Loyalty Matrix

Customer Retention & Repeat Purchase Analytics

Why Repeat Buyers Are Your Real Asset

Acquiring a new customer costs 5-7x more than retaining an existing one. Yet most PPC strategies focus entirely on new customer acquisition.

The Loyalty Matrix connects your ad data to customer lifetime behavior:

Which campaigns acquire customers who buy repeatedly?

Which channels produce one-and-done buyers?

Which customer segments are at risk of churning?

What's the ROI of reactivation vs new acquisition?

Example discovery:

Google Shopping acquires customers with 3.2 average lifetime purchases. Display acquires customers with 1.1 average purchases. Both have similar CPA ($45 vs $50). But Shopping customers generate 3× more lifetime revenue.

Without the Loyalty Matrix, these channels look equally efficient. With it, Shopping is clearly the superior growth channel.

1. RFM Segmentation Model

RFM = Recency, Frequency, Monetary

Each customer is scored 1-5 on three dimensions:

Recency: Days since last purchase

5: Purchased in last 7 days

4: 8-30 days

3: 31-90 days

2: 91-180 days

1: 180+ days

Frequency: Total number of purchases

5: 10+ purchases

4: 5-9 purchases

3: 3-4 purchases

2: 2 purchases

1: 1 purchase (one-and-done)

Monetary: Total spend (percentile-based)

5: Top 20% spenders

4: 60th-80th percentile

3: 40th-60th percentile

2: 20th-40th percentile

1: Bottom 20%

Resulting Segments:

🏆 Champions (R:5, F:5, M:5): Your best customers. Recent, frequent, high-value. Protect at all costs.

💎 Loyal Customers (R:4-5, F:3-5, M:3-5): Regular buyers with solid spend. Nurture with exclusive offers.

🌟 Potential Loyalists (R:4-5, F:1-2, M:2-4): Recent buyers who could become loyal. Critical window for retention campaigns.

⚠️ At Risk (R:2-3, F:3-5, M:3-5): Were loyal but haven't bought recently. Reactivation urgency.

💤 Hibernating (R:1-2, F:1-2, M:1-3): Low activity across all dimensions. May not be worth reactivation cost.

👋 Lost (R:1, F:1, M:1): One purchase long ago. Cheapest to let go.

2. Channel-Loyalty Correlation Map

What it shows: For each acquisition channel, the distribution of customers across loyalty segments.

Visual: Stacked bar chart. Each bar = one acquisition channel. Colors = loyalty segment proportions.

Example findings:

Channel	Champions	Loyal	Potential	At Risk	Lost
Brand Search	22%	35%	18%	15%	10%
Non-Brand Search	8%	20%	25%	22%	25%
Shopping	15%	30%	22%	18%	15%
Display	3%	8%	12%	25%	52%

The insight: Display has the highest Lost rate (52%)—most Display-acquired customers never return. Brand Search produces the most Champions (22%).

Strategic implication: If your goal is lifetime value (not just first purchase), reallocate budget from Display to Shopping and Non-Brand Search.

But be careful: Display might be a necessary top-of-funnel driver (check the Strategist Center for attribution data). Don't cut it without understanding the full journey.

3. Churn Probability Engine

How it works:

The system calculates each customer's average purchase interval from historical data. When a customer exceeds 2× their average interval without purchasing, churn probability rises sharply.

Example:

Customer A buys every 25 days on average

It's been 60 days since last purchase (2.4× interval)

Churn Probability: 78%

The Churn Dashboard shows:

Total customers at risk (churn probability >60%)

Revenue at risk (sum of these customers' historical monthly spend)

Days since last purchase distribution

Segment breakdown (how many Champions are at risk vs Potential Loyalists)

Alert thresholds:

🟡 Warning (1.5× interval): Customer is slowing down. Consider a gentle re-engagement email.

🔴 Critical (2× interval): Customer is likely churning. Trigger reactivation campaign.

⬛ Lost (3× interval): Customer has almost certainly churned. Reactivation ROI is low.

Note: This uses simple interval-based estimation, not ML prediction. It's directionally accurate for businesses with regular purchase cycles (subscriptions, consumables, repeat services). Less reliable for infrequent, high-value purchases (e.g., furniture, cars).

4. Reactivation ROI Calculator

The question: Is it worth running a reactivation campaign for at-risk customers, or should you spend that money on new acquisition?

