How to Measure Customer Satisfaction? KPIs, Methods, and Tips

Every company knows the importance of customer satisfaction. But they do not necessarily know how to measure customer satisfaction itself.

This is the heart of the matter. There are many quantitative and qualitative indicators - some of which have already been discussed on Salesdorado: attrition rates, NPS, customer reviews, etc. - so that sometimes you don't know where to start.

In this article, we won't insult you by explaining why customer satisfaction is important. But we will give you a brief explanation of what customer satisfaction really is and how to calculate it, using both quantitative and qualitative KPIs.

How to define customer satisfaction?Image Satisfaction

The problem is usually to measure what is not said. A very satisfied customer will readily say so, and a very dissatisfied customer will say so even more readily. But how do you measure the satisfaction of the majority of customers?

To measure customer satisfaction you have to measure, and that requires interaction with the customer. But there are many situations in which a satisfied customer is one with whom you never speak.

This introduces a huge bias in the measurement, since we only measure the effectiveness of interactions with customers, so we only measure the extremes, and we remain quite far from the reality on the customer side.

This is very common in software, but also in service: in B2B, we often call on the outside to solve a problem, and especially to avoid having to deal with it.

If we focus on monetary and financial metrics, we can approach the notion of customer satisfaction by the difference between the perceived value of your product or service and the cost.

Raised Prices

It is therefore easy to understand the very direct economic interest in increasing customer satisfaction by all possible means: improving the product, offering more services. The complexity is that we are talking about perceived value.

And even if we can hold on to objective elements, especially in B2B, perception is unique to each individual, and it only takes one painful experience to damage it permanently.

The big issue in measuring customer satisfaction is therefore to try to estimate the value perceived by the majority of customers - even if you are allowed to segment.

We'll spare you the paragraph on why we measure customer satisfaction, and get to the hard part right away 👇

6 quantitative KPIs to measure customer satisfaction

1. The Net Promoter Score

The Net Promoter Score or NPS is one of the most popular measures used to assess customer satisfaction. Basically, it is a measure of customer loyalty that is obtained by asking customers the following question:

"How likely are you to recommend our company/product/service to others on a scale of 0-10?"

  • How does the NPS work in practice?

To calculate the NPS, you can share an NPS survey that allows customers to choose a number on a 0-10 rating scale. Once you have collected all the responses, you divide them into promoters, passives and detractors.

  • Promoters (score of 9 and 10): These are your most loyal customers. They act as brand advocates and contribute to the growth of your business by persuading their friends and family to buy from you.
  • Detractors (score 0-6): These are the people least likely to recommend your brand to others. Not only are they difficult to retain, but they may even drive away potential customers.
  • Passives (score of 7 or 8): They generally take a neutral stance. They are not as enthusiastic as promoters and are also unlikely to damage your brand's reputation through negative word-of-mouth. However, with great effort, you can turn passives into active promoters.

Nps Example

Source: ProprofDesk

We give you the NPS formula:

Net Promoter Score = %Promoters - %Detractors

  • The first advantage is that it's very simple to set up, whether you're in software, service, or product sales, the idea is to send a very simple form - which can be answered with one click.
  • It is also a powerful segmentation lever and it is very actionable. You can do a lot to develop the promoters, work on winning over the passives, and reconcile with the detractors.
  • The problem is that like any feedback form, you will have low response rates and a good proportion of your customers will never respond. Thinking that the sample that responds is representative of the rest of your customers is highly questionable. That's why the measure ignores the passive ones and basically assumes that those who don't respond are passive. But this is a real assumption, difficult to verify.
  • You always have to ask the question at exactly the same time. If you want a measurement that has any meaning at all, you have to make the sample as homogeneous as possible. This means asking the question at the same point in their experience with your product or service for all your customers. But how do you define this moment? We often see X days after registration, or payment, or delivery, etc. But depending on the event you choose, you introduce an important bias. A paying customer is a priori more satisfied than a customer who has never made a purchase. And depending on the timeframe too.

Attention

The time limit is very important: wait too long and you will only be interviewing customers who are satisfied (they would have left otherwise). It's nice but not very useful. Wait too little and you only interview customers who want to please you, or who have not yet found value.

  • The other consequence of the device is that you can only take a snapshot of customer satisfaction at a given moment. At the beginning of the experience it makes a lot of sense, it allows you to measure the quality of your acquisition and onboarding. But the NPS quickly becomes very difficult to use.
  • Since the score depends essentially on when you ask the question in the customer experience cycle, it is impossible to compare it from one company to another. Those who talk about NPS benchmarks are charlatans. Try it out: delay the NPS questionnaire by 2 weeks and you will gain 20 points.

