Almost every company surveys its customers: post-purchase surveys, annual questionnaires, ratings left after an exchange with support. The results are often flattering. Sometimes too flattering.

Because a satisfaction survey only measures what the respondents say. Here are the two biases that skew most satisfaction surveys, and how to organize customer listening that truly reflects what your customers think.

The team behind our flexible, automatable CRM tells you everything.

First bias: those who respond are not the ones who matter

For a measurement to be reliable, the responses must be representative of the customer base. This is rarely the case. Broadly speaking, your customers fall into three groups: the very satisfied, the generally satisfied, and the dissatisfied.

The two extremes respond readily: the first to congratulate you, the last to complain. The middle group, often the largest, rarely responds. It is a silent majority, and yet it is there that many renewals and quiet departures are decided.

Hence a counter-intuitive consequence: a satisfaction survey mainly measures the dissatisfaction of those who want to express it. And a record satisfaction rate calculated on a low response rate does not prove much.

Second bias: not all customers carry the same weight

A survey adds up responses as if they were all equal. They are not. A customer who publishes a review read by hundreds of prospects, who speaks up in your sector’s professional groups or who is consulted by peers before every purchase carries far more weight than another.

Online reviews and professional networks have multiplied this power of influence. A bad experience told publicly by the right person can turn away many potential customers. The topic is developed in our article on customer reviews.

Organizing more reliable customer listening

Ask at the right time

A short question asked right after an interaction, a delivery or a support intervention gets more responses, and more precise ones, than a long annual questionnaire. The memory of the experience is fresh, and the customer sees the link between the question and what they have just experienced.

Reach out to the silent majority

A few phone interviews with customers who never answer surveys are often worth more than hundreds of questionnaires. Choose them from the middle group: regular customers, with no notable incident and no particular enthusiasm. They are the ones who will tell you what could make them leave.

Observe behavior

What customers do often says more than what they declare. Renewal rate, purchase frequency, service usage, the number and nature of support requests: these signals are already in your tools, and they suffer from no response bias.

Identify influential customers

Spot the customers in your portfolio who publish, recommend or are consulted. Their experience deserves special attention, not to grant them special favors, but so that you do not discover their dissatisfaction at the same time as their readers.

Follow up on every response

A survey creates an expectation. A customer who reports a problem and gets no reaction is more dissatisfied after the survey than before. Every negative response should trigger a contact, and every customer surveyed should be able to see, sooner or later, that their comments were put to use.

Linking customer listening to the CRM

All this information is only valuable if it is added to the customer record: survey responses, interviews, complaints, published reviews. Brought together in a single customer view, it makes it possible to cross-check what customers say with what they do, and to alert the right contact at the right time.

Measuring satisfaction is only one step, by the way. Why you should aim beyond it, for trust and engagement, is the subject of our article on moving beyond customer satisfaction.

Your questions about satisfaction surveys

What response rate should you aim for?

It depends heavily on the channel and timing. More important than the rate: knowing who did not respond, and why.

Should the questionnaire be long or short?

Short, almost always. One or two questions and a free-text field get more usable responses than a long form.

Should responses be anonymous?

It depends on the goal. A named response makes it possible to address each case of dissatisfaction; an anonymous response may be more candid.

How often should you survey the same customer?

In moderation. Too many surveys tire customers and lower the quality of subsequent responses.

What should you do about a negative review published online?

Respond publicly and calmly, and deal with the substance privately. Your reply is read by every prospect who reads the review.