For each CRM Key Challenge: a business definition, the variable measured, the Uplift formula, the financial valorisation formula, and a complete worked example.
Challenge 1 — Prospect → First-Time Buyer
Definition: Measure Reelevant’s impact on converting a prospect (never purchased) into a buyer.
Variable measured: First-purchase rate within N days following exposure to an email with a Reelevant block.
Design: A/B test on email opt-in prospects who have never purchased. Exposed = email + Reelevant block.
Uplift conversion = Rate_exposed − Rate_control
Incremental new buyers = Uplift × Exposed Population size
Incremental Value = Incremental new buyers × Average basket (first purchase)
Worked example
| Metric | Value |
|---|
| Exposed prospects | 20,000 |
| Conversion rate (Exposed) | 3.2% — 640 first purchases |
| Conversion rate (Non-Exposed) | 1.9% — 380 first purchases |
| Incremental Value | €29,900 (260 × €115 basket) |
260 additional prospects became buyers thanks to the Reelevant block. With an average first-purchase basket of €115, this represents €29,900 in incremental revenue.
What you CAN affirm
- Reelevant generated 260 additional first purchases on this population over the measured period.
What you CANNOT affirm
- That these new buyers will repurchase — do not project this value over 12 months without estimating the segment’s repurchase rate.
- Use the first-purchase average basket, not the overall average basket.
- Verify that both groups have identical acquisition source distributions.
Challenge 2 — Loyalty Programme Enrolment
Definition: Measure Reelevant’s impact on the loyalty programme enrolment rate.
Variable measured: Enrolment rate within 30 days. Difference in annual value between members and non-members.
Population: Active customers not yet enrolled. CRM email with and without loyalty promotion block.
Uplift enrolment = Rate_exposed − Rate_control
Incremental enrolments = Uplift × Exposed Population size
Incremental Value = Incremental enrolments × (Annual value_member − Annual value_non-member)
Worked example
| Group | Population | Enrolment rate | Enrolled | Avg. annual value | Estimated revenue |
|---|
| Exposed | 50,000 | 6.5% | 3,250 | €130 | €422,500 |
| Non-Exposed | 50,000 | 4.0% | 2,000 | €130 | €260,000 |
| Delta | | +2.5 pts | +1,250 | vs €100 (non-members) | +€37,500 |
1,250 people joined the programme thanks to Reelevant. Members spend on average €130/year versus €100 for non-members. Each incremental enrolment is worth €30 in annual surplus, totalling €37,500 in Incremental Value.
What you CAN affirm
- Reelevant drove 1,250 additional loyalty enrolments, each worth €30/year in incremental spend.
What you CANNOT affirm
- That the member/non-member value difference is entirely causal — members may be naturally more engaged (selection bias).
- To neutralise selection bias, compare customers with similar histories BEFORE enrolment.
- Do not annualise without verifying average tenure in the programme.
Challenge 3 — Status Upgrade (Loyalty Tier Migration)
Definition: Measure Reelevant’s impact on migrating members from status A (e.g. Silver) to status B (e.g. Gold).
Variable measured: Migration rate to the higher status within 60–90 days.
Unit value: Difference in annual spend between status B and status A.
Uplift migration = Rate_migration_exposed − Rate_migration_control
Incremental migrations = Uplift × Exposed Population size
Incremental Value = Incremental migrations × (Annual revenue_status_B − Annual revenue_status_A)
Worked example
| Metric | Value |
|---|
| Silver members exposed | 30,000 |
| Migration rate (Exposed) | 8.4% — 2,520 → Gold |
| Migration rate (Non-Exposed) | 5.1% — 1,530 → Gold |
| Incremental Value | €79,200 (990 × €80 annual gap) |
990 additional Silver members upgraded to Gold thanks to Reelevant. A Gold member spends on average €80 more per year than a Silver member. Incremental Value = 990 × €80 = €79,200.
What you CAN affirm
- Reelevant accelerated 990 status upgrades, each worth €80/year in additional spend.
What you CANNOT affirm
- That the spend difference is not partly due to seasonality (holidays, sales).
- Exclude members close to the natural migration threshold (self-selection bias).
Challenge 4 — Retention Rate Increase
Definition: Measure whether Reelevant blocks increase the proportion of customers who purchase at least once in a given period.
