consider the appearance of the facility to be a hygiene factor. Once it is good enough, it has no impact on your purchase. I don’t see how it could be a motivator. The same applies to the “comfort of the area or office where you agreed to the price of your vehicle.” What is the added value in getting a score above a five, for example? Indeed, perhaps it should be uncomfortable so you want to negotiate quickly and leave.

Of course, many other companies do the same thing. According to employees at Walmart, Old Navy, Staples and Best Buy, they are measured on receiving 10s and interrogated about receiving anything less. The slip of paper shown below is what one US supermarket allegedly gave customers with their receipts:

Conclusion

There are many ways to cheat or bias your feedback system. This is just one. For a number to be credible, the people who are asking you to fill out the surveys must have no direct motivation to achieve a specific score. They should only be motivated by the desire to improve.

Next time

In the next article I will consider the case of AppDynamics; a company that almost gets its customer research right but seems to fail at the last hurdle.

August 6, 2019

NPS (20) – How not to do customer research – 20th article in the series on NPS

Maurice FitzGerald / Blog, Customer Satisfaction, Net Promoter System, NPS / British Airways, Chief Customer Officer, Customer Experience, Customer Satisfaction, CX, Net Promoter System, NPS, strategy /

While I am sure it is not their intent, the way British Airways surveys both its general and special customers is the worst I have ever seen from any company, from a customer perspective. Perhaps paradoxically, it is one of the richest possible from an internal British Airways perspective. It provides them with a rich set of reports and metrics. While it includes part of the Net Promoter System, it has far too many screens and questions. They have never contacted me about my issues or improvement suggestions. What follows is a critique that includes screen captures from their September 2016 Executive Club survey. Please bear in mind that I am not entirely negative about it all. I just feel it works from an old paradigm, is internally focused, and designed primarily to fit customer experience strategy and reporting software supplied by KPMG Nunwood.

Summary

While the covering email from Nunwood says the survey will take 15 minutes, there are 134 screens to go through, and it is hard to see how it could be done in 15 minutes. The survey invitation is signed by James Hiller, with no job title. A Google search indicates he manages the Executive Club. Nothing is said about what they have improved based on past surveys.

As distinct from a pure survey about the Executive Club itself, competitors are also covered, as are BA and competing airlines. The huge amount of data collected probably fits a comprehensive internal reporting structure, and provides no particular value to customers. It violates the fundamental principle of customer research in that it fails to provide more value than it extracts. Here is how it starts:

Cher Monsieur Fitzgerald,

Given that you are an Executive Club Member, we would welcome your views on your experiences with British Airways and we would like to understand how they have influenced the way you feel about us.

Whether positive, neutral, or negative, your feedback is welcome and highly valued, as it will help us to shape our products and services in the future, and understand what our most valuable customers need from us.

The survey will take about 15 minutes to complete.

Will I be able to provide the input I want to provide?

Before starting to work through the many screens, I decided what I wanted to say. I wanted to make three points:

Transiting through Heathrow Terminal 5 for long-haul travel from Geneva is far more complex than transiting via Zurich, Munich, Frankfurt, Amsterdam and even Paris. This is because passengers have to go through a pointless duplicate security check inbound from the USA / Canada, and a pointless duplicate outbound check when arriving from countries that use the same security standards as the UK. While I did not know this at the time I filled out the survey, the UK Civil Aviation Authority requires this second check and does not accept screening done anywhere else. I suppose this may date from an era when it was too difficult to separate the ‘clean’ and ‘not quite so clean’ transit passengers from each other. Whatever the reasoning, it seems pointless today. The layout of Terminal 5 could easily be modified to create this separation should the law change. BA should apply the necessary pressure, in my opinion.
It is next to impossible to use the Executive Club website to reserve some long-haul destinations using miles. Entire months have no availability. I am not talking about high-demand routes like Christmas in Australia. My latest experience with trying to reserve Australian flights in Business Class was for October, seven weeks in advance, and there were no flights on any date, either to Sydney or Melbourne. December and January show no availability either, which is less surprising. During a Christmas vacation, I looked at availability for South Africa the following Christmas. Nothing.
I wanted the opportunity not just to complain, but also to make improvement suggestions. Among the suggestions, I wanted to be able to mention that BA should tell us what they are doing with the input, and should contact people who make suggestions.