Reactivation ROI Formula:

Reactivation ROI = (Reactivated Customers × Predicted LTV) / Reactivation Campaign Cost

vs New Acquisition ROI:

New Acquisition ROI = (New Customers × Predicted LTV) / New Acquisition Cost

Example:

Metric	Reactivation	New Acquisition
Campaign Cost	$2,000	$5,000
Customers Reached	500 at-risk	10,000 prospects
Conversion Rate	8% (40 reactivated)	1% (100 new)
Predicted LTV	$400 (known history)	$300 (estimated)
Total Value	$16,000	$30,000
ROI	8:1	6:1

In this example, reactivation has 33% higher ROI because these customers already know your brand—conversion rates are higher and LTV is more predictable.

When reactivation doesn't work: If customers churned due to product dissatisfaction (not just inattention), reactivation campaigns have very low response rates. Check NPS or review data before investing.

Recommended Monthly Workflow

Loyalty analysis is a monthly exercise. Customer segments shift slowly—checking daily adds noise.

First Monday of Every Month (30 minutes):

Step 1: Segment Migration (10 min)

Compare this month's segment distribution to last month's. Are Champions growing or shrinking? Is the At Risk segment expanding?

Step 2: Channel Quality Check (10 min)

Review the Channel-Loyalty Map. Has any channel's loyalty distribution changed? Did a recent campaign bring in higher or lower quality customers?

Step 3: Churn Review (5 min)

Check the Churn Dashboard. How many customers crossed the 2× interval threshold this month? Is the revenue at risk growing?

Step 4: Reactivation Decision (5 min)

If At Risk segment grew >10% month-over-month, run the Reactivation ROI Calculator. If ROI exceeds 3:1, trigger a reactivation campaign.

Data Requirements: Requires GA4 or CRM data with customer-level purchase history. Minimum 6 months of transaction data for reliable segmentation. 12+ months is ideal.

When to Use This Dashboard (vs. Other Tools)

Use Loyalty Matrix when you want to:

Understand customer lifetime behavior by acquisition channel

Identify at-risk customers for reactivation

Compare channel quality beyond first-purchase CPA

Justify retention budgets vs acquisition budgets

Don't use Loyalty Matrix when you want to:

Optimize campaign-level performance (use PCC)

Calculate CAC and LTV (use Unit Economics)

Analyze creative performance (use Creative Lab)

Who Should Use This:

✅ E-commerce managers (repeat purchase optimization)

✅ Retention marketers (churn prevention)

✅ CMOs (channel mix strategy based on LTV)

❌ PPC practitioners (too downstream for daily optimization)

❌ Lead-gen businesses (single-purchase model doesn't benefit as much)

How Often: Monthly. Quarterly for strategic channel reallocation decisions.

Technical FAQ

Q: What data sources does the Loyalty Matrix need?

GA4 transaction data (minimum), ideally supplemented with CRM data. Needs customer-level purchase history with timestamps, amounts, and acquisition source. Minimum 6 months of data; 12+ months is ideal.

Q: How are the RFM score thresholds determined?

Recency and Frequency use fixed thresholds (configurable in settings). Monetary uses percentile-based scoring relative to your customer base—so a '5' always means top 20% of YOUR customers, regardless of absolute spend.

Q: Can I customize the loyalty segments?

Yes. The default segments (Champions, Loyal, Potential Loyalists, At Risk, Hibernating, Lost) use standard RFM combinations. You can adjust the score ranges that define each segment in Settings.

Q: How reliable is the churn probability estimate?

It's directionally accurate for businesses with regular purchase cycles (subscriptions, consumables, repeat services). Less reliable for high-value infrequent purchases like furniture or cars. The model uses simple interval-based estimation, not ML prediction.

Q: Should I stop advertising to 'Lost' customers?

Generally yes—the cost to reactivate a Lost customer (R:1, F:1, M:1) is often higher than acquiring a new one. Exception: if your product has changed significantly since their purchase, a 'we've improved' campaign might work.

Q: How does this connect to my PPC campaigns?

The Channel-Loyalty Map shows which ad campaigns produce high-LTV customers vs one-and-done buyers. Use this to shift budget toward channels that acquire Champions, even if their first-purchase CPA is slightly higher.

Q: Can I create remarketing audiences from these segments?

Yes. Export any segment as a customer list and upload it to Google Ads as a Customer Match audience. Common use: create a 'Potential Loyalists' remarketing list with special offers to push them toward the Loyal segment.

Q: What's the difference between this and Google Analytics' 'Returning Users' metric?

GA4's returning users metric is binary (new vs returning) and session-based. The Loyalty Matrix uses multi-dimensional scoring (Recency × Frequency × Monetary) at the customer level, giving you much richer segmentation and actionable groupings.