Variable measured: 90-day retention rate (% of customers with at least one purchase during the period).
Unit value: Average annual revenue of a retained customer × additional retention probability.
Uplift retention = Rate_retention_exposed − Rate_retention_control
Incremental retained customers = Uplift × Exposed Population size
Incremental Value = Incremental retained × Average annual revenue (retained customer)
Worked example
| Metric | Value |
|---|
| Customers exposed | 25,000 (active, at risk of disengagement) |
| Retention (Exposed) | 74.5% over 90 days |
| Retention (Non-Exposed) | 70.2% over 90 days |
| Incremental Value | €193,500 (1,075 × €180 annual revenue) |
Thanks to Reelevant, 1,075 more customers were retained (purchased at least once). With an average annual revenue of €180 for a retained customer, this represents €193,500 in Incremental Value.
What you CAN affirm
- Reelevant retained 1,075 additional customers over the 90-day window.
What you CANNOT affirm
- That 90-day retention fully represents annual retention — ideally measure over 12 months.
- Use the annual revenue of customers with a similar profile, not the global average.
- Verify retention is not inflated by concurrent promotions.
Challenge 5 — Reactivation of Inactive Customers
Definition: Measure whether Reelevant increases the repurchase rate of inactive customers (no purchase for X days).
Population: Customers with no purchase for 90–180 days. Still email-active (opt-in, not unsubscribed).
Unit value: Revenue generated over the 12 months following reactivation.
Uplift reactivation = Rate_reactivation_exposed − Rate_reactivation_control
Incremental reactivated = Uplift × Exposed Population size
Incremental Value = Incremental reactivated × Average post-reactivation revenue (12 months)
Worked example
| Metric | Value |
|---|
| Inactive customers exposed | 15,000 (90–180 days without purchase) |
| Reactivation (Exposed) | 9.8% — 1,470 reactivated |
| Reactivation (Non-Exposed) | 6.2% — 930 reactivated |
| Incremental Value | €97,200 (540 × €180 post-reactivation revenue) |
540 additional inactive customers repurchased thanks to the Reelevant block. Estimating that a reactivated customer generates €180 on average over the following 12 months, the Incremental Value is €97,200.
What you CAN affirm
- Reelevant reactivated 540 additional customers from the 90–180 day dormant segment.
What you CANNOT affirm
- That the post-reactivation revenue estimate is exact — wait 6–12 months of actual data if possible.
- Exclude customers who would have purchased anyway (e.g. seasonal recurring buyers).
- Cross-reference with the second-purchase probability to avoid over-estimating residual value.
Challenge 6 — Churn Reduction
Definition: Measure whether Reelevant reduces the proportion of customers who permanently stop purchasing.
Variable measured: Churn rate = proportion of customers with no purchase after N months of inactivity.
Unit value: Preserved LTV: average annual revenue × estimated residual lifetime.
Churn decrease = Rate_churn_control − Rate_churn_exposed
Incrementally saved customers = Churn decrease × Exposed Population size
Incremental Value = Saved customers × Preserved LTV
Preserved LTV ≈ Average annual revenue × Estimated residual lifetime (in years)
Worked example
| Metric | Value |
|---|
| At-risk customers exposed | 10,000 (high churn score) |
| Churn (Non-Exposed) | 22.0% — 2,200 lost |
| Churn (Exposed) | 17.5% — 1,750 lost |
| Incremental Value | €135,000 (450 saved × €300 preserved LTV) |
Reelevant saved 450 customers who would otherwise have churned. Estimating that a preserved customer generates €300 of value over their residual lifetime, the Incremental Value is €135,000.
What you CAN affirm
- Reelevant prevented 450 customers from churning during the measurement period.
What you CANNOT affirm
- That a “saved” customer will not churn again 6 months later — measure effective retention at 12 months.
- Preserved LTV relies on residual lifetime hypotheses — bound them conservatively.
- Exclude customers recovered via other actions (promotional emails, customer service) to isolate Reelevant’s effect.
Challenge 7 — Purchase Frequency Increase
Definition: Measure whether Reelevant increases the average number of transactions per customer over a given period.
Variable measured: Average number of purchases per customer over 90 days or 12 months. Use t-test or Mann-Whitney.
Valorisation: Incremental frequency × average basket = incremental revenue per customer × Exposed Population size.
