ROI Case File No.305 | 'WineTrade's Invisible Customers'

Chapter 1: The Nightmare of 150,000—Price Revision Hell

The week after resolving ElectraCore's Design Thinking incident, a consultation arrived from Tokyo regarding an alcoholic beverages and food wholesale company's EC operations. Episode 305 of Volume 25 "The Pursuit of Certainty" tells the story of abandoning the illusion of treating all customers equally and identifying valuable customers.

"Detective, we operate 5 EC sites with 150,000 customers. However, we're overwhelmed responding to all customers with every manufacturer price revision and new product registration. Average 180 price changes monthly, 50 new product registrations. The field has reached its limit."

Keiko Yamada, EC business manager at WineTrade Inc., originally from Minato-ku, visited 221B Baker Street with an exhausted expression. In her hands were stacks of manufacturer price revision notices and, in stark contrast, complaint records stamped "Response Delay."

"We operate alcoholic beverage and food wholesale in Tokyo. Main business is wholesale to restaurants and retail stores, but we started EC business 5 years ago. Wine specialty site, sake site, craft beer site... 5 sites selling to 150,000 customers."

WineTrade's EC Operations Crisis: - Establishment: 1988 (Alcoholic beverages/food wholesale) - Annual Revenue: ¥5.2B (of which EC ¥1.2B) - Number of EC Sites: 5 - Registered Customers: 150,000 - Products Handled: 8,200 SKUs - Price Revisions: Monthly average 180 - New Product Registrations: Monthly average 50 - EC Operations Staff: 8 - Complaints due to Response Delays: Monthly average 42

Deep anxiety filled Yamada's voice.

"The problem is giving the same response to all customers. When manufacturer price revision notices arrive, we search for applicable products among 8,200 items and change prices on all 5 sites. Then we email all 150,000 customers. 'Wine XX price has changed.'"

Typical Price Revision Response (per case):

Monday Morning: Notification from Manufacturer A: "20 French wine brands, 5-10% price increase from next month"

Monday PM - Tuesday: - Identify applicable products: 20 brands × 5 sites = 100 locations - Price change work on each site: Manual input - Time required: Average 15 min per brand → Total 25 hours

Wednesday: - Create customer email: "Price Revision Notice" - Sent to: All 150,000 people - Email open rate: 2.8%

Thursday - Friday: - Customer inquiry response: "Why the price increase?" - Complaints: "Too many emails" - Time required: About 10 hours

1 price revision requires 35 hours of work total

180 monthly price revisions = 6,300 hours monthly work (8 staff × 160 hours monthly = only 1,280 hours available)

Result: Work can't keep up, response delays

"It's impossible. Chased by price revisions, new product registrations delay. No time for customer response. Unless we change the policy of giving same response to all customers, this situation won't improve."

Chapter 2: The Illusion of Equality—Those Who Try to Save Everyone Save No One

"Ms. Yamada, what policy guides current customer response?"

To my question, Yamada answered.

"'Treat all customers equally' policy. All 150,000 are important customers. So price revisions and new products are notified to everyone. But as a result, work can't keep up."

Current Approach (Egalitarian Type): - Policy: "Treat all customers equally" - Response: Notify everyone of price revisions and new products - Problem: Work volume overwhelming, impossible to respond

I explained the importance of recognizing customer value differences.

"Treating all customers equally is a beautiful ideal. But in reality, customer value differs. RFM Analysis—Recency, Frequency, Monetary. Classifying customers by these 3 axes reveals valuable customers."

⬜️ ChatGPT | Catalyst of Vision

"Don't try to save everyone. Classify customers with RFM, prioritize valuable customers"

🟧 Claude | Alchemist of Stories

"Equality is illusion. Customers have ranks. Acknowledging that is true customer orientation"

🟦 Gemini | Compass of Reason

"RFM is customer value measurement technique. Identify premium customers on 3 axes: Recency, Frequency, Monetary"

The three members began analysis. Gemini developed the "RFM Analysis Framework" on the whiteboard.

RFM's 3 Indicators: 1. Recency: When was last purchase 2. Frequency: How frequently do they purchase 3. Monetary: How much have they purchased cumulatively

"Ms. Yamada, let's classify WineTrade's 150,000 customers with RFM."