The short answer is that I was indeed able to provide the input. The more complete answer is that I was forced to bury this critical information inside many unimportant and irrelevant (to me) answers. I am sure the 100+ irrelevant answers correspond to the way some staff inside British Airways are measured.

The start

I expected BA to provide a choice of languages for the survey, and they do. I suppose it is a bit surprising that Spanish is not among the languages. It is the second-most common first language in the world with about 400 million speakers, after Chinese, with 1.2 billion.

NPS questions

The recommendation question is next as shown below. Oddly the ‘overall satisfaction’ question is asked as well. After a question about emotions, the NPS “Why?” question is asked.

Logically, the following question should be about what BA should improve. That question is never asked about British Airways, though it is asked about the Executive Club, close to the end of the survey. When you do not ask for improvement suggestions, you tend not to get any from Promoters. Remember that Promoters are the customers who admire you most, and want you to succeed. They tend to provide the most valuable improvement suggestions.

Lots more questions

Questions about my feelings about the airline follow. I suppose these provide overall statistics and somebody is measured on the trend. It is hard to see how someone would work directly on improvements from the answers, as there is no opportunity to explain the input. The questions all use rating scales and include:

How likely would you be to travel with British Airways in the future?
To what extent do you consider British Airways a convenient option when choosing an airline to fly with?
Taking everything into consideration, overall how satisfied are you with British Airways?
How satisfied are you with the value for money British Airways offers?
How would you say you feel about British Airways?

Information about competitors

A useful section follows as shown below. After asking what airlines I have used for long- and short-haul travel in the last 12 months, the survey goes on to ask which airlines are my top choices for such travel. It then asks why the first-choice airline is in first place. I consider this section to be the only thing that is potentially useful beyond the basic NPS questions.

Oddly, they sort of throw this away by asking you to choose from a list of possible responses for not flying with BA. None of the preselected reasons was my main one, so I filled in the Other box. I feel they should have asked this as an open text question.

Then you are asked again about your overall satisfaction with BA. You can’t see the answers to prior questions and my second answer to the same question was one category lower than the first. Maybe because the survey was making my opinion more negative. The wording of the question was not exactly the same as the first one, though I don’t recognize them as different questions.

Lots of rating questions

A seven-point scale is then used to gather answers on flying touchpoints, emotions associated with BA and opinions about their communication. Finally, the section on the Executive Club starts, at last. Remember that the email invitation came from the head of the Executive Club, not the head of British Airways. Overall satisfaction and recommendation questions are followed by another extensive batch of seven-point rating scales Each rating question is presented individually, and you can review your answers to each section, possibly making changes.

Finally finally

A breath of hope is provided by the question “And finally, thinking about your experiences of any frequent flyer programmes, how could British Airways Executive Club be improved?” I mistakenly thought this was the last question. Unfortunately, I was then asked seven questions about demographic information that BA should already have in my Executive Club profile. This is one of the disadvantages of subcontracting surveys. The confidentiality agreement you have with your customers probably forbids you passing their details to a third party. However, the way BA has handled this with Nunwood is not optimal. Since they already had my name, there would be no reason not to provide them with a unique code, that would then allow BA to match my name and demographic details with their own database. The name on its own might be a duplicate.

Speculation and conclusion

I recommend a visit to the KPMG Nunwood website at www.nunwood.com. Their offers include customer experience strategy, journey mapping, NPS support and other comprehensive approaches to improvement. One of the approaches, the ‘Six Pillars’ explains their satisfaction surveys. The excessive number of questions satisfies the needs of the Six Pillars model, and fully exploits their Fizz reporting software.

