The NPS Doubling Curve for Viral Growth
Optimize NPS instead of conversion, because each promoter tier doubles how many people they tell
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 92%
Wise discovered that word-of-mouth referrals don't rise linearly with satisfaction — they double at each NPS band. Moving customers from 6-7 to 8-9 to 9-10 each doubles the number of people they tell. Because a lift in NPS compounds the viral coefficient across the whole customer base, it delivers far more ROI than a one-off conversion-rate win.
Origin
Nilan Peiris and the Wise team developed this by overlaying their referral/invite data onto Net Promoter Score survey data in the company's early years. NPS itself originated with Fred Reichheld/Bain.
Core principles
- 01Word-of-mouth is the cheapest scalable channel that reaches an entire audience
- 02A conversion-rate win is a one-off hit; an NPS win compounds the viral coefficient forever
- 03Referral behaviour is non-linear — it doubles between NPS bands
- 04Optimize the metric that changes the slope of growth, not the intercept
How to run it
- 1
Instrument word-of-mouth as a real number
Run an attribution model overlaying referral data and cookie/web tracking, then survey a sample of customers asking how they heard about you and map that onto direct traffic. Ask directly — 'how did you hear about us' — rather than inferring.
Pro tip Stitch the acquisition question into the product flow itself (an interstitial pop-up), the way Airbnb did, so measurement is seamless and ongoing.
- 2
Run a recommend-to-a-friend NPS survey
Ask 'would you recommend us to a friend' on a 0-10 scale. Group into detractors (0-6), neutrals/passives (7-8), and promoters (9-10).
- 3
Overlay referral data on NPS bands
Cross-reference how many people each respondent actually invited/converted against their NPS band. Look for the doubling pattern: neutrals tell some, 8-9s tell double, 9-10s double again.
- 4
Make NPS the metric the business rallies behind
Instead of chasing conversion rate, set moving NPS (e.g. 30 to 50) as the north star, because it multiplies how many new customers each existing customer generates.
Watch out NPS comes from a low base in finance — banks average around -30 — so early market NPS of 20-30 is normal; the target for true advocacy is 70+, higher than the iPhone or Google Search.
In the wild
When Wise overlaid referral data on NPS survey data they saw very low invite and conversion rates from 1-6s. But 6s-to-7-8 group doubled the number of people they told, 8-to-9 doubled again, and 9-to-10 doubled again.
→ Wise reoriented its product strategy around raising NPS, and word-of-mouth grew to about 70% of new-user acquisition — roughly 700,000 of the million new users per quarter.
Common mistakes
Optimizing conversion rate as your growth metric
Moving conversion rate by 10% is a one-off hit; it doesn't change how many people each customer tells. Teams stop at 'it works' and polish conversion, missing the compounding channel.
Reading only the NPS number, not the comments
The score tells you where you are; the free-text comments tell you what to fix. Wise emailed every comment to the whole company weekly for years and that is how the price/speed/ease pillars emerged.
Is it for you?
Best for
Product and growth leaders at low-margin or commodity businesses that must grow through word-of-mouth rather than paid acquisition
Not ideal for
Businesses with high per-user margin and cheap, scalable paid channels where paid CAC economics dominate, or products people won't naturally discuss
From the transcript
“when you've got people from sixes to this seven and eight group they doubled the number of people they told eight to nine they doubled…”
“and nine to ten they doubled again”
“if you move the MPS like from 30 to 50 you increase the viral coefficient of your customer base so every customer that goes through…”
“we acquired 70 of the users that found out about wise last month through Word of Mouth”
From the episode
How to drive word of mouth
Nilan Peiris (CPO of Wise)