In the search for the perfect customer feedback metric there is an ongoing arms race. It started with the venerable Customer Satisfaction. Then came Net Promoter®. More recently we have seen Customer Effort Score, and another new word of mouth metric recently released. What about Net Emotional Value (NEV) or Net Value Score?
Nobody wants to be seen dead in last year’s customer feedback metric (oh the embarrassment, I could just die) but with so much choice it naturally leads to the question: which customer feedback metric is the best?
(Spoiler alert: Net Promoter Score® comes out ahead of pack but read on for the reasons.)
Why do we need a metric in the first place?
The first question you really need to ask yourself is: Why do I even need a customer feedback metric? Well there are three reasons:
1. Proxy for Customer Loyalty
Most of the time customer feedback is about understanding what you need to change to improve customer loyalty, and referral propensity. So you need an index, proxy or measure for the loyalty of each respondent.
2. Benchmark Progress
As you continue to improve your customer experience you need a measure against which you can compare yourself. How much better, or worse, are we than last quarter? You might also want to benchmark yourself against other companies.
3. Dependant Variable
In statistical terms you also need a useful dependent variable that you can use to determine how each of your independent variables (service and product attributes to you and me) impacts on customer loyalty and how important is each one.
The Options
At the moment there are really four common options, plus a couple of new players on the block, for the best customer feedback metric. Let’s look at each of them in turn.
Customer Satisfaction
This is the maiden Aunt of Customer Feedback Metrics and it’s been around for many, many years and is simple to use. If you’ve seen it once you seen it a million times. Kindly, familiar and always ready to fill its place in the survey question line up.
It is pretty simple to understand:
How satisfied were you with your recent product?
The key downside of Customer Satisfaction is its lack of predictive power. Generally satisfied customers don’t often behave much differently to generally dissatisfied customers. That begs the question; why ask if what you are getting is not useful?
Basically it does not perform task 1. Proxy for Customer Loyalty very well.
Net Promoter Score(R) or NPS
This was officially launched into the public realm in late 2003 in a Harvard Business Review. Not happy with the predictive power of Customer Satisfaction, Frederick Reichheld, Bain and Satmetrix tested a range of questions with the goal of identifying a question that would predict future actual future purchasing patterns. The result was Net Promoter Score.
How likely would you be to recommend our company to a friend or colleague on a scale of 0 – Very Unlikely to 10 Very Likely.
By subtracting the percentage of 0-6 scores from the percentage of 9 and 10 scores you arrive at the NPS. This was shown to be more highly correlated, more often, with future revenue growth.
Download a more complete introduction to the Net Promoter methodology here.
Customer Effort Score
Not satisfied that the NPS was as good as it could get, a competitor to the companies that created NPS did some research and identified another question that they claimed also provide a good proxy for real customer loyalty:
How much effort did you personally have to put forth to handle your request on a scale from 1 (very low effort) to 5 (very high effort)
Once again this question was shown by the authors to be a good proxy for loyalty, potentially as good as NPS.
For more information on this question see this Harvard Business Review Article
Statistically Accurate Scores
Here, I’m lumping the very many company or industry specific multi-part scores. Typically these scores use advanced statistical techniques to look at the response from two or more questions and create a composite score:
10% of Customer Satisfaction + 23.6% of would buy again for the first time + ….
Implemented correctly these scores can be more accurate than any of the other scores. They do however require advanced statistical skills to implement and can be difficult for staff to understand and use.
Flavour of the month question
Almost every month a new question is launched: Net Emotional Value (NEV), Net Value Score, WoMI (Word of Mouth Index). Each of these vies to be an even better predictor of loyalty. Each of the developers is looking for a silver bullet in the science of finding the perfect customer feedback metric.
I am not going to go into details for each of these flavour of the month metrics mostly because they change so frequently they tend to violate the “general acceptance” and “body of support material” tests.
Criteria for Selection
So which is the best metric?
That really depends on what you are looking to achieve. Some people will suggest that you should use the score with the highest accuracy. That is not necessarily the case as there are many elements of the score that will influence how easy it is to succeed in implementation.
Comparing the Options
When comparing the different option what attributes are important. From our experience the following attributes are the most important in a customer feedback metric.
Predictive Power: is the score an effective predictor of customer loyalty?
Ease of Understanding by All Staff: Can it be explained to and understood by all staff quickly and easily?
Ease of Use: Can the score be used to drive change in the organisation?
Ease of Calculation: can the score be calculated quickly and easily?