I speculate that the survey and software are part of a larger consulting engagement with British Airways that requires a ‘slice and dice’ approach to data for distribution to a variety of internal departments. I feel this is an internally-focused way of proceeding. I would love to know what their response rates are, and speculate that they are under 10% overall. I suppose that higher-status members will be more likely to give feedback, and their input is critical, since they generate the most revenue for the company. I believe an interesting opportunity would be lower-status members who have higher status on other airlines. My recommendation for BA is an A/B split test, asking just the basic few NPS questions as a ‘pulse’ survey in between the main ones. I believe they would get better response rates and clearer actions to prioritize. Of course, they would lose the ‘slice-and-dice’ metrics they probably use to measure people. And finally, I have of course lost all my status and privileges with BA since retiring, so don’t know whether they have changed the way they survey Executive Club members since then. Please feel free to reply below if you have new information.

July 31, 2019

NPS (19) – How to calculate the lifetime value of a customer – 19th article in the series on NPS

Maurice FitzGerald / Blog, Customer Satisfaction, Net Promoter System, NPS / Chief Customer Officer, Customer Experience, Customer Satisfaction, CX, Net Promoter System, NPS, strategy /

While you can find many methods of doing the calculation on the web, there is only one correct answer. The value of a customer to your company is the total cash flow they generate for you over their entire lifetime. The single calculation method for this is Discounted Cash Flow; the Net Present Value of all cash flows out into the future. No other method is correct, though there are some acceptable approximations. The theory and experience of NPS practitioners is that the operating profit per customer is highest for Promoters and lowest for Detractors. Do not confuse revenue and operating profit. You can have customers that generate lots of revenue, but even more cost. They have negative lifetime cash flow, unless you can turn them around.

The main problem you will face

Unfortunately, most companies are not able to calculate profitability by customer. Financial reporting usually works by product line, not by customer. Asking your finance person to provide you an accurate cash flow from each of your top 100 customers individually will probably be a waste of time. Businesses that depend on large contracts are an exception, and you may well be able to calculate the cash flow from each contract.

Suggested method

In the absence of accurate standard reporting, a good starting point is to work with your finance person to calculate a ‘proxy P&L’ for each of your different customer categories. Unfortunately, there are accepted accounting standards for many things, but not for the precise methods used to calculate the lifetime value of a customer. Segmenting customers in a useful way is a critical step. The first distinction to be made is between customers who have an ongoing contractual relationship with your company, and those who buy things, but have no contractual relationship. The former category is easy to work with. Past contract renewal rates can be used to determine future expectations. Direct Marketing principles need to be applied to the latter category. Direct Marketers think in terms of ‘RFM’: Recency, Frequency, and Monetary value.

Other customer segmentations that are useful will depend on your business model. If you sell both direct and via resellers, lifetime value expectations will be different. They will also differ between very large customers who depend on your company to run their business, and others. If you have the data, NPS categories can be used to refine renewal rate expectations. Here are examples of what needs to be used in the calculation:

List prices.
Contra-revenue, meaning discounts you give from the list price. While large customers tend to attract higher ‘standard’ discounts, you will generally have easier discount negotiations with Promoters than with Detractors. Since contra-revenue is not part of any standard P&L, it is often forgotten.
Net revenue, meaning list price less contra-revenue.
Cost of Goods Sold. This includes costs of ‘free’ replacement items and additional merchandise that you may have to supply Detractors to help turn them around. If you have a dedicated escalations team, you can easily calculate the average cost per escalation and use that to assign a total escalation labor cost. While ‘Cost of Goods Sold’ sounds like it should include the cost of selling and marketing, it does not. Selling, General and Administrative costs are ‘below the line’, which means they are not counted in the Gross Margin calculation, but are considered when calculating operating profit.
Where Detractors or Passives leave you sooner than Promoters, you can assign a cost of attracting a new customer. Many sales teams are split into ‘Hunters’ who go after new customers and ‘Farmers’ who develop deeper relationships over time and are measured on retention and additional business for existing customers. Farmers generally cost about a quarter of the Hunter cost per revenue dollar. If you don’t split your sales teams this way, agree a reasonable rule of thumb with your finance partner while you try to measure it more accurately. Marketing costs can be split that way too.
If you happen to have any indication that you get some of your new customers by referrals from your happiest existing customers, it is worth agreeing a way of taking this into account. However, if you already have a strong enough differentiation between Promoters, Passives and Detractors for your purposes, I suggest this simply be listed as an additional factor that you have not yet quantified. Attaching a reliable number to it can be tricky.
Beware of the law of small numbers. If the most positive and negative numbers come from unusually small groups in your analysis, you should find a way of combining them with a larger group until you can validate the reality of what you have seen.

