You don't know which users would refer if asked
Your power users, high-NPS scorers, and frequent users are the most likely to refer. Ask them specifically instead of asking everyone.
When to use
You have engaged users and a way to identify them. You want more efficient referral outreach.
Hypothesis template
If we identify and ask [likely referrers] specifically, referral rate will increase by [target]% because targeted asks outperform broadcast requests.
Method
The problem: You show a referral prompt to everyone. Most users aren't engaged enough to refer. The prompt gets ignored and becomes noise.
How to do it:
- Score users on referral likelihood: usage frequency, NPS score, support interactions (positive), social shares
- Identify top 20% as "likely referrers"
- Send a personal ask: "You're one of our most active users — would you share Golden Gecko with one person who'd benefit?"
- Make the ask specific: "one person" is easier to act on than "share with your network"
- Provide a ready-to-send message they can customize
- Track: referral rate from targeted asks vs broadcast
Key insight: Asking one engaged user directly converts 10x better than showing a referral banner to 100 random users.
Success metrics
- •Referral rate from targeted asks
- •Conversion rate vs broadcast referral
- •Referred user quality
- •Referrer identification accuracy
Prerequisites
- User engagement scoring
- A personal outreach channel (email)
- A referral link or invite mechanism
Common pitfalls
- •Asking users who aren't engaged enough
- •Making the ask too broad ("share with everyone")
- •Not providing a ready-to-send message
- •Asking too frequently (once per quarter max)
Source: Common growth pattern. Targeted referral asks convert 5-10x better than broadcast prompts.
Suggested ICE scores
Ready to run this experiment?
Sign up free to use this playbook with step-by-step guidance and track your results.