Chapter 3: Dissecting 150,000—Customers Divide into 5 Segments

Phase 1: Data Extraction (1 week)

First, analyzed purchasing data from the past 2 years.

150,000 Customer Data: - Customer ID - Last purchase date - Purchase count - Cumulative purchase amount

Phase 2: RFM Score Calculation (3 days)

Assigned each customer R, F, M scores of 1-5 points.

Recency Scoring: - 5 points: Purchased within 30 days - 4 points: Within 31-90 days - 3 points: Within 91-180 days - 2 points: Within 181-365 days - 1 point: 366+ days ago

Frequency Scoring: - 5 points: 12+ times annually (monthly+) - 4 points: 6-11 times annually - 3 points: 3-5 times annually - 2 points: 2 times annually - 1 point: Only 1 time annually

Monetary Scoring: - 5 points: Cumulative ¥500K+ - 4 points: Cumulative ¥200-499K - 3 points: Cumulative ¥50-199K - 2 points: Cumulative ¥10-49K - 1 point: Cumulative under ¥10K

Customer RFM Score Examples: - Customer A: R5 F5 M5 (Best customer) - Customer B: R1 F1 M1 (Dormant customer) - Customer C: R4 F3 M4 (Good customer)

Phase 3: Customer Segmentation (1 week)

Based on RFM scores, classified 150,000 into 5 segments.

Segment 1: VIP Customers (Champions) - Definition: R≥4, F≥4, M≥4 - Count: 8,400 (5.6%) - Characteristics: Recent purchase, frequent purchase, high-value purchase - Average Annual Purchase: ¥280K - Segment Total Sales Contribution: ¥2.35B (196% of EC sales)

Segment 2: Loyal Customers - Definition: R≥3, F≥3, M≥3 - Count: 22,500 (15%) - Characteristics: Regular purchase, medium-high value purchase - Average Annual Purchase: ¥120K - Segment Total Sales Contribution: ¥2.7B (225%)

Segment 3: Potential Loyalists - Definition: R≥3, F=2-3, M=2-3 - Count: 31,500 (21%) - Characteristics: Recent purchase, low frequency but growth potential - Average Annual Purchase: ¥38K - Segment Total Sales Contribution: ¥1.2B (100%)

Segment 4: At Risk Customers - Definition: R≤2, F≥3, M≥3 - Count: 18,000 (12%) - Characteristics: Previously loyal but no recent purchases - Average Annual Purchase: ¥8K (declining) - Segment Total Sales Contribution: ¥144M (12%)

Segment 5: Lost Customers - Definition: R≤2, F≤2, M≤2 - Count: 69,600 (46.4%) - Characteristics: No long-term purchases, low-value purchases - Average Annual Purchase: ¥200 - Segment Total Sales Contribution: ¥14M (1.2%)

Phase 4: Shocking Facts (1 day)

Visualizing RFM analysis results revealed shocking facts.

Distribution Beyond Pareto's Law:

Discovery: - Top 20.6% of customers generate 421% of EC sales - Bottom 46.4% (dormant) contribute only 1.2% of sales

Yamada was stunned.

"So we were giving same response to all 150,000, but actually top 30,000 generate most sales..."

Chapter 4: Priority as Strategy—Respond from Valuable Customers

Phase 5: Segment-Specific Response Strategy Design (2 weeks)

Based on RFM analysis, changed response policy by customer segment.

Segment 1: VIP Customers (8,400)

Response Policy: - Top priority response - Price revisions: "Advance notice" (provide purchase opportunity before increase) - New products: "Pre-release information" (1 week before general sale) - Assign dedicated staff

Measure Example (Price Revision): - Email 1 week before increase: "XX wine price increasing next month. Purchase now at old price" - Follow-up after increase: "Introduce other recommended wines"

Segment 2: Loyal Customers (22,500)

Response Policy: - High priority response - Price revisions: "Normal notification" - New products: "Priority information" - Propose regular purchase plans

Segment 3: Potential Loyalists (31,500)

Response Policy: - Medium priority response - Price revisions: "Notify only applicable product purchasers" (don't send to all) - New products: "Information only for interest categories" - Measures to increase purchase frequency (coupon distribution)

Segment 4: At Risk Customers (18,000)