If you are not able to get to this level of detail, calculating expected lifetime revenue per customer is a lot better than doing nothing.

If you can only work on one thing for a contract business

If all this is too sophisticated for the time you have available, here is the single area you should focus on, if your business depends on annual contracts. Just agree the value of a one-percent improvement in contract renewal rates. It is easy to calculate, and you probably have the data on past renewal rates. If you work in a large company, you have de facto experiments in place. By this I mean you can simply observe different measurable business practices or levels of customer-centric performance that have happened by accident, rather than by design. Look at how they have affected renewal rates. If you seem to have nothing relevant, test new practices in part of your company and compare renewal rates. You can test multiple things simultaneously, provided you keep a control group, where nothing has changed. Improved renewal rates were how we justified Customer Success teams at HP Software. These teams worked with customers, free of charge, to make sure they were using software effectively and achieving the benefits they expected.

Summary

Compared to the others, Promoters stay with you longer, buy more products, escalate less, are less price sensitive and are less expensive to sell to and service. There are many businesses that have marginal overall profitability. If you have customers that have been generating negative margins for some time, you will have to decide whether you can turn them around or whether you should drop them as clients. Clarity is more important than always getting the answers you would most like to hear. Once you have agreed the lifetime values of the different NPS customer categories, you and your CFO (or delegate) should present it together to your leadership team for approval. Even if you are not able to do it right away, it should be part of your formal NPS implementation plan. Your credibility depends on it. Your company’s long-term commitment to NPS also depends on it.

Looking forward

The next article in the series will be a counter-example, meaning an example of what you should not do when conducting customer research. The first one will be from British Airways.

Learning check

Here are some true / false questions to check whether I have been successful in communicating my main messages in the last few blog posts. The answers are at the bottom of the page.

The cost of renewing an existing contract is typically one-tenth the cost of acquiring a new customer.
If your overall brand NPS trend is better than that of your competitors, it is safe to assume that all parts of your business will benefit in the same way.
Bain experience suggests that the average time lag between a relative NPS improvement and new product revenue improvement is 6 to 18 months.
Since contra-revenue is not in a standard P&L statement, it should not be considered when calculating the lifetime value of a customer.
To avoid bias, all customers should be considered equal in your research, no matter how much they spend with you or your competition.

Answers to the learning check questions:

True: Renewing a contract is mainly an administrative effort and costs far less than hunting for and acquiring a new customer.
False: You need to do the analysis for each part of your business for which you produce a P&L.
True: This is to do with the average purchase cycle. Even if they love you, people are relatively unlikely to buy the same thing again right away. This varies widely by type of product.
False: Promoters tend to get lower discounts than less happy customers and discounts from list prices are not in your P&L in most accounting systems.
False: You need to distinguish between customers who spend a lot and those who don’t spend much on your type of product or service when calculating lifetime value. Large spenders tend to be more demanding than small spenders, and therefore tend to give lower NPS scores.

July 26, 2019

NPS (18) – Relationship between NPS trends and revenue / market share – 18th in the series on NPS

Maurice FitzGerald / Blog, Net Promoter System, NPS / customer culture, Customer Experience, Net Promoter System, Surveys /

This is the 18th article in a series on the Net Promoter Score and Net Promoter System. The subject this time is a critical one: the relationship between NPS trends and revenue. The topic has been somewhat controversial over the years, though I am not entirely sure why.

