Monetization Analysis

Repeat Purchase Analysis

Diagnose why repeat-purchase rate drops, assess repeat-purchase health, and compare repeat-purchase behavior across user groups, turning "why users don't renew" from a vague guess into data-driven attribution.

IndustriesGame SubscriptionShort-Drama MembershipTool SubscriptionE-commerce

Agent · Skills Live

skill · loadedindustry+100

Run this Skill on my data

Retention heat-grid

cohort × day

D1D3D7D14D30

Wk1

Wk2

Wk3

Analysis readyevidence attached

Repeat Purchase Analysis is a repeat-purchase behavior diagnosis capability in ThinkingAI's industry Skill library. It helps operations teams diagnose why repeat-purchase rate drops abnormally, assess current repeat-purchase health, and compare repeat-purchase behavior across user groups, turning "why users don't renew" from a vague guess into data-driven attribution. This is ThinkingAI's decade of industry experience, packaged into ready-to-use Skills.

Live walkthrough

AgentSave as AgentExport session

repurchase-analysis

Our short-drama app's VIP monthly-pass repeat-purchase rate dropped to 48% in March, after holding steady around 65%. The content team says it's the library, the UA team says it's not their fault. Help me pin down the root cause before the end-of-month review.

Data validation: confirm the mapping of the renewal event vip_renew to the order table

Reconciles with the order system within 0.2%, data is trustworthy

Locate the month-by-month trend

January 64.8%, February 63.9%, March 48.0%: a sudden break, not a slow decline

Compare the repeat-purchase interval distribution (December baseline vs March)

Segment comparison: channel × user age × price tier

Channel B repeat rate 56% → 32%, other dimensions stable

Cross-reference the UA change log and output attribution

Diagnosis done. It's not a library problem, it's a user-mix problem. In March channel B ran a "$1 first month" promotion, lifting new-user volume 2.5×, but this cohort's repeat rate is only 32%, and their share of expiring users spiked to 45%, so the weighted total was dragged straight from 65% down to 48%.

1. Start with the shape of the repeat-purchase interval distribution

VIP monthly-pass repeat-purchase interval distribution after expiry (December baseline vs March)　Unit: %

December baselineMarch

What collapsed is "renew on the expiry day" (35% → 24%), the signature of the most loyal, strongest-subscription-habit segment being diluted; the hesitant segment past 3 days barely moved. Users didn't become more hesitant, the "renew on expiry" crowd got diluted.

2. Segment comparison pinpoints the source

Segment	Share of March expiring users	December repeat rate	March repeat rate
Channel A users	55%	67%	61%
Channel B users	45% (only 18% in December)	56%	32%
Weighted total	100%	65%	48%

← Scroll to see more

Check: December = 82%×67% + 18%×56% = 65%; March = 55%×61% + 45%×32% = 48%, the weighting reconciles exactly. Attribution chain: channel B's "$1 first month" promotion → brought in a large number of one-off spenders not enrolled in auto-renewal → this group leaves at expiry (32% repeat) → combined with their share rising 18% → 45% → total repeat rate 65% → 48%.

3. Ruling out content and pricing

Both lines were checked: pricing didn't change in March; channel A users' repeat rate went 67% → 61%, a mild drop uncorrelated with the library update cadence. If it were a library problem, the drop wouldn't concentrate only on one channel's promotional new users. The content team isn't to blame this time.

Heads-up

48% is already below the healthy range for short drama (55%–70%), but don't rush to kill the promotion: (1) change "$1 first month" to "$1 first month + auto-renew checked by default + $9.9 transitional second-month price" to guide one-off users toward a subscription habit; (2) in the review, present "promotional new-user repeat" and "organic new-user repeat" separately, or April's total will keep being dragged by the mix and you'll have the same argument all over again.

On your data

That was a simulated run

Leave your work email and we will run a live walkthrough on your real business data.

The problem

The lifeblood of a subscription product is its repeat-purchase (renewal) rate, and a single 5% drop can mean a 20% revenue shrink over the next 3 months. But the causes of a drop are often tangled together: did a price change hurt the value proposition? Is the new version's content not compelling enough to renew? Did a competitor pull users away? Or do users from a particular channel simply have low renewal intent to begin with? Facing a decline, analysts often look only at the total trend without segmenting user groups, leaving the root cause blurry.

What it does

Four analysis dimensions covering the full repeat-purchase path: churn diagnosis, trend analysis, health assessment, and segment comparison

Validate the data before analyzing: mandatorily query the project's real event and property lists to ensure the repeat-purchase event name and property mapping are accurate

Built-in industry benchmarks: game subscription, short-drama membership, tool Pro, and e-commerce repeat purchase each have different health standards

Output root-cause attribution plus optimization recommendations: give concrete interventions for the anomalous dimension diagnosed

When to use it

A game's monthly or seasonal pass renewal rate drops suddenly and you need to determine whether it is a pricing, content, or channel issue

A short-drama platform's VIP membership repeat-purchase rate is below the industry benchmark and needs a health assessment and optimization direction

A tool product's Pro subscription renewal rate fluctuates and needs analysis of cyclical patterns and segment differences

An e-commerce platform's repeat-purchase rate declines and needs comparison of repeat-purchase behavior across channels and user tiers

Assessing whether a new pricing strategy has a positive effect on repeat-purchase rate after launch

In the field

Case

A short-drama platform · VIP monthly membership repeat-purchase diagnosis

VIP monthly membership repeat-purchase rate fell from 65% to 48%. After validating the renewal event name, the Skill decomposed it: the drop concentrated in March, and channel B new users had a repeat-purchase rate of only 32%, clearly below channel A's 58%, with the current rate below the short-drama industry benchmark. The root cause was a high share of one-time-spend users brought in by channel B, and the team adjusted channel spend and designed a first-renewal offer accordingly.

FAQ

What is the difference between repeat purchase analysis and LTV analysis?

Repeat purchase analysis focuses on behavioral attribution of why users renew or rebuy, while LTV analysis focuses on the numerical calculation of user lifetime value.

Is it only for subscription products?

No. E-commerce repeat purchase (a user's second purchase of a physical item) also applies. Any repeat-purchase behavior can be analyzed.

Is a repeat-purchase drop always a product problem?

Not necessarily. It could be declining channel quality, a price change, competitor diversion, or seasonal fluctuation. You must break it down by dimension to pinpoint the root cause.

Precise LTV through curve fitting

LT Prediction (User Lifetime Prediction)

Estimate average user lifetime in days

LTV Payback Period Prediction

Predict whether spend will pay back

LTV Analysis and Monitoring

Diagnose the root cause of LTV drops