FICILCOM Inc. LogoFICILCOM Inc.
  • Company
  • Company Overview
  • Mission · Vision · Values
  • Guidelines
  • Services
  • NeX-Ray
  • Xtrategy
  • Trial Job Change
  • Tsurugi
  • Recruit
  • Culture
  • Benefits
  • Process
  • FAQ
  • Open Positions
Contact
  1. Home
  2. Blog
  3. CRM, LTV & Customer Management
  4. How to Analyze Customers: RFM, LTV, and Persona Examples

How to Analyze Customers: RFM, LTV, and Persona Examples

顧客分析の基本フレーム|RFM・LTV・ペルソナを連携させる方法

Published: 05/22/2026

Last Updated: 10/08/2026

Category: CRM, LTV & Customer Management, Marketing Strategy

Authors: Shusaku Yosa

Ready to put this into practice?

Start Xtrategy for free→

1-minute signup, free, cancel anytime

Customer analysis combines transaction or usage records with research to decide whose problem to address and what to do next. RFM describes purchasing recency, frequency, and monetary value; LTV estimates customer value; personas help explain context and decisions. No single score proves a customer’s future value or the correct next offer.

Work from the decision to the population, period, definitions, analysis, action, and review. Keep revenue separate from profit, people separate from accounts, and a 90-day observation separate from a lifetime forecast.

Choose a method that matches the question

Method

Useful question

Cannot establish alone

RFM

How recently, often, and much did customers purchase?

Reasons, margin, or future retention

LTV and cohorts

What value does a comparable group produce over time?

Unobserved lifetime value or independent campaign causality

Personas and qualitative research

What situation, needs, and barriers shape decisions?

Population shares or guaranteed messaging success

ABC analysis

Where is revenue or another measure concentrated?

Every reason for concentration or future priority

Use ABC analysis for concentration, persona research for evidence about decisions, and persona marketing to translate findings into campaigns. They offer different views rather than interchangeable customer rankings.

Fix scope and definitions before scoring

  1. Choose one decision, such as repeat-purchase support or initial adoption.
  2. Define the customer unit: person, company, or contract, including merges and missing identifiers.
  3. Set the reference date and observation window.
  4. Specify eligible orders and treatment of cancellations, discounts, returns, shipping, and tax.
  5. Check duplicates, missing joins, outdated attributes, and unconfirmed orders.

An annual-contract business and a frequent-purchase business cannot use an identical inactivity rule without examining purchase cycles. New customers also have less observation time. Preserve privacy and contact restrictions when creating an analytical copy of the data.

Fictional RFM example with explicit three-level boundaries

The reference date is September 30, 2026; the 90-day window is July 3–September 30. R is days since the last eligible order, F is completed orders in the window, and M is revenue excluding tax after discounts and returns. This fictional dataset has no cancellations or returns.

Use three illustrative levels: R ≤30 days scores 3, 31–60 scores 2, and ≥61 scores 1. F ≥6 scores 3, 3–5 scores 2, and 1–2 scores 1. M ≥JPY 60,000 scores 3, JPY 30,000–59,999 scores 2, and below JPY 30,000 scores 1. These are fixed fictional thresholds, not tertiles, quartiles, five-level scores, or universal standards.

Customer

R days / F orders

M revenue

R–F–M score

A

5 / 10

JPY 100,000

3–3–3

B

15 / 8

JPY 80,000

3–3–3

C

25 / 5

JPY 60,000

3–2–3

D

45 / 4

JPY 40,000

2–2–2

E

65 / 2

JPY 20,000

1–1–1

F

85 / 1

JPY 10,000

1–1–1

Microsoft’s RFM configuration documentation illustrates the need to choose periods, divisions, and monetary treatment. Lower R means a more recent purchase, so it receives the higher score here. Higher M does not prove higher profit.

E and F are not confirmed churned customers. Check purchase cycles, usage, contracts, and seasonality. A and B do not automatically need discounts either: investigate what support or additional value would actually help, then consider margin and retention.

Separate a 90-day value observation from lifetime LTV

This is a separate fictional cohort of 100 customers, not the six-person RFM table. Every customer has the same 90-day observation age. Net revenue excluding tax is JPY 1,000,000; gross profit after cost of sales is JPY 600,000; additional customer support costs are JPY 100,000; acquisition costs are JPY 300,000. Support is not already deducted in gross profit.

