What are the churn warning signs?

Declining order frequency, shrinking order sizes, and longer gaps between purchases are the three biggest predictors of churn.

Who churned this month?

Your monthly report lists every customer who was active last month but had no transactions this month, with their full history.

What's my monthly churn rate?

Bottomline calculates your churn rate automatically by comparing active customers at the start and end of each month.

Are my repeat customers actually coming back?

Bottomline tracks repeat purchase rates and flags customers whose buying frequency is declining.

How much is a customer worth over their full relationship?

Bottomline calculates customer lifetime value from your transaction history so you know what each customer relationship is actually worth.

Were there warning signs before they churned?

Almost always, yes. Customers rarely leave without leaving a trail. Declining order sizes, longer gaps between purchases, reduced scope. Here's how to look back at your data and spot the signals you missed.

7 min read

The short answer

In most cases, yes. Research shows that 70-80% of customer churn is preceded by measurable behavior changes in the 2-4 months before departure. The three most reliable signals are declining order frequency, shrinking order size, and increasing time between purchases. Your transaction history has these patterns. You just have to look.

Why finding the warning signs retroactively still matters

You already lost the customer. Why bother looking back? Because the same patterns are likely happening right now with other customers who have not left yet.

If you discover that Riverside Dental's order frequency dropped from monthly to every 6 weeks to every 8 weeks before they finally stopped, you can scan your current customer base for anyone showing the same pattern. A phone call to a customer whose frequency is slipping costs you 20 minutes. Replacing them after they leave costs you months of sales effort.

The goal is not to do a post-mortem for its own sake. The goal is to build a playbook: these are the warning signs, this is how early they appear, and this is the intervention that works.

The three behavioral patterns that predict customer churn

These are the signals that show up most consistently in transaction data before a customer leaves:

1. Declining order frequency. A customer who used to buy monthly is now buying every 6 weeks, then every 2 months.

2. Shrinking order size. Their average order was $2,800. Now it is $1,900. They are pulling back before they pull out entirely.

3. Increasing gap since last purchase. Their typical gap between orders was 28 days. Their current gap is 52 days and counting.

Any one of these is a yellow flag. Two together is a red flag. All three means the customer is almost certainly on their way out.

How to check for warning signs in QuickBooks Online

1
Go to the churned customer's profile
Search for the customer name. Click into their profile to see their full transaction history.
2
List their transactions for the last 6-12 months
Look at all invoices and sales receipts. Note the date and amount of each one.
3
Calculate the gaps between transactions
In a spreadsheet, list each transaction date and calculate the number of days between consecutive transactions. Is the gap growing?
4
Track the average order size over time
Plot the transaction amounts chronologically. Is there a downward trend in the last 3-4 months compared to the 6 months before that?
5
Compare to their historical baseline
If a customer used to order every 25 days at $2,800 and their last three orders were at 35 days, 42 days, and $1,900, the warning signs were there.

Total time: 15-20 minutes per customer. If you lost 5 customers, this analysis takes over an hour. Multiply that by every month and it becomes unsustainable.

How to check for warning signs in Xero

1
Go to Contacts → find the churned customer
Open their contact record and click into their activity tab to see their full invoice and payment history.
2
Review their invoice history for the last 6-12 months
Note dates and amounts. Export to a spreadsheet if there are many transactions.
3
Run the same gap and size analysis
Calculate gaps between invoices and track average order size over time. Look for the same three patterns: longer gaps, smaller orders, declining frequency.

Total time: 15-20 minutes per customer. Same as QuickBooks. The individual customer review is the bottleneck in both systems.

Why retrospective churn analysis is almost never done manually

15-20 minutes per churned customer. With 4-6 churned customers per month, that is 1-2 hours of manual transaction review.
The real value is proactive, not retrospective. Knowing the warning signs after the fact is educational. Detecting them in real time on current customers is where the money is. But scanning every active customer for declining patterns manually is not feasible.
Pattern detection requires math your accounting tool cannot do. Calculating rolling averages, comparing current behavior to historical baselines, and flagging deviations is statistical work that no standard accounting report provides.

Or let Bottomline detect warning signs before customers leave

Bottomline analyzes the transaction patterns of every customer every month. When a customer's behavior deviates from their baseline (longer gaps, smaller orders, declining frequency), they get flagged as at-risk before they churn.

Churn risk analysis: Riverside Dental

Order frequencyBaseline: Every 28 daysNow: Every 42 days

Average order sizeBaseline: $2,800Now: $1,900

Days since last orderBaseline: 28 day avgNow: 52 days

3 of 3 warning signals triggered. This customer churned the following month.

From a real Bottomline report. Warning signs are detected by comparing current behavior to each customer's historical baseline.

The difference between hindsight and foresight is automation. Running this analysis on every customer every month is impossible manually. Bottomline does it automatically and surfaces only the customers who need attention.