We estimate the average customer who leaves a fintech, insurance or health provider takes about $826 in revenue with them that could have been kept. For a business that loses 10,000 customers a year, that is about $8.3m left on the table. Most of these customers took weeks to go and showed it along the way. 87% heard nothing from their provider before they decided.

The findings below come from our new report, Ghosted, a survey of 1,000 US adults who left a fintech, insurance or health provider in the last year.

Finding 1: We estimate the average customer who left took $826 in revenue that could have been kept

We asked 1,000 US adults who had left a fintech, insurance or health provider what they paid, what would have kept them, how likely that was to work and how much longer they would have stayed. We only counted customers a provider could have caught in time: they took a week or longer to leave, did something the provider could see and weren’t contacted before they decided.

Across the 765 respondents who paid something, that comes to about $826 in revenue per customer who left. For a business that loses 10,000 customers a year, it would be about $8.3m. Treat it as an estimate: it uses the high end of each range, with $500 or more a month counted as $500 and more than 5 years counted as 5 years. Most customers contribute nothing to it, and a small group of high spenders who say they would have stayed for years make up most of the total. A business’s own revenue per customer gives a better figure.

In their own words:

Finding 2: Most customers take weeks to leave

We asked 1,000 US adults who had left a fintech, insurance or health provider how long it was between first thinking about leaving and leaving. Only 7% left the same day. 72% took a week or longer. 40% took more than a month.

Base: all respondents (n=1,000). 2% weren’t sure.

Finding 3: They show it

79% did something their provider could see before they left. The most common signals were quiet ones: using the product less, asking about fees or how to cancel, and moving money out.

Base: all respondents (n=1,000). Respondents could pick more than one.

Customers who got in touch about a problem often had to keep trying. 56% of them contacted their provider more than once about the same problem. 30% had to explain their situation again each time.

Finding 4: Almost none heard from their provider in time

12% of customers heard from their provider before they decided to leave. 23% heard from them only after they had decided. 64% never heard from them at all.

Base: all respondents (n=1,000). 1% weren’t sure.

The pattern holds across categories: 85% of fintech customers, 92% of insurance customers and 80% of health customers weren’t contacted before they decided. It shows in how customers feel. 84% say companies only pay attention to them when they try to leave.

Finding 5: What a provider says when it gets in touch matters

Among customers who were still thinking about leaving or looking at other companies, the kind of contact made a difference.

Base: people who seriously thought about leaving and were still only thinking about it or looking around when they made their final choice.

Fixing a problem was the only kind of contact with a higher stay rate than no contact. Trying to sell something else had the lowest. When we asked what would have kept them, fixing the problem the first time was also the top answer (19%), ahead of matching a competitor’s price (11%). In their own words:

The bases here are small, so treat the gaps as directional. Getting in touch on its own didn’t make the difference: compared at the same stage of their decision, people who were contacted stayed at about the same rate as people who weren’t.

What this means for businesses

The money is in the weeks before a cancellation. By our estimate, a business losing 10,000 customers a year leaves about $8.3m on the table.

Your customers tell you before they leave. Most spend weeks deciding, and most leave a signal along the way: lower usage, a question about cancelling, money moving out, the same problem raised twice.

You probably aren’t acting on it yet. 87% of customers heard nothing before they decided. By the time a cancellation request arrives, most have already made up their mind.

The fix matters more than the pitch. Customers who had a problem fixed stayed at the highest rate. Customers who were sold something else stayed at the lowest.

Lorikeet Dreamweaver

This is the gap we built Lorikeet Dreamweaver for. It increases customer lifetime value by giving every customer their own account manager. A Dreamweaver does the four things the data points to.

• You set the goal. Retention, a renewed policy, a refilled prescription. Each customer gets a Dreamweaver working toward it over weeks and months. You don’t build a segment, a trigger or a sequence.

• It notices and reaches out before they decide. It wakes when something happens, like a change in product usage or a support request, looks at everything about that customer and decides the best next step. It reaches out only when it helps.

• It starts with the customer’s problem. It’s the same Concierge that already handles your support, so it can solve the issue the customer raised, with the same guardrails, consent and contact rules.

• It remembers the whole person. What they said last time, when they reply, which channel they use. 30% of customers who got in touch about a problem had to explain their situation again every time. With a Dreamweaver, they don’t. And it knows when to stop: it drops a customer the moment the goal is met.

Read the full report: Ghosted covers all seven findings, including the revenue providers could have kept.

Try Dreamweaver: the first 100 sign-ups get $1,000 of free credits.

How we ran it

Lorikeet surveyed 1,000 US adults who had chosen to leave or not renew a fintech, insurance or health product in the last 12 months. The survey carried no brand and did not mention AI. Results are unweighted. The $826 figure is an estimate based on the data collected.