Data Analysis Tips·5 min read

Customer Retention Analysis: 3 Metrics to Understand Customer Loyalty

Customer acquisition costs keep rising — retaining existing customers is more cost-effective than acquiring new ones. Learn to measure loyalty through repeat rate, purchase frequency, and customer lifetime value, and identify churn signals and retention strategies.

Why Retention Matters More Than Acquisition?

A business axiom: acquiring a new customer costs 5-7x more than retaining an existing one. Yet most effort still goes into 'getting traffic' — promotions, ads, social media — while 'whether they come back' gets little attention. If you add 100 customers monthly but lose 80 existing ones, growth is slow. Reduce churn from 80% to 40% and you've effectively gained 40 free customers. Retention analysis answers: are your customers staying? If not, why? And how can you keep them?

Metric 1: Repeat Rate — How Many Customers Come Back?

The most direct loyalty metric. It answers: of all customers who've purchased, how many purchased more than once?

How to Calculate Repeat Rate

Repeat Rate = Customers with 2+ Purchases ÷ Total Customers × 100%. Example: 500 customers in a month, 150 are returning (previously purchased), repeat rate is 30%. Note the distinction between 'repeat rate' and 'returning customer share': repeat rate is customer-level (% of customers who returned); returning share is transaction-level (% of transactions from returning customers). Track both. If returning customer share is 60%, most revenue comes from loyal customers — good for stability, but risky if they start churning.

What's a Good Repeat Rate?

Varies widely by industry: F&B monthly repeat rate of 20-40% is normal (50%+ is strong), retail 15-30%, coffee/tea beverages 30-50%. Don't just benchmark against industry averages — watch your own trend. Repeat rate declining for 3 consecutive months signals trouble, even if the absolute number is above average. AI tools automatically calculate repeat rate trends and alert on declining patterns.

Metric 2: Purchase Frequency — How Often Do They Visit?

Repeat rate tells you 'did they come back?' Frequency tells you 'how often?' Two repeat customers with different patterns — one visits 3x per week, another 1x per month — have very different values.

Frequency Distribution Matters More Than Average

Don't just look at 'average frequency' — look at the distribution: How many customers purchased once? 2-3 times? 4-6 times? 7+ times? You'll typically find: 60-70% purchased only once, 20-25% purchased 2-3 times, 5-10% purchased 4+ times. That small high-frequency segment contributes disproportionately to revenue — your core customer base. With this distribution, strategy becomes targeted: For one-time customers (60-70%), the goal is a second visit — largest growth opportunity; For 2-3x customers, increase frequency to at least monthly; For high-frequency customers, focus on retention and VIP service to prevent churn.

Purchase Interval Analysis

For repeat customers, analyze average purchase interval — time between purchases. Shortening intervals means increasing stickiness; lengthening intervals means customers are 'fading out.' A more refined approach is cohort analysis: group customers by first purchase month, then track each cohort's monthly repeat rate. For example, 100 customers who first purchased in January: 30 returned in Month 2, 25 in Month 3, 20 in Month 4 — this decay curve is your 'customer churn curve.' If curves flatten for newer cohorts, your product and service are improving.

Metric 3: Customer Lifetime Value (LTV) — How Much Is a Customer Worth?

LTV measures total revenue a customer generates from first to last purchase. It's the ultimate customer value metric.

How to Calculate LTV

Simplified formula: LTV = Average Ticket × Monthly Purchase Frequency × Average Customer Lifespan (months). Example: 50 yuan average ticket, 2 purchases/month, 6-month average lifespan = LTV of 600 yuan. With LTV you can make two critical judgments: First, LTV-to-acquisition-cost ratio. If acquiring a customer costs 100 yuan and LTV is 600, ratio is 6:1 — healthy. Below 3:1 means acquisition strategy needs optimization. Second, compare LTV across customer segments — by channel (Xiaohongshu vs Douyin customers), by first purchase amount, by first purchase category. Focus acquisition on high-LTV segments.

Three Levers to Increase LTV

From the formula, three ways to increase LTV: Increase average ticket (spend more each visit) — through upselling, combo deals, cross-selling; Increase purchase frequency (visit more often) — through membership programs, regular pushes, limited-time offers; Extend customer lifespan (stay longer) — through loyalty programs, personalized service, churn prediction and retention. Of the three, extending lifespan has the biggest impact — each additional month adds a full month of spending. Investment in churn prevention has the highest ROI.

From Analysis to Action: Building a Customer Health Dashboard

Repeat rate, purchase frequency, and LTV should be monitored regularly, not checked once and forgotten. Build a customer health monitoring system: weekly tracking of repeat rate and returning customer share; monthly review of frequency distribution and cohort churn curves; quarterly LTV updates. More importantly, set up churn alerts — when a customer's purchase interval exceeds 1.5x their average, automatically trigger retention actions (coupons, exclusive events). Traditional analysis requires extensive pivot tables and manual calculations. AI tools handle it automatically: upload transaction data, get repeat rate trends, frequency distributions, LTV calculations, and cohort analysis. No formulas needed — the full loyalty picture in 5 minutes.

Want to try it yourself?

Upload your spreadsheet and see what's in your data in 5 minutes.

Analyze My Data Free

Related Articles

AI Data Analysis2026-06-01·5 min read

AI Data Analysis: Upload a Spreadsheet, Get Insights in 30s

AI data analysis is transforming how businesses handle data. This guide covers the basics and shows how DataFish completes the full analysis pipeline in 30 seconds.

Productivity2026-05-28·4 min read

Excel for Data Analysis? 5 Pain Points and How AI Fixes Them

Excel data analysis too slow and tedious? Here are the 5 most common pain points and how AI cuts the time from hours to 30 seconds.

Industry2026-05-25·5 min read

Restaurant Chain Data Analysis: Store Comparison to Insights

A practical guide for restaurant chain data analysis: using AI to quickly complete store rankings, category analysis, time-slot analysis, and business recommendations.

← Back to Blog