If you work for a for-profit organization, every activity should contribute to your operating profit. If you personally have no credible background in finance, your company Chief Financial Officer is critical to your success. There are two main areas where you must be aligned: proving the relationship between NPS trends and revenue trends and calculating the lifetime value of a customer.

Relationship between NPS and growth

You should ask your CFO to assign a person to work with you to validate the relationship between NPS and growth (or decline) in market share. Expect to propose the methodology and ask that person to assist with the data; both your own and that from your competitors. Organization changes and M&A can make comparisons difficult and a Finance person will be well-equipped to read and understand relevant P&L statements. Assuming you are basing the comparison on NPS data from a reliable benchmark provider, there should be no major difficulty with the analysis. It is natural that you will be able to segment your own data better than that of competitors. NPS trends usually, but not always, predict market share trends. In Bain’s experience, NPS trends explain and predict between 20 and 60% of market share trends. While NPS trends are an important growth predictor, they are of course not the only thing that matters. You or a competitor might get distracted by a merger with another company, for example, and that might counteract any NPS gain. There are also types of businesses for which NPS does not provide any predictive value. An example follows.

Calculating the relationship between NPS and growth

The only reliable basis for calculating the relationship between NPS and growth is a well-executed double-blind benchmark survey. The reason no other survey type provides a reliable data source is because no other research gives you unbiased data on your competitors. Imagine a situation where you only have a single competitor, and you know nothing about their NPS trend. Your own score moves from 22 to 35 over three years. You pat yourself on the back and assume increased market share will follow. Maybe. If your competitor has improved from 10 to 32 over the same period, they will probably take share from you. The thing that matters in the relationship between revenue and NPS is your trend compared to your main competitor’s trend. If you are in a highly-fragmented market it is OK to compare yourself to the overall average.

Use cubic regression

Since you will not have a huge number of data points for your analysis, cubic regression will usually provide a better data fit than linear regression, quadratic regression or simple correlation. All are easy to test in Excel. Assuming there is a relationship between relative NPS and revenue, a good rule of thumb is that cubic regression will show it to be about twice as strong as a linear fit. An advantage of using multiple regression (rather than correlation), is that you can test the relationship between other factors like an economic recession or GDP changes within your overall model. You can get surprises. The graphs in the following paragraphs show the relationship between relative Net Promoter Score and growth for one HP business, as measured over a period of eight years. The absolute numbers have been modified somewhat from the confidential data, and the trend indications are accurate. We got 20,000 competitive benchmark survey responses per year.

Analyzing subsets of a business

Within an overall business, you can and should examine trends for subsets. In one HP Software business, we looked at the effect of NPS trends on license sales, professional services sales, and remedial support contracts. While we found the results somewhat surprising, they do agree with what Fred Reichheld and Rob Markey have found over the years, namely that NPS tends to have little or no predictive value for support / service center types of businesses. My interpretation of this lack of predictive value is that NPS can’t predict revenue trends for ‘hygiene factor’ businesses. These are the businesses that can anger customers but have no real ability to delight them. Let’s discuss each of the three businesses in turn. The first graph shows the results for the overall business. Note that these numbers are not the actual confidential HP numbers, but are all notionally correct.

Licenses

Product purchased follow a life cycle. Customers buy things, and if they are happy they will buy more after some time. In Bain’s experience, the average time lag between a movement in NPS relative to competition and a corresponding movement in product sales is about 18 months. This graph shows approximately what it was like for a major HP software business:

Professional services

Consulting services include training, installation services and other services associated with using software as effectively as possible. The services are not proprietary, which means customers have alternatives. The effect of this is that changes in NPS produce changes in professional services revenue more quickly and more durably than for product licenses, as can be seen in this graph:

Support

Reichheld and Markey have long held that the recommendation question is not a good predictor for most support services, meaning services where the perceived value is the ability to call on someone to help you with a product when you have a problem. They don’t discuss why. I believe it is because the work of service centers is a hygiene factor. Once it is good enough, there is no real benefit to making it great. The recommendation question works best for products and services that can truly be great. This is what the data showed for HP Software support services, backing up that point. There is no significant correlation between overall software brand NPS scores and software support revenue. Businesses that only or mainly do service center work would be an exception. The non-relationship applies to businesses where the support center exists to do just that, answer questions about a product or service after it has been sold. Personally, I suggest using Customer Effort Score for research about support services.