Response Policy: - Reactivation measures - Price revision notifications: "None" (not interested) - Return campaign: "20% OFF for returning customers"

Segment 5: Lost Customers (69,600)

Response Policy: - Minimum response - Price revision/new product notifications: "None" - Only annual "Please return" campaign - Don't allocate resources

Phase 6: New Price Revision Flow (Implementation 1 month)

Previous Flow (Notify everyone): 1. Price change work: 25 hours 2. Email all customers (150,000) 3. Inquiry response: 10 hours 4. Total: 35 hours

New Flow (RFM Segment-specific):

Step 1: Advance notice to VIP customers (8,400) - Work time: 2 hours - Email open rate: 38% (13× previous 2.8%) - Pre-purchase rate: 12% (purchase before increase)

Step 2: Normal notification to Loyal customers (22,500) - Work time: 3 hours - Email open rate: 18%

Step 3: Notify only applicable Potential Loyalists (31,500) - Extract applicable: Only customers who purchased this product before (about 2,800) - Work time: 1 hour - Email open rate: 22%

Step 4: No notification to At Risk/Lost customers (87,600) - Work time: 0 hours

New Flow Total Work Time: 6 hours (17% of previous)

Chapter 5: Choice as Results—Narrowing Customers Improves Overall

Phase 7: Results After 6 Months

Dramatic Work Time Reduction: - Price revision response: Monthly 6,300 hours → 1,080 hours (83% reduction) - New product registration response: Monthly 400 hours → 120 hours (70% reduction) - Reduced time strengthens customer response: Phone response time +280 hours/month

Customer Satisfaction Improvement:

VIP Customer Reactions: "Helpful to inform before price increase. Bought in bulk" "Happy about pre-release information for new products. Feels like special treatment" - VIP customer satisfaction: 4.2 → 4.8 - VIP customer purchase frequency: 12× yearly → 15× yearly (+25%)

Loyal Customer Reactions: "Information better organized than before, easier to understand" - Loyal customer satisfaction: 3.8 → 4.3

Potential Loyalist Reactions: "Unrelated product notifications decreased, emails easier to read" - Email open rate: 2.8% → 22% (about 8×)

Lost Customers: - Stopped notifications, complaints "Too many emails" decreased: 42 cases/month → 2 cases/month

Business Results After 12 Months:

EC Business Growth: - Annual Revenue: ¥1.2B → ¥1.56B (+30%) - Breakdown: - VIP customers: ¥2.35B → ¥2.94B (+25%) - Loyal customers: ¥2.7B → ¥3.24B (+20%) - Potential Loyalists: ¥1.2B → ¥1.38B (+15%)

Customer Count Changes: - VIP customers: 8,400 → 10,500 (+25%, promoted from Loyal) - Loyal customers: 22,500 → 25,200 (+12%, promoted from Potential) - At Risk customers: 18,000 → 12,600 (30% returned)

Lost Customer Reduction: - 69,600 → 58,200 (email list cleanup)

Organizational Change: - EC operations staff: 8 → 12 (increased, gained capacity) - Response delay complaints: 42/month → 3/month (93% reduction) - Staff satisfaction: 2.8 → 4.5 ("Work became manageable")

Yamada's Reflection:

"Until RFM analysis, I thought all 150,000 customers had equal value. But actually, top 30,000 generated most sales.

Trying to treat everyone equally, we ultimately satisfied no one. However, prioritizing VIP customers, VIP customers appreciated special treatment, Potential Loyalists were satisfied with fewer unnecessary notifications, and resources for Lost customers were reduced.

Classifying customers wasn't discrimination, it was optimization."

Chapter 5: Detective's Diagnosis—Choice Creates Value

That night, I contemplated the essence of RFM analysis.

The ideal of "treating all customers equally" is beautiful. But in reality, you must respond with limited resources.

WineTrade tried to give same response to all 150,000 and satisfied no one. However, by classifying customers with RFM and prioritizing VIP customers, overall satisfaction improved.

"Equality is illusion. Customers have value differences. Acknowledging that and prioritizing valuable customers is true customer orientation."

The next case will also depict the moment RFM visualizes customer value.

"Those who try to save everyone save no one. Classify customers with RFM, prioritize valuable customers. That is strategy"—From the Detective's Notes

Related Files