General Acceptance: Is the score generally well accepted in the business community?
Body of Support Material: Is there a good body of material in the public domain on which organisations can build their implementation approaches?
Ease of Loyalty Driver Analysis: How easy is it to perform loyalty driver analysis?
Comparing the Options
Let’s score each of the metrics. Yes, this is a subjective rating but I’m happy to debate these scores with anyone.
Leave a comment if you disagree with a rating here and I’m happy to discuss.
|
Metric
|
CSAT
|
NPS
|
CES
|
Statistical Score
|
Flavour of the Month
|
|---|---|---|---|---|---|
| Predictive Power | Low | Good | Good | Best | Generally Un-proven |
| Ease of Understanding by all staff | Good | Good | Good | Low | Varies |
| Ease of Use | Good | Good | Medium | Good | Varies |
| Ease of Calculation | Good | Good | Good | Low | Varies |
| General acceptance | Good | Good | Increasing | Low | Low |
| Body of support material | Good | Good | Low | Good | Very low |
| Ease of loyalty driver analysis | Good | Good | Low | Good | Varies |
The Winner Is: Net Promoter Score
NPS is not the most accurate predictor of customer loyalty but it is the overall most effective approach. Overall it is as easy to user as customer satisfaction while also being more accurate.
Conclusions
Net Promoter Sore is not perfect by any means.
Some researchers argue that the “net” approach reduces the information in the score, others highlight that it is not as accurate a predictor as could be achieved, and still others say that it is not a good predictor of word of mouth promotion. On the other hand some others say the net value makes the score more predictive.
All this may be true but the most import attribute of any customer feedback indicator is that it can be used by the business to improve and become more customer focused. NPS has proven that it can do this better than the alternatives.
Do you disagree with any of my scores or selection criteria? Let me know in the comments below.
I take issue with the premise of the article – that there is one metric that companies should adopt and their biggest choice is which one.
In particular, NPS and Customer Effort Score are measuring two different things. Propensity to refer may incorporate a customer’s perception of the effort required to do business with the company, but also a lot more – subliminal brand feelings in particular.
Customer Effort Score is therefore more limited, but has two key benefits that are not discussed in the article. One, it can be measured more or less objectively. It’s possible to find high-effort points in your interactions without even asking your customers.
Second, it’s highly actionable. Once you’ve identified these high effort areas, you can make specific efforts to fix them. Reducing customer effort is not a vague goal – it can be pursued systematically.
In discussions I’ve had with customer experience folks, the consensus that emerges is that BOTH NPS and CES are valuable metrics. CES is best at measuring the quality of a specific interaction. NPS is better at assessing overall perception, and incorporating those subjective judgments that all customers carry, whether they’re aware of it or not.
…customer behavior, or the downstream effect of customer experience, on a granular basis:
http://www.customerthink.com/blog/customer_effort_score_and_nps_gangnam_style_metrics
http://www.customerthink.com/blog/sorry_nps_i_m_not_buying_it
http://www.customerthink.com/blog/its_time_to_get_really_serious_about_performance_metrics_what_works_what_has_challenges_and_why
John,
Thanks for taking the time to comment.
I agree that CES provides a subset of those influencers that can be identified by NPS.
It is also true that you can use more than one metric in the business. One of my concerns is that most organisations have enough trouble getting one metric implemented properly. Until the first is implemented correctly there is not much point adding more. So if you are only going to use one, which should it be?
However, I tend to disagree with the other points you make.
CES is objective: I assume that mean that some CES related elements of the customer experience can be measure objectively, e.g. hold time or first call resolution. This is true but the customer's response to the CES question is subjective. Two different customers going through the same experience can easily respond in different ways. So while some of the drivers of CES may be measured objectively CES is intrinsically subjective.
CES is more actionable: it is our experience when using CES that it is harder to build the driver analysis than when using NPS. You can certainly use continuous improvement processes to drive change. We have just found it easier to identify the root causes when using NPS than CES.
Michael,
It would be nice if you summarized the arguments for your case in the comments rather than directing people to a range of posts. Without that summary it is difficult to identify the basis for your statement.
However, you will note that I never suggested that NPS helps you to understand customer behavior. I stated that it was a usefully good proxy for customer loyalty behavior, i.e. it is a good measure of customer behavior. This I think is relatively well proven.
In order to understand customer behavior you have to combine NPS (or CES for that matter) with a range of other feedback and information, using NPS as the dependent variable.