One-Sided, Two-Sided, or Broken: Auditing a Referral Program's Weak Link
Objective: Audit a live referral program's incentive structure and recipient landing page against the lesson's two highest-leverage levers, and recommend which to fix first.
FirstCry's referral program currently gives only the referring parent a discount code; the friend who signs up gets nothing extra beyond the normal new-customer offer. Growth wants a recommendation before the next planning cycle.
Score the current program against the two-sided-incentive and recipient-landing-page findings from the lesson, and recommend the single highest-leverage fix.
Before you start
What you'll need
Free path (everything below is enough to finish)
Fast, shareable, no setup needed for a one-page recommendation
The process
2 steps
Step 01 of 02
Dropbox's two-sided program (both sender and recipient got 500MB) drove a K of 1.225 and scaled Dropbox from 100,000 to 4,000,000 users in 15 months; ReferralCandy's meta-analysis of 500+ programs found two-sided incentives produce 3x higher participation than sender-only rewards.
FirstCry's program gives the referrer a Rs.300 discount and the referee nothing beyond the standard new-user offer. Sender participation (the referrer share rate) is 8%. Using the 3x figure, what participation rate would a genuinely two-sided version realistically target, and is that alone enough to fix the loop?
Procedure
- Record the current structure: referrer gets Rs.300, referee gets nothing extra
- Apply the 3x participation multiplier: 8% x 3 = an estimated 24% participation rate under a two-sided structure
- Flag that participation (i) is only half of K; note that conversion (c) still needs a separate audit
Current: one-sided, Rs.300 to referrer only, 8% participation Projected: two-sided, matching Rs.300 credit to referee, ~24% participation (3x ReferralCandy benchmark) Open question: participation lift alone doesn't tell us what happens to conversion on the recipient side
Healthy
The team recognizes a participation lift on i still needs a c-side check before declaring the fix complete.
Unhealthy
The team ships the two-sided incentive and assumes K triples, without checking whether the recipient landing page converts those extra invites.
What this means
A one-sided incentive is a real, fixable gap, but it only raises i; it says nothing about whether more invited friends actually convert once they click through.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Leadership wants to greenlight the two-sided incentive as the whole fix | Pair the incentive change with a recipient landing page review before calling the project done | 5 min |
Step 02 of 02
Redpoint's 2024 benchmark found referral landing page conversion varies 3-10x across comparable SaaS products, while sender participation only varies 1.5-2x; the recipient page is usually the bigger lever, not the sender incentive.
FirstCry's referral link currently sends every recipient to the generic homepage, not a page mentioning the friend who referred them or the specific discount. Given the Redpoint finding, which of the two fixes (two-sided incentive vs. a personalized recipient page) should be built first?
Procedure
- Note sender-side variance ceiling: roughly 1.5-2x from incentive changes
- Note recipient-page variance ceiling: roughly 3-10x from landing page personalization
- Recommend building the personalized recipient page first, then layering the two-sided incentive on top
Fix A (two-sided incentive): expected variance ceiling ~1.5-2x on participation Fix B (personalized recipient page): expected variance ceiling ~3-10x on conversion Recommendation: ship Fix B first, it has the larger addressable range and doesn't require a discount budget change to test
Healthy
The recommendation sequences the higher-ceiling, lower-cost fix (the landing page) ahead of the incentive change that requires a finance sign-off.
Unhealthy
The team defaults to the incentive change first because it's the more visible, easier-to-explain fix, even though the data says the page has more headroom.
What this means
When two fixes compete for one planning cycle, the one with the wider verified variance range and the smaller budget ask should usually go first.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Two valid fixes are competing for the same sprint | Sequence by variance ceiling and approval friction, not by which fix is easier to describe in a meeting | 5 min |
Final deliverable
A one-page recommendation memo naming the highest-leverage fix, with the supporting variance comparison.
See a reference example
Squarespace referral audit (excerpt) Current: one-sided credit, generic recipient landing page Fix A ceiling: ~1.8x (two-sided incentive) Fix B ceiling: ~6x (personalized recipient page showing the referrer's plan tier) Recommendation: build Fix B this sprint; revisit the incentive structure next quarter once conversion data confirms the page fix worked
Success criteria
You're done when you can:
- Correctly applies the ReferralCandy 3x figure to the current participation rate
- Correctly compares the two fixes' variance ceilings from the lesson's Redpoint data
- Recommends a single sequenced fix, not both simultaneously