Looking forward

The next article in the series will be on how to calculate the lifetime value of a customer.

As is often the case, the above is a slightly-edited version of a chapter in one of our books; in this case Net Promoter – Implement the System All of our books are available in paperback and Kindle formats from Amazon stores worldwide, and from your better book retailers.

World map used for blog on cultural differences

July 19, 2019

NPS (17) – Does culture affect NPS / customer survey outcomes? – 17th in the series on NPS

Maurice FitzGerald / Blog, Net Promoter System, NPS / customer culture, Customer Experience, Net Promoter System, Surveys /

This is the 17th article in a series on the Net Promoter Score and Net Promoter System. This time I cover cultural differences in the way people respond to surveys.

As a person who has lived and worked in seven countries, I believe cultural differences exist and matter. My career has taken me regularly to countries in Africa, Asia, Europe and North America and I have directly observed that people behave differently, especially in business situations. A lot has been written on the subject. My favorite book on the subject is When Cultures Collide by Richard D. Lewis ^[1]. Cultures and Organizations by Hofstede and Hofstede^[2] is also excellent. I got into the habit of studying the When Cultures Collide chapter on a new country before visiting for the first time. I suppose my most educational experience was in Jordan.

Comfort zones in different countries

Along with all the science, Lewis tells entertaining stories about the effect of comfort zones on behavior in groups. In Northern European and North American culture, the ‘comfort zone’ extends to arm’s length. What this means is that people from these cultures become uncomfortable when someone is less than a full arm’s length away from them. They will tend to back away from someone the feel has come too close. In Southern Europe, the comfort zone is half that, meaning elbow’s length.

Lewis describes multinational meetings, for example European Union social gatherings, where the Southern Europeans back all the Northern Europeans into the corners of the room by the end of the evening. His chapter covering Jordan mentions that the local comfort zone is half again what you find in Italy. He also mentions what happens when you are being successful in your discussions. He was not wrong.

Opening HP’s office in Jordan

I went to meet the Minister for Information and Communication Technology together with an American colleague. We wanted to negotiate a series of things to do with opening an office and potentially a software center in Amman. My American colleague had not read the Lewis book, and I did not say anything about what I was going to do in advance. I sat just left of the minister. When the time came to have the critical discussion, I moved to about 20 cm/8 inches from his face, and …this is critical… did not lose eye contact at any time while we spoke. I could see my American colleague getting twitchy while this was going on. And yes, the sign of success then happened. The minister put his arm around my shoulders. We had never met before. My colleague’s eyes were popping out. Success! I had learned how to do the male cheek kissing version that is specific to Jordan, but did not get to try it out as we did not socialize beyond the meeting. On a side note, the best coffee I have ever had in my life was in that office. Something to do with flavoring it with cardamom seeds, I think.

Culture matters, but does it affect customer surveys?

After many years using double-blind surveys at HP and seeing responses that were consistently different by country, I became certain that the sole explanation for the differences was culture. Exhibits 3.4 and 3.5 show what the NPS scores looked like by country for the combination HP and its major enterprise hardware and software competitors respectively, by country. This excludes consumer PCs and printers. The data is for a three-year period and the sample sizes are absolutely massive.

So therefore

The only possible conclusion from this data is that the Japanese, Koreans, Australians and Dutch are grumpy people and don’t want to give top grades to anyone. Right? The only possible conclusion? That is certainly what I thought before going to meet Rob Markey for the first time in August 2015. Before getting into what he told me, think about it this way. The data above represents customer views on the major IT suppliers around the world. Almost all are US companies. Almost all try to do business in other countries exactly the way they do it at home in the USA. Almost all believe that this is the best and only way of doing business and that any other way of doing business is inferior, or perhaps just costs more.

