Referral Program

/referral-program — customer-referral loop design

Design a customer-referral mechanic where existing users become a measurable growth channel. Distinct from affiliate / partner programs (different intent, different incentive math).

Doctrine inherited (Step 7 — 0626 rollout, locked 2026-06-04)

Output complies with:

• output-tenets.md, output-simplicity.md, marketing-psychology.md
• Step 6 calibration: see [[feedback_execution_doctrine_refinements_step6]]

Refinements applied to this skill:

Why bother — the benchmark

Per source data (cite-verified, MIT):

• Referred customers show 16–25% higher LTV than typical customers.
• Referred customers refer at 2–3× the rate of typical customers (compounding).
• Referred customers have 18–37% lower churn.

Net: a referral loop with even modest uptake is one of the highest-ROI growth investments available — but only if the trigger moment, share mechanism, and incentive math are all dialed.

The loop — 4 stages

Trigger Moment → Share Action → Convert Referred → Reward → (back to Trigger)

If any stage breaks, the loop stalls. Audit each stage independently when diagnosing a stalled program.

Workflow

Step 1 — Trigger Moment selection

The high-intent moment when a customer is most likely to share. Candidates ranked by typical conversion:

Sharp rule: never trigger on signup or trial-start. Customer hasn't felt the value yet; share fires hollow.

Step 2 — Share Mechanism — hierarchy by conversion

Order share mechanisms by conversion rate (highest first):

1. In-product native share (one-click within the product) — highest conversion.
2. Personalized link with name + context (e.g., "Matteo invited you").
3. Personalized email referral (sender adds note in their own voice).
4. Generic share via email / SMS (template).
5. Social share (LinkedIn, X) — high reach, low conversion per share.
6. Referral code (manually entered) — lowest conversion; reserve for offline channels.

Sharp rule: every program needs at minimum the top 3. Generic-code-only programs underperform by 5–10×.

Step 3 — Incentive structure

Three patterns:

Sharp rule: double-sided is the default. Single-sided fails when the referred customer has zero curiosity reason. Tiered is for $1k+/year products where compounding is worth the complexity.

Cost modeling: target referral cost-per-acquisition (CPA) ≤ 30% of LTV. Above that, the program drains margin. Below that, the program isn't generous enough to drive participation.

Step 4 — Launch checklist

• Trigger moment instrumented (event fires in analytics when the moment hits).
• Share mechanism live (in-product + personalized link, minimum).
• Incentive backend ready (reward delivery automated, not manual).
• Tracking attribution (referred-customer flag set at conversion; downstream LTV / churn dashboards segment by it).
• Email sequence (3-touch): Day 0 invitation, Day 7 nudge if no shares, Day 30 thank-you for any successful conversions.
• Anti-fraud rules (self-referrals blocked via email/IP match, duplicate accounts flagged).

Step 5 — A/B variations + optimization

Test variables in priority order:

1. Trigger moment (try alternatives, measure share rate per trigger).
2. Incentive amount (10%, 30%, 50% of LTV — find the elbow).
3. Share-mechanism mix (in-product only vs. in-product + personalized link).
4. Messaging copy (the referrer's invitation language).
5. Reward delivery cadence (instant vs. milestone-locked).

Worked example — Genesys-internal first

Use case: Genesys customers (GTM Engineers running our skills) refer peers from the niche.

• Trigger moment: when a Genesys customer has run 3+ skill-orchestrator chains over 30 days (clear sustained-use signal).
• Share mechanism: in-product banner ("Refer a GTM Engineer — both get $100 credit") + personalized link with the customer's name baked in.
• Incentive: double-sided $100 (referrer gets $100 credit on next renewal; referred gets $100 off first month).
• Cost check: assuming average LTV $5k, CPA target = $1.5k. Total $200 program cost = 4% of LTV. Sustainable.
• Anti-fraud: new account email must differ from existing customer's email + work-domain check.

After 90 days, productize to clients with viral-loop potential (ClientCo member→member, ClientCo adviser→adviser).

{Client / Genesys} referral program design

Trigger moment

• Chosen: {moment}
• Why: {evidence the moment correlates with willingness to share}
• Instrumentation: {analytics event}

Share mechanism

• Tier 1 (in-product): {design}
• Tier 2 (personalized link): {design}
• Tier 3 (email referral): {design}

Incentive structure

• Type: {single / double / tiered}
• Amount: {value, referrer + referred}
• LTV check: {CPA / LTV ratio}

Launch checklist

• Trigger event instrumented
• Share mechanism live
• Incentive backend automated
• Attribution + tracking
• Email sequence (Day 0, 7, 30)
• Anti-fraud rules

A/B test queue

• Test 1: {variable}
• Test 2: {variable}

KPIs

• 30-day target: {shares / conversions / referred LTV}
• 90-day target: {compounding rate, churn delta}

---

## Anti-patterns

- ❌ Trigger on signup or trial-start. Hollow share.
- ❌ Referral code-only program. Conversion stays at 0.5–1%.
- ❌ Single-sided when the referred customer needs a reason to click.
- ❌ Incentive cost > 50% of LTV. Drains margin.
- ❌ Manual reward delivery. Friction kills compounding.
- ❌ No anti-fraud — self-referrals erode trust + tank CPA math.

---

## Integration with other skills

- **Upstream:** `/icp-behavioural` defines who refers and why; `/product-messaging` defines the value-prop the referrer sells.
- **Downstream:** `/lifecycle-marketing` runs the 3-touch email sequence; `/analytics-tracking-plan` instruments the trigger + attribution events.
- **Companion:** `/ab-testing` runs the optimization tests after launch.

---

## Attribution

This skill adapts patterns from [`coreyhaines31/marketingskills/referrals/SKILL.md`](https://github.com/coreyhaines31/marketingskills/blob/main/skills/referrals/SKILL.md) (MIT license, © Corey Haines / Conversion Factory). Adapted to Genesys operator voice. Lane B: Genesys-internal first deployment.

---

## Final ship gate

Run `/premortem --output` before ship. See [`/premortem` skill](../../../meta/orchestration/premortem/SKILL.md) for the 5 execution domains (will-it-resonate / will-it-convert / will-it-stay-on-brand / will-stakeholder-push-back / will-it-degrade-over-time) and output template.

Trivial-case escape: `## Premortem\nNo failure modes — trivial change` satisfies the contract for genuinely trivial outputs.

---

## Persuasion & stickiness pass

Output complies with [persuasion-and-stickiness.md](../../../../rules/persuasion-and-stickiness.md) — Cialdini's 7 persuasion levers + Heath's SUCCESs. Deploy the 1-2 Cialdini levers that fit the reader's barrier (never all seven; every lever must be TRUE), run the SUCCESs diagnostic (Simple / Unexpected / Concrete / Credible / Emotional / Stories) over the near-final draft, then the rule's pre-ship gate.

---