Customer Churn Analysis | Who Left, and Why?

Acquiring a new customer is almost always more expensive than keeping an existing one. Yet most businesses never look at the customers they’ve lost — they just try to find new ones. But the data from the ones who left is the most valuable clue you have for spotting who’s about to leave next.

In a churn analysis, I extract the common traits of your past customers who left: at what stage they dropped off, which product they never bought, how long they’d gone quiet. Once that pattern is found, we can spot early which of your current customers are showing the same signs.

How do we define churn?

The first thing to nail down before starting is the definition of “lost.” In a subscription business this is clear-cut (they canceled); but for a regular retail or e-commerce customer, “lost” is a threshold you have to set — for example, no purchase in the last 90 days.

We set that threshold together, based on your industry and purchase frequency. The wrong threshold means either false alarms or a real loss noticed too late.

Finding common traits

Looking at the data on customers who left usually reveals a recurring pattern: people who never came back after a specific product, people who never placed a second order after their first, people who opened a support ticket and were left waiting for a response. These patterns are hard to spot by eye, but stand out clearly once the data is examined in bulk.

The goal isn’t to find someone to blame, it’s to find the point of intervention. For example, if the finding is “customers who never place a second order are the ones we lose,” the fix is focusing on the period right after that first purchase.

Risk score and retention plan

Based on the pattern found, we score your still-active customers by risk level: high risk, medium risk, low risk. This list shows you who to reach out to first — instead of trying to reach everyone at once.

For each risk group, I put together a concrete retention suggestion: a personal call, a special offer, or just a short survey asking “why might they have left.” You or your team handle the execution — I provide the plan and the reasoning.

How we start: a free 20-30 minute initial call first, where we talk through what you want and what’s realistic. If it makes sense, I send scope and price in writing, then we begin. Reach me via the contact page.

Frequently Asked Questions

I don’t have a subscription model, I run regular sales. Does churn analysis still apply?

Yes. For non-subscription businesses, the “lost” definition is set a bit more loosely — for example, a customer who normally buys every 45 days going quiet for 120 days could count as churn. We set the threshold together based on your own purchase cycle.

Can a lost customer actually be won back?

Some, yes. Customers lost to forgetting or price competition, rather than dissatisfaction, can often be won back. For customers who left unhappy with the product or service, you need to fix the root cause first, or the win-back effort just goes to waste.

Do I need a CRM for this analysis?

Not required, but it helps. A simple sales spreadsheet or e-commerce order history is enough to start. If you have a CRM, we can include customer interaction history too, for a richer analysis.

How often should this be updated?

For most businesses, re-running it every 3 months is enough. If you have a large customer base and change fast, monthly tracking is more accurate. After the first run, I’ll also show you how to update the risk list on your own.