Proving the value.

Everything before this earns the right to a budget conversation. This chapter turns state movement into financial language a CFO accepts, because a believable number beats the biggest number every time.

The three value lenses

Every value report uses the same three categories, always shown with state composition:

• Value at risk: revenue sitting in L3, L4, and L7 that current behaviour says will leave.
• Value protected: revenue retained because a repair or recovery action worked.
• Value created: new revenue from conversion (L5), reactivation (L6), and deepening (L2 to L1).

The proof discipline

• Every claim needs a named baseline method: period comparison, matched cohort, holdout, or trend adjustment. No baseline, no claim.
• Every claim carries an evidence grade: A (high confidence), B (medium), C (directional only). Grade C never leads an executive ROI story.
• Shared credit is declared: when loyalty actions and campaigns overlap, say so, rather than double-counting.

The value tree

State movement links to money through a simple tree: states drive behaviours (retention, conversion, reactivation, recovery, spend mix, advocacy), behaviours drive revenue and margin, and a v1 formula exists for each branch. The playbook ships the formulas below; the operator supplies the inputs.

The six v1 formulas

Simple and auditable beats sophisticated and opaque. Use contribution rather than revenue where possible, match the horizon to the business cycle (typically 6 or 12 months), and use conservative assumptions where causality is partial.

1. Retention value protected (L3, L4, L7): customers retained or recovered due to intervention × expected contribution per customer over the horizon − intervention cost.
2. Conversion value created (L5): incremental converted customers × (first-transaction contribution + expected follow-on contribution within the horizon) − programme cost. Separate one-off promo conversions from quality conversions.
3. Reactivation value created (L6, L7): incremental reactivated customers × average reactivation contribution over the horizon − reactivation cost. Track L6 and L7 separately; trust-led reactivation is not generic win-back.
4. Recovery value saved (L4, high-value L7): value at risk before recovery − value lost after the recovery period − recovery cost. If value-at-risk estimation is immature, report recovery completion rate, post-recovery repeat rate, and sentiment restoration instead.
5. Margin-efficiency improvement (L2, L3, L5): change in contribution margin + reduction in incentive cost + reduction in cost-to-serve − action cost. A lower conversion rate at better margin can still be a win.
6. Advocacy value (L1, recovered L4 and L7): count referral and advocacy behaviour, track referred conversion where available, and report advocacy value separately from direct revenue unless attribution is strong. Never monetise advocacy aggressively without evidence.

Which states feed which value

• Value at risk concentrates in L3, L4, and L7: the revenue those relationships hold, weighted by the defection likelihood their movement history shows.
• Value protected is claimed when a repair or recovery playbook demonstrably holds L3, L4, or L7 revenue against a baseline.
• Value created is claimed for L5 conversion, L6 reactivation, and L2 deepening into L1.

Choosing a baseline method

• Period comparison: simplest and weakest; only where seasonality is stable.
• Matched cohort: compare actioned customers with similar unactioned ones; the pilot workhorse.
• Holdout: the strongest proof; hold back a random slice of a state from the playbook and compare.
• Trend adjustment: when history is short, adjust the expected trajectory and declare the assumption.

The monthly value report

Five blocks, every month: the value headline with its grades, what moved and why, action performance against value, the risk picture, and the asks. It feeds the Monthly Loyalty Review's decision footer rather than living as a separate ceremony.

The five-line executive narrative

Every monthly value story fits five lines: where value sits, what moved, what we did, what it was worth (with grade), and what we decide next. If it takes more than five lines, the story is not ready.

An illustrative example: "£1.8m of annual value sits in L4 and L7 (Grade B). Recovery playbooks moved 240 guests back to L1 or L2 this quarter. That protected an estimated £310k against a matched-cohort baseline (Grade B). Conversion tests in L5 created £90k (Grade A, holdout). We ask for the analyst hire to extend coverage to the second brand." Illustrative numbers, real structure.

The KPI stack

• Executive: value at risk, value protected, value created, L1 share of value, and the movement trend.
• Operating: state movement rates, SLA adherence, action completion, Provisional-to-Firm conversion.
• Proof: baselined outcome comparisons at 30, 60, and 90 days per state playbook.

Maturity, honestly

Value proof matures in stages: pilot (directional, small scope, Grade B and C), scale (matched baselines, Grade A and B), portfolio (full value tree reporting). Claiming portfolio-grade numbers from a pilot destroys the credibility the framework exists to build.

What good looks like

Leadership funds the next phase from the value report, not from enthusiasm. The loyalty conversation moves from cost centre to managed asset.

Chapter checklist

• Choose your baseline method per playbook before actions launch.
• Build the value tree with your finance partner, not around them.
• Draft your first five-line narrative from pilot data, graded honestly.

Worksheet 8Print this worksheet →

The value proof sheet.

Purpose: one row per value claim, so nothing reaches an executive without a baseline and a grade. A believable number beats the biggest number.

Configurable

The six v1 value formulas are in chapter 8. Use them as written; do not improvise formulas.

The sheet

Claim	State movement behind it	Baseline method	Formula used (chapter 8)	Inputs required	Evidence grade	Owner

Baseline methods: period comparison, matched cohort, holdout, or trend adjustment. Grades: A high confidence, B medium, C directional only. Grade C never leads an executive ROI story, and shared credit with campaigns is declared, not hidden.

The five-line executive narrative

1. Where value sits:
2. What moved:
3. What we did:
4. What it was worth (with grade):
5. What we decide next:

Completion check

No row has an empty baseline or grade column, every claim maps to a state movement, and the narrative fits five lines. If it needs more than five, the story is not ready.