Insights / Checklist / Consumer behaviour analysis
ChecklistReading loyalty and churn
Understanding who stays and who leaves guides where to invest. Define what a lost customer means for you, watch the early signals and talk to people who leave.
Winning a new customer usually costs far more than keeping an existing one, yet many businesses know much more about where new customers come from than about why existing ones leave. Churn is reported as a single percentage, if at all, and loyalty is assumed rather than measured.
Understanding who stays and who leaves guides where to invest: in products, service, pricing or simply in contacting the right customers at the right time. This checklist sets out the questions to answer, the signals to watch and the conversations to have.
Define what a lost customer is
For a subscription or contract business, churn is clear: the customer cancels or does not renew. For most others, customers simply stop buying without telling you. You have to decide when silence means they have gone.
Base the definition on normal buying patterns. A customer who usually orders office supplies every month and has not ordered for three months has probably left. A customer who buys a sofa every eight years has not. Look at the typical gap between purchases for each type of customer or product, and set a threshold that marks a clear break from it. Many businesses use two levels: “at risk” when a customer goes noticeably longer than usual without buying, and “lapsed” when the gap is well beyond normal.
Consider partial churn too. A business customer who still orders but has moved half their spending to a competitor is not lost, but is leaving. Tracking share of spend, where you can estimate it, or a fall in order value against the customer’s own history, catches this.
Watch the early signals
Customers rarely leave without warning. The signals are usually in data you already hold, though often spread across systems that nobody looks at together.
- Falling order frequency or value compared with the customer’s own history.
- Narrowing purchases: buying fewer categories or only discounted items.
- Complaints, returns or service problems, particularly unresolved ones.
- Reduced engagement: fewer logins, emails unopened, loyalty points unused.
- Changes in the customer’s own situation, such as a new buyer at a business customer.
- Price checks or requests for quotes on items they normally reorder without asking.
Bring these signals together in a simple view for account managers or customer service. A short weekly list of customers showing two or more warning signs is far more useful than a sophisticated model nobody acts on.
The loyalty and churn checklist
- We have a written definition of a lapsed customer, based on normal buying gaps for each customer type.
- We track churn and retention by segment, product and acquisition channel, not just in total.
- We compare cohorts: customers acquired in each period, and how many remain after six and twelve months.
- We monitor partial churn, such as falling order value or narrowing purchases.
- Early warning signs from orders, service and engagement are combined in one view.
- A named team reviews at-risk customers weekly and decides who to contact.
- We hold exit conversations with a sample of customers who leave.
- Exit reasons are recorded consistently and reviewed monthly.
- Fixable causes are assigned to owners with deadlines.
- We measure whether contacting at-risk customers changes what they do, against a group we did not contact.
Talk to customers who leave
Data shows that customers are leaving. Conversations show why. Contact a sample of lapsed customers, ideally by phone and by someone not responsible for their account, and ask what changed. Keep it short and genuinely curious; the aim is to learn, not to win them back on the call.
Causes are often more fixable than expected: a delivery problem that was never resolved, a change of contact who did not know about the account, a product discontinued without an alternative being offered, a competitor’s offer that could easily have been matched. Record reasons in consistent categories so patterns become visible over months.
Separate fixable from unavoidable
Some customers leave for reasons you cannot influence: they move away, close the business or no longer need the product. Track these separately, so effort goes to the causes you can change.
Act, then measure the effect
Every insight needs an owner and an action. If late deliveries drive churn in one region, the fix sits with operations. If customers leave after a price rise, commercial teams need to see it. If one product range sees repeated defections, buying should know.
When you contact at-risk customers with offers or service calls, hold back a small comparable group and do not contact them. Comparing the two shows whether the effort works, or whether those customers would have stayed or left regardless. Without that comparison, retention activity is easy to overvalue.
Our exit calls kept surfacing the same problem with invoicing. Fixing it did more for retention than any loyalty offer we had tried.
Commercial director, wholesale distribution client
Where to start
This month, agree a definition of a lapsed customer and calculate how many customers lapsed in the last year, by segment. Pick the two or three warning signals easiest to pull from your systems and build a weekly at-risk list. Then call ten recently lapsed customers and write down what they say. That alone will tell you where to focus next.
Source: Intelliversal analysis.