Uplift frequency = Frequency_exposed − Frequency_control (purchases/customer/period)
Incremental revenue per customer = Uplift frequency × Average basket
Total Incremental Value = Incremental revenue per customer × Exposed Population size
Worked example
| Metric | Value |
|---|
| Customers exposed | 20,000 (active buyers) |
| Frequency (Exposed) | 3.8 purchases/year |
| Frequency (Non-Exposed) | 3.2 purchases/year |
| Incremental Value | €960,000 (20,000 × 0.6 × €80) |
On average, customers exposed to Reelevant made 0.6 more purchases per year. With an average basket of €80, this represents €48 of incremental revenue per customer, totalling €960,000 across the 20,000 exposed customers.
What you CAN affirm
- Reelevant increased average purchase frequency by 0.6 purchases/year on this segment.
What you CANNOT affirm
- That this frequency gain persists beyond the measurement period.
- This case is strongly linked to average basket and annual revenue — avoid consolidating with those to prevent double-counting.
- Normalise frequency by the number of active days per customer.
- Verify significance with a bilateral t-test — frequency distributions are often asymmetric.
Challenge 8 — Average Basket (AOV) and Annual Spend
AOV (Average Order Value): Average transaction amount. Measures whether Reelevant generates higher-value purchases.
Annual Spend: Total revenue per customer over 12 months = Frequency × Average basket. Synthesises both effects.
Uplift AOV = AOV_exposed − AOV_control
Incremental Value (AOV) = Uplift AOV × Number of transactions (Exposed Population)
Annual Spend incremental = Annual_Spend_exposed − Annual_Spend_control
Incremental Value (AS) = Annual Spend incremental × Exposed Population size
Worked example
| Metric | Exposed | Non-Exposed | Uplift | Incremental Value |
|---|
| Average basket | €87.50 | €74.20 | +€13.30 | +€13.30 × 60,000 = €798,000 |
| Frequency/year | 3.8 | 3.2 | +0.6 | +0.6 × €80 × 20,000 = €960,000 |
| Annual Spend | €332 | €237 | +€95 | +€95 × 20,000 = €1,900,000 |
CRITICAL: Average basket and frequency compose Annual Spend. Never add all three — choose the final level (Annual Spend) for consolidation. Adding all three means counting the same value 3 times.
What you CAN affirm
- Annual Spend per exposed customer is €95 higher than for Non-Exposed customers.
What you CANNOT affirm
- That the Annual Spend increase is entirely attributable to Reelevant — it includes all downstream effects (confounders).
Challenge 9 — LTV (Customer Lifetime Value) Impact
Definition: LTV is the total revenue generated by a customer over their entire relationship with the brand. Reelevant can increase it by boosting frequency, basket size, or relationship duration.
| Approach | Description |
|---|
| Observed LTV | Measured on a cohort over 12–36 months. Direct Exposed vs Non-Exposed comparison. |
| Projected LTV | Estimated from observed frequency, basket, and retention rate, projected over time. |
Simplified LTV = Annual Spend × Average customer lifetime (in years)
Incremental LTV = LTV_exposed − LTV_control
= (Annual_Spend_exposed × Lifetime_exposed) − (Annual_Spend_control × Lifetime_control)
Total Incremental Value = Incremental LTV per customer × Exposed Population size
Worked example
| Metric | Exposed | Non-Exposed |
|---|
| Annual Spend | €332 | €237 |
| Customer lifetime | 3.2 years | 2.8 years |
| LTV | €1,062 (332 × 3.2) | €664 (237 × 2.8) |
| Incremental LTV | +€398/customer | |
Each exposed customer generates on average €398 more over their lifetime. On a cohort of 10,000 customers, this represents €3.98M in total incremental LTV. Caution: this projection assumes observed effects persist.
What you CAN affirm
- Exposed customers show €398 higher LTV on the measured cohort.
What you CANNOT affirm
- That projected LTV gains will materialise — uncertainty grows with the projection horizon.
- LTV is a synthesis metric: if you use it for consolidation, do not also add frequency, basket, or retention.
- Bound customer lifetime via a survival analysis (Kaplan-Meier curve) if possible.
- Distinguish observed LTV (reliable) from projected LTV (growing uncertainty).
- Incremental LTV is only valid if both groups were followed for the same duration.
