Churn Prediction: 3 Months Without Purchase = 85% Won't Return (Reactivate Now)
- How Churn Prediction Works
- Early Warning Indicators
- Reactivation Strategy Effectiveness
- AskBiz Churn Alerting
- Building a Simple Churn Score Without a Data Scientist
- Worked Example: The SGD 2,400 Weekly Reactivation Run
- Common Mistakes That Waste Reactivation Budget
- When Churn Prediction Should Change Your Business, Not Just Your Emails
Retail: customer last purchase Jan, now April (3 months). Historical data: 85% of customers inactive 3 months never return (churn). Send reactivation email April (before 3-month mark): 25% open rate, 5-10% click through, 2-3% purchase rate. Reactivation value: SGD 100 order × 2-3% = SGD 2-3 revenue per reactivation attempt. Cost: email SGD 0.01. ROI: 200-300x. Send 1000 at-risk emails = SGD 2-3K recovered revenue.
- How Churn Prediction Works
- Early Warning Indicators
- Reactivation Strategy Effectiveness
- AskBiz Churn Alerting
- Building a Simple Churn Score Without a Data Scientist
How Churn Prediction Works#
Analyze historical customer data: when did they stop buying? If inactive 3 months = 85% never return. If inactive 2 months = 40% return. If inactive 1 month = 10% never return. Build rule: if customer inactive 2 months, flag as "at risk." Trigger reactivation campaign.
Early Warning Indicators#
(1) Purchase frequency drop: bought monthly, now 2-month gaps. (2) Order value drop: avg SGD 100, last order SGD 30 (lower engagement). (3) Email engagement: stopped opening emails. (4) Support tickets: complaints increase before churn. Track all signals = earlier intervention.
Reactivation Strategy Effectiveness#
Email at 1-month inactive: 5-10% respond, 2-3% convert. Email at 2-month inactive: 3-5% respond, 1-2% convert. Email at 3-month inactive: 1-2% respond, 0.5-1% convert. Best window: 2-month mark (still fresh enough to reactivate, not yet written off). Offer: 10-20% discount SGD 100 order = 0% margin at that point, but recovers lifetime value (if reactivation works).
AskBiz Churn Alerting#
Predicts churn risk. "20 customers now 2 months inactive: predicted churn 40%. Reactivation campaign recommended: 10% discount on next order. Send emails today. Expected: 3-5 conversions (SGD 300-500 recovery). Monitor: if conversion rate <1%, messaging is off (try different offer/subject)."
Building a Simple Churn Score Without a Data Scientist#
You don't need machine learning to get 80% of the value. Pull your order history and calculate, for every customer, days since last purchase. Bucket customers into three groups: active (0-30 days), at-risk (31-60 days), and likely-churned (61-90+ days). Then look backward: of everyone who was ever in the 31-60 day bucket last year, what percentage bought again within the next 30 days versus never returned? That historical percentage is your churn probability for the bucket, and it's the number you plug into email targeting. The mechanics are just: recency bucket → historical conversion rate for that bucket → expected value of intervention (conversion rate × average order value − campaign cost). Refresh the buckets weekly. This spreadsheet-level model is what most SMBs should start with before paying for anything more sophisticated.
Worked Example: The SGD 2,400 Weekly Reactivation Run#
A Singapore homeware retailer with 6,000 active customers found 240 fell into the 2-month at-risk bucket in a given week. Historical data showed this bucket converts at 2% within 30 days of a reactivation email with a 15% discount, at an average order value of SGD 100. Expected recovered revenue: 240 × 2% × SGD 100 = SGD 480 in direct reactivation revenue, at a campaign cost of roughly SGD 5 (email platform + discount margin absorbed into order value). Run this every week for a year and the compounding effect on retained customers, not just the one-off order value, pushed total annual recovered revenue to an estimated SGD 2,400 — because a meaningful share of reactivated customers went on to buy a second and third time without further discounting. The lesson: track not just the immediate conversion but whether reactivated customers stay active afterward.
Common Mistakes That Waste Reactivation Budget#
The most common mistake is waiting too long — sending the first touch at 3 months instead of 2, when response rates have already fallen from 3-5% to 1-2%. The second is leading with a discount every time, which trains your best customers to wait for markdowns before repurchasing; reserve discounts for the second touch and lead the first with a simple "we miss you" or new-arrivals email. The third is treating all at-risk customers identically — a customer who historically spent SGD 500/year deserves a different, higher-value intervention (a phone call, a personalised offer) than one who spent SGD 40 once. AskBiz's churn alerting segments at-risk customers by historical lifetime value alongside inactivity, so your highest-value at-risk customers surface at the top of the list rather than getting the same generic 10% code as everyone else.
When Churn Prediction Should Change Your Business, Not Just Your Emails#
If your at-risk bucket is growing month over month even after reactivation campaigns, the problem usually isn't the email — it's the product, price, or service experience driving people away in the first place. Cross-reference churned customers against support tickets, return rates, and order value trends before their last purchase. A spike in complaints or a drop in order value in the two months before someone goes inactive is a leading indicator, not just a lagging one — and it points at a fixable root cause (a stockout on their usual item, a price increase, a bad delivery experience) rather than a reactivation-messaging problem. Treating a systemic churn increase as an email-copy problem wastes budget on symptoms while the underlying cause keeps generating new churn every month.
People also ask
What discount should I offer?
10-20% for first reactivation. If first doesn't work, try 30-40% (more aggressive). Balance: don't want to train customer to expect discounts only.
How often should I reactivate the same customer?
Once per 3-month period. If customer churns twice despite reactivation, stop (not recoverable). Focus on preventing churn in first place (better value).
Our team combines expertise in data analytics, SME strategy, and AI tools to produce practical guides that help founders and operators make better business decisions.
Predict & Prevent Churn (Recover 20-30% At-Risk Customers)
AskBiz predicts churn at 1/2/3 month inactivity milestones. Auto-triggers reactivation campaigns. Tracks recovery rate. Try free.
Connects to Shopify, Xero, Amazon, QuickBooks, Stripe & more in minutes