Materi 11 · Advanced Techniques

Customer Analytics & RFM

Tidak semua customer sama. Pelajari RFM segmentation, CLV calculation, churn prediction, dan cohort retention untuk bangun marketing & retention strategy data-driven.

⏱ 28 Menit🎯 Intermediate👥 Customer

1. KENAPA Customer Analytics?

Pareto: 80/20 Rule

20% customer mengontribusi 80% revenue di hampir semua bisnis. Tanpa segmentation, kamu treat semua customer sama — waste budget marketing & lose VIP. Customer analytics = unlock differentiation.

2. RFM SEGMENTATION

3 Dimensi RFM R = Recency (kapan terakhir order?)
F = Frequency (berapa kali order?)
M = Monetary (total revenue?)
// Score 1-5 per dimension. Total 125 kombinasi → simplify ke 11 segmen.

Cara Hitung Score

Quintile method: bagi customer ke 5 grup equal per dimensi.
Recency: 1 = sangat lama tidak order, 5 = baru order.
Frequency: 1 = jarang, 5 = sering.
Monetary: 1 = spend rendah, 5 = spend tinggi.

3. SEGMEN RFM POPULER

Champions
R5 F5 M5
Loyal
R5 F4 M4
Potential
R5 F3 M3
New Cust
R5 F1 M1
Promising
R4 F1 M1
Need Att
R3 F3 M3
Sleeper
R3 F2 M2
At Risk
R2 F4 M4
Cant Lose
R1 F5 M5
Hibernate
R2 F2 M2
Lost
R1 F1 M1
Lost VIP
R1 F4 M5
About
To Sleep
Cant Eval
R3 F1 M1
Loyal
Slow Down

Tiap segmen butuh treatment berbeda: Champions kasih VIP perks. At Risk kirim win-back campaign. Lost tinggalkan (cost-benefit jelek).

4. STRATEGI PER SEGMEN

SegmenKarakterStrategi
ChampionsRecent + frequent + high spendReward loyalty, ask referral, early access
LoyalHigh frequency consistentCross-sell, upsell, exclusive content
Potential LoyalistRecent + medium frequencyMembership program, target campaign
New CustomerJust made first orderOnboarding, welcome series, second order push
At RiskHigh value tapi inactivePersonalized win-back, special discount
Can't LoseWas VIP, sekarang inactive1-on-1 outreach, premium offer
HibernatingLow value + inactiveRe-engagement low cost, last attempt
LostLama tidak active, low valueTinggalkan. Fokus akuisisi baru.

5. CUSTOMER LIFETIME VALUE (CLV)

CLV Sederhana CLV = Avg Purchase × Purchase Frequency × Customer Lifespan
// Untuk e-commerce / retail
CLV Subscription CLV = ARPU × (1 / Churn Rate)
// Untuk SaaS / subscription

LTV/CAC Ratio

Selalu hitung LTV/CAC. Sehat: ≥ 3. < 1 = rugi setiap acquisition. 1-3 = bisa diperbaiki. ≥ 3 = scalable, push acquisition lebih agresif.

6. CHURN PREDICTION

7. PERSONA & BEHAVIORAL CLUSTERING

Beyond RFM

RFM = transactional. Tambah behavioral data: kategori pilihan, hari/jam aktif, channel preference, response rate. K-Means clustering pada feature ini = persona berbasis data, bukan asumsi marketing.

8. STUDI KASUS

Sephora's Beauty Insider Tier System

Sephora pakai customer analytics untuk 3-tier program: Insider, VIB, Rouge. Tier ditentukan oleh annual spend. Top 10% Rouge generate 36% revenue, dapat free shipping, exclusive event, makeover gratis. Hasilnya: retention Rouge 85% tahunan vs Insider 35%. Pelajaran: differentiate treatment by value.

📝 Tugas Praktik

  1. Pakai dataset transaksi (Online Retail di Kaggle).
  2. Hitung RFM score per customer dengan SQL atau pandas.
  3. Klasifikasikan ke 8-11 segmen RFM populer.
  4. Hitung CLV per segmen.
  5. Visualisasi RFM matrix heatmap: sumbu R vs F, color = M.
  6. Rancang 1 strategi marketing berbeda untuk 3 segmen kunci (Champion, At Risk, Lost).

Rangkuman