The NPS is a great way to measure and track the satisfaction of new customer cohorts over time. But like any measure, it should be taken for what it is.

2. CSAT: Customer Satisfaction Score / CServiceSatisfaction

It is a very popular measure: think of the happy and unhappy men in airport toilets. It is distinguished from NPS because it seeks to measure customer satisfaction with a very specific action.

Example Csat

Source: ProprofDesk

The most typical case is an interaction with customer service: for example, a customer may find it easier to contact the telephone support team than to find a relevant help article in a poorly structured knowledge base. Most customer support software has the CSAT questionnaire and measurement functionality built in.

There is no doubt that successful brands score highly on customer satisfaction. CSAT allows you to collect customer feedback on your products and services and therefore to correct exactly where you need to improve.

  • How does CSAT work?

CSAT surveys usually give customers the opportunity to rate their experience on a scale of 1 to 5. To calculate the CSAT, we only take into account responses of 4 (Satisfied) and 5 (Very Satisfied).

CSAT = [Number of positive responses / Total number of responses] x 100

Let's assume that out of 10 respondents, 8 customers answered 4 or 5.

CSAT score = 8/10 X 100 = 80%.

To remember

A CSAT score of 75% and above is a good indicator of success, but it varies by sector.

3. Customer Health Score

CHS is a popular measure used by customer success teams to determine whether customers are healthy (likely to stay) or at risk (likely to churn). It helps you to carefully segment your audience and focus attention on at-risk customers.Chs ImageCustomer success teams are often faced with a dilemma: either they spend more time on customers who are going to churn anyway, or they spend less time on high-value customers who demand a consistent quality experience.

  • How does the CHS work?

It depends a little on the objective of your business: for example, you may want to monitor the churn rate of your customers, or on the contrary the success of yourupsell strategies.

Therefore, you can use different drivers to build the score. At Salesdorado, we recommend you to use the RFM concept (Recency, Frequency, Monetary Value), which is very classical in relationship marketing, but whose basic idea is also applicable to CHS.

  • Recency: When was the last time your customer used the product/service (or better still, when did they derive value from the product or service if you can measure it)?
  • Frequency: of use, of value generation, of exchange, etc. In the service sector we are quite aware of the frequency of interactions with a customer
  • Monetary value: You can think of it as the value of a customer (LTV) but to apply it to CHS, it would be the monetary value of your product or service.

For example, if you sell clicks to an advertiser and you generate a lot of customers for that customer, the customer health will clearly be much higher than when you sell clicks to a customer who gets no customers. One might be tempted to think that this is just a ratio - as long as the customer acquisition costs are good, the advertiser will be happy. But in reality volumes (in addition to acquisition costs in line with targets) are essential.

To remember

To get an overview of your clients' health, you can include as many elements as you like.

4. Churn

We talked a lot about it in our article on churn rate. The churn rate, also known as customer attrition, is the percentage of customers who stop doing business with your company.

Let's be clear: churn is inevitable. The purpose of the attrition rate is to see if it has reached an alarming level for your company.

  • How does the churn rate work?

Churn Rate

To calculate the churn rate for a given period, you take the number of users at the beginning of the period as well as the number of users at the end of the period.

A simple example: you start the month of April with 300 users and by the end of the month you have 150 users.

(300-150)/300 = 50% churn rate.

  • In most cases, it is too late to act, so it is not a very operational indicator - although you should obviously at least try to winback.
  • A customer base is usually very heterogeneous. Churn is the least frequent signal of customer dissatisfaction - it usually occurs after a number of negative interactions. And because it is the least frequent signal, you have little volume of signals. And so you can mechanically segment less well.
  • With a very heterogeneous customer base, being satisfied with a low overall churn when you are bleeding into a segment with high acquisition costs and high churn can be catastrophic. And conversely, panicking about a high overall churn due to a failed acquisition campaign can lead to very bad interpretations.
  • In addition, it is very often very badly calculated, hence our proposals to optimise the calculation of the attrition rate.

5. First response time

First Response Time (FRT) refers to the time (minutes, hours or days) it takes your team to respond to a customer problem or request.

The link between this measure and customer satisfaction is not very complicated: today, a customer expects a brand to respond as quickly as possible, hence the interest in reducing your average first response time, on all channels, to meet customer expectations and improve their satisfaction.

  • How does the RTF work?

It is quite easy to calculate when your team receives few customer enquiries.

Average FRT = [Sum of total time to first response in a given period ] / Total number of tickets resolved in that period

For example, suppose your team resolved 500 tickets during the month of May and the total time to first response was 20,000 seconds.

