Find the Break: Auditing a Referral Program's Funnel Data
Objective: Given real funnel-stage numbers for a live referral program, identify which of the two highest-failure nodes (the share moment or tracking/attribution) is actually broken, and separate that from a reward-design problem.
You're a growth marketer at Mailchimp. The referral program has a healthy volume of eligible users but a share rate and conversion rate the team suspects are underperforming, and leadership wants the actual bottleneck, not a guess.
Walk the funnel stage by stage against the lesson's two highest-failure nodes and recommend a fix scoped to the real bottleneck, not the reward.
Before you start
What you'll need
Free path (everything below is enough to finish)
A 6-stage funnel with simple ratios needs no dedicated analytics tool to diagnose
Paid upgrades (optional, faster/deeper)
Once the broken node is confirmed, a standing funnel report catches regressions after the fix ships
No access? Re-run the Google Sheets funnel export weekly
The process
2 steps
Step 01 of 02
The lesson's funnel diagram identifies two highest-failure nodes: the share moment never surfacing, and tracking failing on mobile (particularly iOS Safari stripping link decoration). Both are engineering problems, not incentive problems.
Eligible users to share-prompt-seen conversion is 91%, but share-prompt-seen to link-shared is only 4%. Click-to-signup conversion, when it's tracked, is a healthy 22%. Which node is actually broken?
Procedure
- Import referral-funnel-export.csv with raw counts per stage
- Compute stage-to-stage conversion rate for all 6 transitions in the funnel
- Compare the two lowest conversion rates against the lesson's two highest-failure nodes (share moment, tracking)
- Rule out reward size as the cause if the drop happens before the reward is ever seen
Mailchimp referral funnel, stage conversion (excerpt) Eligible -> Prompt seen: 91% Prompt seen -> Link shared: 4% <- broken Link shared -> Friend clicks: 68% Click -> Signup: 22% Signup -> Tracked correctly: 95%
Healthy
Prompt-seen to link-shared conversion sits in the 15-30% range typical of an unburied share moment.
Unhealthy
A 4% prompt-seen to link-shared rate, while every stage after the share itself converts normally.
What this means
The drop happens at the share action itself, before any reward or tracking logic runs, so this is a share-moment placement or friction problem, not a reward-size or attribution problem.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| 91% see the prompt but only 4% act on it | Test moving the share prompt to fire immediately after a completed campaign send, not from a settings-page link | half day |
| Team's first instinct is to raise the reward amount | Do not touch the reward until the share-moment placement test has run, the drop happens before the reward is seen | 5 min |
Step 02 of 02
Rewards delivered immediately after the referred friend's first action outperform delayed rewards by 2.6x in repeat referral behavior (Extole, 2023), because the immediate reward confirms the system works.
Tracked signups convert to reward issuance in 95% of cases, but the reward currently pays out on a 14-day delay after signup. Repeat-referral rate among rewarded users is only 9%. Is reward timing a secondary issue worth fixing here?
Procedure
- Filter to users who received a reward
- Compute what percentage of rewarded users share a second referral link within 30 days
- Compare against the lesson's 2.6x immediate-vs-delayed benchmark
Rewarded users, repeat-share behavior Current (14-day delayed reward): 9% share again within 30 days Extole benchmark (immediate reward): ~2.6x higher repeat rate expected
Healthy
Repeat-share rate among rewarded users tracks close to the immediate-reward benchmark once delivery is fast.
Unhealthy
A 14-day gap between the referred signup and the reward, with a repeat-share rate well below benchmark.
What this means
Reward timing is a real, secondary lever here, worth fixing after the share-moment placement, not instead of it.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Rewards pay out 14 days after the qualifying signup | Move reward issuance to fire immediately on qualifying signup confirmation, as a second-priority fix behind the share-moment test | dev ticket |
Final deliverable
A funnel-stage conversion table with the broken node identified, evidence for why it's the share moment (not the reward or tracking), and a prioritized two-item fix list.
See a reference example
Freshworks partner-referral funnel, diagnosis (excerpt) Eligible -> Prompt seen: 88% Prompt seen -> Link shared: 3% <- broken node Link shared -> Click: 71% Click -> Signup: 19% DIAGNOSIS: Share moment is buried in account settings, 3 clicks deep. Not a reward or tracking issue. FIX 1 (this week): Surface share prompt in the post-onboarding success screen. FIX 2 (after Fix 1 ships): Move reward payout from 10-day delay to immediate.
Success criteria
You're done when you can:
- Computes conversion rate for every funnel stage, not just the two suspected weak points
- Correctly rules out reward size as the cause when the drop happens before the reward is seen
- Separates the primary fix (share-moment placement) from the secondary fix (reward timing) and orders them