The perception is widespread

Do a web search on cultural differences in Net Promoter Scores, and you will find a consensus that the cultural differences in scoring are real and look something like the graph above. Even Bain consultants used to agree. Back in 2006, Jayne Hrdlicka, Edmund Lin, Gary Turner, Bruno Lannes, YeonHee Kim and Shintaro Hori published How to win—and retain—loyal customers in Asia and Australia . It includes many examples of NPS score ranges from different industries. Most of them have negative average scores. There are also some ‘normal’ scores in the mix, which should not be possible if the cultural paradigms are 100% correct. By this I mean that either ‘all Australians are grumpy when responding to surveys’ or they are not. Both cannot simultaneously be true.

There is data that supports the existence of cultural differences

Per.ceptive (Customer Monitor) is an Auckland-based company that publishes consumer NPS benchmarks for Australia and New Zealand, among other work. Their work on consumer benchmarks seems to support the ‘grumpy Australian’ viewpoint. The only other possible explanation would seem to be that all types of Australian companies don’t care much about their customers. The report in Exhibit 3.6 is from August 2016.

There is limited though reliable data that says the opposite

If the bar graphs really reflected cultural differences, then there would be no benchmark NPS scores in these countries that would show anything different in any industry. Not only is there in fact benchmark data from the ‘bottom four countries’ that has similar scores to those observed in the United States, but some of the data is for high-tech. ING Direct Australia scores a 38 in double-blind sampling. Local IT outsourcing suppliers working in Dutch score in the 30s in the Netherlands, though Dutch Telco KPN has an 11.

OK, so these are cultural anecdotes

Here are a few short examples that gradually changed my views about what was going on:

You should adjust to local conditions if you are not from the country. I remember being beside Francesco Serafini, then the HP Managing Director for Europe, Middle East and Africa, when the CEO of Philips Electronics phoned him. They had recently transferred their IT help desk operations to us. Nobody had told the CEO that one way cost had been reduced was by eliminating Dutch as a language for telephone support. He was angry, to say the least. Angry with us and angry with his CFO who had signed the contract and had not told him about the change.
According to a former Apple employee I met socially, the Tokyo and Beijing Apple stores had the lowest NPS scores among all stores worldwide, at one point in time. They did deep research into local habits after asking themselves whether the way they did business in New York and San Francisco was appropriate in China and Japan. The first thing to change was the introduction of ‘fast checkout’ for a set of commonly purchased items that were put against one wall. The customers could avoid the friendly-but-lengthy typical Apple experience and get out of the store quickly. NPS scores rose to and even above the company average quite quickly.
I met Bill Thomas of EquaTerra at an IT event. EquaTerra had just published the customer satisfaction benchmark numbers for the Benelux for the IT Outsourcing industry. All multinationals (HP, IBM, Accenture) did badly, and many local providers did very well. I asked him to tell me the top reason why. He said it was because all the local providers had their support teams physically in the Benelux and they all spoke to customers in their languages and accents. The multinationals all operated offshore, providing support in English for the Netherlands and Luxembourg, and non-Belgian-accented French for the Belgian customers. He said this was a pattern in other countries too. The multinationals may have had lower costs, but the costs were quickly forgotten after the contracts were signed.
A local radio station reported on an esoteric study on how people wanted to be addressed by doctors in hospitals. In the United States and the UK, the vast majority of patients wanted doctors to use their first names. In France none did, preferring their family names. I suspect the answers would be even more extreme here in the French-speaking part of Switzerland where you are expected to know someone before going further than “Sir” or “Madam.” Indeed, in a recent newspaper article a number of Swiss rail passengers were complaining that the staff inspecting their annual rail passes thanked them and used their family names, which they considered totally unacceptable.
During a discussion about NPS scores in Australia, a person I met at a Satmetrix conference asked me from where we provided telephone support for Australian customers. I told him it was done from India for the business in question. He told me that many Australian IT companies had tried that, but had moved support back onshore to have people on the phone with Australian accents, despite the cost. HP’s consumer support people did something similar when moving UK telephone support back to the UK from South Africa.