Measure

Fictional calculation

Meaning

90-day contribution

JPY 600,000 − 100,000 = JPY 500,000

Profit after the defined additional support cost

Contribution per customer

JPY 500,000 ÷ 100 = JPY 5,000

Observed 90-day value, not lifetime value

CAC

JPY 300,000 ÷ 100 = JPY 3,000

Same acquisition population and cost scope

Observed contribution-to-CAC ratio

5,000 ÷ 3,000 ≈ 1.67

Not lifetime LTV/CAC or a universal viability threshold

After acquisition cost

JPY 500,000 − 300,000 = JPY 200,000

Not company net profit including all overhead and tax

Declare whether value means revenue, gross profit, or contribution. Stripe’s SaaS business guide discusses a simplified revenue-and-churn approach to LTV, which needs assumptions such as reasonably stable inputs. Do not relabel revenue-based estimates as profit or treat zero observed churn as infinite customer value.

See LTV and CAC for detailed scope and revenue, profit, and cash for the underlying distinctions. Compare cohorts at the same observation age rather than a new group’s first month against an established group’s year.

Turn findings into helpful actions and review them

Before discounting for a new customer who has stopped using the service, investigate setup or usability problems. High-frequency customers with heavy support demands may need a different service arrangement rather than more selling. Check needs, margin, and retention before proposing upsells or cross-sells.

Research should include relevant differences such as new, continuing, stopped, and non-buying customers, not only the largest spenders. Separate statements, observations, and assumptions, including user and approver perspectives. Select research methods for the question rather than assuming one fixed interview count is sufficient.

  • Save dates, population, order definitions, and monetary scope.
  • Keep scoring boundaries and tie handling explicit.
  • Use the starting eligible population for customer-retention measures.
  • Compare cohorts at equal observation ages.
  • Review costs, support burden, and contact preferences as well as revenue.

Keep new revenue out of existing-customer revenue-retention measures. Record price changes, seasonality, missing data, and small samples rather than attributing every change to the intervention. Update the analysis when customer circumstances change.

Connect the next decisions

For related decisions, see ABC calculations, using personas in campaigns, revenue, profit, and cash distinctions.

Xtrategy provides company, contact, and deal records alongside project planning and monthly budget-versus-actual management. Review the Xtrategy feature list to decide whether those management functions fit your team’s workflow.

 

 

 

 

Table of Contents

  1. Choose a method that matches the question
  2. Fix scope and definitions before scoring
  3. Fictional RFM example with explicit three-level boundaries
  4. Separate a 90-day value observation from lifetime LTV
  5. Turn findings into helpful actions and review them
  6. Connect the next decisions

 

Company

  • Company
  • Company overview
  • Mission · Vision · Values
  • Guidelines

Services

  • Services

Blog

  • Blog
  • Categories
  • Authors

Estimate

  • Estimate simulator

Careers

  • Careers
  • Culture & Work Style
  • Benefits & Systems
  • Hiring Process
  • FAQ
  • Open Positions

Policies

  • Privacy Policy
  • Anti-Social Policy
  • Information Security Policy

Contact

  • Contact

Social

  • X
  • LinkedIn
  • Facebook
  • Pinterest
© 2026 Ficilcom Inc.

Related Articles

パイプライン管理とは?案件の進捗を可視化して予測精度を上げる
CRM, LTV & Customer Management08/20/2026

What Is Pipeline Management? Stage Definitions, Stalled Deals, and Forecast Examples

Define pipeline stages using customer evidence, investigate stalled deals, calculate a hypothetical in-period bookings f...

Shusaku YosaRead more
中小企業のCRMの選び方|判断軸5つと導入時につまずく点
CRM, LTV & Customer Management08/19/2026

Choosing a CRM for a Smaller Company | Five Criteria and Where Adoption Stalls

How smaller companies should choose a CRM when nobody can be assigned to it full time. Covers five criteria — setup time...

Shusaku YosaRead more
アドボケイト(推奨者)の育て方|事例と施策
CRM, LTV & Customer Management06/30/2026

How to Build Customer Advocacy: Identify, Invite, Support, and Measure

Build a voluntary customer advocacy workflow: identify candidates, offer useful participation options, manage permission...

Shusaku YosaRead more