How these pages work
Every diagnosis here begins by testing the number before reading the customers behind it. A churn problem starts by confirming the definition and the base have not changed, then reads churn by cohort rather than as one blended rate; a survey problem starts with the score's own range of noise and a comparison of the people who answered with the people who were asked; a support problem starts with contacts per customer by reason over time; a loyalty problem starts by comparing members with similar non-members over the same period; a complaints problem starts by tagging a month of complaints by root cause; and a win-back problem starts by reading who returned against a holdout. Once the fault is visible, a diagnosis table lists what you can see, the check that confirms each cause and the fix. Each cause then gets a short section with the fix, a way to verify it and the habit that prevents it, followed by a routine that keeps the whole set under watch.
The pages assume the methods are taught elsewhere and link to them. How churn and retention are defined and read by cohort is in the churn and retention metrics guide; relationship and transactional surveys are in the voice of the customer guide; the work of sorting every contact by its contact reasonThe main reason a customer contacted support, tagged by the agent from a fixed, defined list so volume and other service metrics can be read by cause. is in the tagging tutorial; and the friction that makes customers call twice is in the customer effort guide.
Find the page for your symptom
Choose the row that matches what you are seeing. Diagnosis pages in this section are also listed beneath this introduction automatically; the last three rows send problems that belong somewhere else to the page that owns them.
Measurement, experience or service: decide which question you are asking
Nearly every symptom in this section answers one of three questions, and naming the question first saves most of the diagnosis. A measurement question asks whether the number moved before any customer did: a churn rateShare of customers at the start of a period who left during it. whose definition or base changed, recent cohorts hidden inside a blended rate, a survey score moving within its normal noise or answered by different people, a loyalty reading with no comparison group, contacts counted but never tagged. An experience question asks whether customers are getting what they were promised: first value never reached, a marketing promise the product breaks, effort rising somewhere in the journey, rewards for purchases that would have happened anyway, a delay customers hear about only when they chase. A service question asks what happens when something goes wrong: failures never recovered, front line staff with no written remedies, no procedure, replies that vary by agent, a repeat contactA customer contacting the business again about the same problem, the clearest sign that the first answer did not end the customer's work. because the first answer did not resolve the issue, complaints closed without a root cause.
The test is to settle measurement first, because it is the cheapest to check and the most embarrassing to miss. If the definition, the base, the respondents and the timing all held, the movement is real, and the next question is whether it traces to what customers were promised or to what happened when something went wrong. Where the cause is experience, drawing one customer's path from purchase to the point of failure locates it; the churn page does this with Nielsen Norman Group's introduction to journey mapping, which describes a journey map as a visualisation of the process a person goes through to accomplish a goal. Where the cause is service, read the contacts and complaints themselves rather than the score that summarises them.
The questions also chain. Unrecovered failures feed complaints, complaints feed the bottom of the survey scale, and both feed churn, so the same root cause can appear on several of these pages; fix it once, where it starts, and watch the downstream numbers follow.
When to audit instead
A diagnosis page fixes one fault. When several of these symptoms arrive together, when nobody can say why customers leave, or when retention has never been reviewed as a system rather than a campaign, an audit is the better tool. The customer retention audit scorecard rates how churn is measured and why customers leave, through onboarding, service and cancellation, to win-back, loyalty and listening, and ranks the weaknesses that lose the most customers before any new offer is launched; and the customer reviews audit scorecard rates how reviews are earned, answered, shown and passed to the people who can fix what they describe. The measurement section and the service and support section place the guides these pages draw on side by side.