More Bain experience

I went to my meeting with Rob Markey armed with my data and convictions. He persuaded me that I was wrong, at least partly. He described his experience working with an unnamed American company’s operations in Japan. The Japanese operation had had the worst NPS scores in the company for years. To get a new answer, they asked themselves a new question, “What if the problem is not the supposedly grumpy Japanese? What if the problem is us, and the way we do business in Japan?” That was a transformative question. Bain and their client did deep research on the way Japanese customers wanted to be served. They adjusted the way they did business from top to bottom, and it does not look like any other country. Japan now consistently scores in the top three countries on the corporate NPS chart. Going back to the anecdotes, the company in question had had a policy of always addressing customers by their first names, even in Japan. Totally unacceptable. That was one of the first things they changed.

Suggestions

What all this means is that cultural differences exist in responses to NPS surveys, and the reasons may not be entirely what you think. At a pure survey level, yes there are cultures that are reluctant to give you the top score. This is a good reason to use the zero to 10 scale for NPS ratings rather than 1 to 5. People who have a cultural reluctance to give you a ten may still give you a nine. There are cultures where customer-centricity has not evolved to the extent it has in the United States. An American company offering US-style service there may do very well compared to local expectations. This may be what we see happening in Mexico and Brazil in the IT industry, for example. There are other countries where behaving in the excessively friendly American way may be totally inappropriate. If I were to suggest remembering one thing when you see low scores in a country, ask yourself, “What if the problem is not the customers? What if the problem is us?” Answer that question before your competitors and you will win.

Communicating multi-country competitive customer experience scores

If you have competitive benchmark scores and graphs like the ones shown above, you should never present them in the way I have. For most companies, competition is local. It does not actually matter much that you and your competitors average an 8 in Australia and a 50 in India. I suggest showing individual country competitive comparisons without showing the rating scale at all. Show the tick marks on the graph but no numbers, or just relative numbers compared to your score. What you want in Australia is for your NPS trend to be better than that of your competitors, no matter what the numbers. Comparing country scores to each other mainly wastes time and delays getting on with improvements. Only removing the numbers from the scoring scales can help avoid your audience making irrelevant comparisons.

^[1]Richard D Lewis: When Cultures Collide, 2005, Nicholas Brealey Publishing, ISBN 978-1904838029

^[2] Geert Hofstede, Gert Jan Hofstede and Michael Minkov: Cultures and Organizations, 2010, McGraw Hill Education, ISBN 978-0071664189

Learning check

Decide whether each of these statements is true or false. Answers below.

1. Even the font you use in an email asking customers to take a survey can make a big difference.
2. To make sure you are doing a good job, you should always send a survey request to every customer who phones your service center.
3. Embedding at least the rating question in a survey request email improves response rates by about 2.5 times.
4. Culture makes no difference to survey responses, so it is always fair to compare countries’ NPS performance with each other.
5. Low survey response rates tend to give overly optimistic NPS scores.
6. The more questions there are in your survey, the more likely people are to complete the survey, as it is clearer that you value their opinions.

[Answers below.]

Looking forward

The next article in the series will be about the all-important question of whether or not NPS trends predict revenue and market share trends. Deep stuff.

Answers to the learning check questions

True. Using a serif font like Times New Roman in emails improves response rates compared to using a non-serif font like Arial.
False. You should only survey if you have the people, funding and plans to act on the input.
True, at least in the author’s experiments. There are software vendors who say that it triples response rates.
False. Do your best to make all competitive comparisons local. Comparing NPS scores between countries is not helpful.
True, and the effect is substantial. You can only consider an NPS survey to be truly representative of the customer population when you have response rates above about 50%. Such numbers are rare for transactional surveys, another reason to be selective when using them.
False. The longer your survey questionnaire, the more people will drop out before completing it. Compared to a 3-question survey, about 6% of the people who start a 20-question survey will not complete it, rising to 10% for a 40-question survey.

1 2 3 4