FRT = 20,000 seconds ÷ 500 tickets resolved = 4000 seconds or 66.66 minutes.

Attention

Now, things can get complicated when your team handles thousands of tickets every month. In such a scenario, ticketing software like Zendesk can help you automate this process and view your team's average first response time with a single click.

If there are two important things to remember here, it is not so much the KPI as what it highlights:

  • You have to reply quickly - It seems obvious, but it is very common to read emails that are two weeks late, in which they explain that they wanted to check something before replying.
  • If you can't answer right away: say that you are on the case, show that you understand the problem, explain why it takes longer to resolve the situation, and explain how you will do it. Over-communicating in complicated situations never fails

6. First Contact Resolution

But the problem with the FRT is that it often encourages customer support teams to respond with poorly designed templates that do not advance the situation at all.

At least we responded quickly.

But put yourself in the customer's shoes: if you need to contact support, you want a full answer, right away.

If you have to contact the company several times to get a simple solution to their problem, chances are you won't go back to that company again.

You can't solve all the problems in one exchange, but making it a KPI helps to combat this bias.

For phone calls or live chats, first contact resolution means that the customer's issue is resolved before they hang up the phone or end the chat session.

For example, to improve this rate, you can give your agents enough knowledge and authority to solve problems on their own without having to refer to their superiors.

  • How does the RCF work?

Calculating your RCF is simple if you know the number of tickets you resolved in the first interaction. First-Contact-Resolution-How-To-Calculate-Rate

Another example: suppose your team received 8500 customer tickets in one month and resolved 4200 of them on the first contact. In this case, FCR = 4200/8500 X 100 = 49.4%.

Qualitative measures of customer satisfaction

1. Price elasticity

The best way to measure the perceived value of your offer is to see how much your customers are willing to pay.

Not how much they tell you they are willing to pay.

Hence the measurement of price elasticity: it measures the responsiveness of the quantity demanded or supplied of a good to a change in its price. It can be described as elastic, when consumers react to price changes, or inelastic, when consumers react less to price changes.

The question to ask yourself: if you double your prices, how many of your customers will leave, and in how long?

Price-Elasticity

Source: Four Weeks MBA

2. Customer reviews

Image Customer ReviewsYou don't need to know anything about it: when we talk about customer reviews, we are referring to the subjective comments of a customer who has bought and used your product or service. Basically, they can be found in reviews of online shops, websites and other online services. customer review platformsproduct comparison sites, social networks, etc.

Even so, customer reviews are super important and can determine the success or failure of your brand's offerings. In fact, the majority of customers will actively browse hundreds of reviews before buying.

Going further

If you still need convincing, we suggest our article on How to get more good customer reviews (and why)

Pro tip

We recommend reviewflowz.com which is a free solution to aggregate all your customer reviews.

Reader Interactions

Comments

  1. Will says

    I would also include usage metrics, not just declarative. For example the stickiness of the product, in time spent per day, week or month, the frequency of use, the volume of interactions with the product / service, and even potentially some less direct / more complicated elements: how many times or clicks to go from one state to another, or to use such or such functionality.
    From my point of view, declarative data is good, but it should always be taken with care because what users say is often different from their behavior.

    • Axel says

      Hello Will, thanks for your comment. It's true, what you call declarative is what we define as bias of extremes in the introduction, and that is clearly a problem.

      However, approaching perceived value through things "that we know better than the customers" - that seems to me to be an extremely dangerous slope. It's the best way to build a product that nobody wants, in a way that nobody wants.

      How many times or clicks to go from one state to another, it allows you to work on your product or your operations, and potentially generate more value for a certain segment of customers: those who do this action very often, AND who want to do it quickly. It's extremely reductive to reduce customer satisfaction to these kinds of metrics in my opinion, and the bias is huge.

      As for the frequency of use or the time spent, one could easily say the same thing. It is actually quite contradictory with your second point. Is it better to spend a lot of time on a software to perform an action, or little time?

      Basically I think what you meant was that you need to measure the value actually generated by the product or service. For example, the number of high value-added actions performed by a customer over a period of time. This is a classic method in SaaS, by the way. But that too poses a lot of problems. Is the value actually realized perceived? And is it perceived by the right person (the one who pays)?

      The only thing that really affects customer satisfaction is perceived value. For that to happen, you obviously have to generate value, but anyone can do that.

      The challenge, in my opinion, is to understand under what condition(s) the value generated is perceived by the right person.

Leave a comment

Your e-mail address will not be published.