How Wise built 70% word-of-mouth growth through product obsession

Executive overview

Most companies treat word of mouth as a marketing outcome. Wise treats it as a product problem. Nilan Peiris, CPO at Wise, explains how 70% of their new users come through word of mouth — not through referral mechanics or brand spend, but by building a product so good it creates advocates.

The core mechanism: map NPS to actual referral behaviour, identify the product attributes that move people from passive users to promoters, then invest relentlessly in making those attributes 10x better than anyone else.

To drive word of mouth, you must give users an experience they didn't know was previously possible.

Why Wise had to grow through word of mouth

• Low margin on each transfer means limited budget for paid acquisition
• Money is a commodity — brand-led marketing for a cheaper product is structurally weak
• Word of mouth has near-zero distribution cost and scales to the full addressable market
• 70% of new customers in any given month discovered Wise from a friend; of ~1M quarterly new users, 700K came through word of mouth

Using NPS as a product metric, not a survey

• NPS correlates directly with referral behaviour: moving from a 6 to a 7–8 doubles invite rate; 8–9 doubles again; 9–10 doubles again
• This compounds: raising NPS increases the viral coefficient of the entire customer base, with far higher ROI than optimising conversion rate
• NPS comments — not just scores — revealed the consistent driver: price, speed, ease of use
• Wise emailed every NPS comment to the whole company weekly for the first four years
• Entering a new market at 5.9% vs a 6% competitor got users but no advocacy; dropping to 8–10x cheaper triggered word of mouth

The 10x product bar

• Most teams stop building once the product works; some then focus on conversion rate — that's not enough
• To get recommendation, users need an experience they didn't know was possible before
• This means solving structural, systemic problems that take years — not incremental improvements
• Ask: what is the theoretical minimum cost, maximum speed, minimum friction for this use case?
• Work backwards from the ideal rather than iterating forwards from the current state
• Examples: dropping transfer fees from 6% to 0.35%; enabling instant transfers; building central bank accounts in Singapore, Australia, and the UK

Cost-based pricing as a word-of-mouth engine

• Every cost — people, risk, partner fees — is allocated back to the transaction or route that generated it
• 20% of customers generate 80% of costs; those customers pay more, everyone else pays less
• Engineering investment targets the three cost buckets: people (customer service), realised risk, and partner fees
• As costs fall, pricing drops and Wise enters new market segments

Getting customers to help solve hard problems

• Singapore required face-to-face verification for every customer — Wise ran it as an unscalable on-the-ground operation, then got customers to lobby the regulator
• One year of lobbying produced the world's first eKYC licence in Singapore — selfie-based verification
• Mission authenticity matters: a rebrand email with no CTA, no sign-up button, just the mission statement, generated more new customers than any other marketing when recipients forwarded it
• People recommend partly for rational reasons (better product) and partly for emotional ones (belief in the mission)

Product marketing: closing the gap between value delivered and value perceived

• Users often believe they saved money but don't believe the specific number
• A PM iterated a visualisation graph (bank fees vs Wise fees) with customers in a coffee shop until the reaction was "I'm never using my bank again"
• Adding that graph plus a share button to the post-transfer success screen produced a 3x increase in sharing
• Instant transfers need to feel instant: a "whizzy animation" confirming arrival in the recipient's account also drove a measurable jump in referral rate
• Closing the delta between what was done and what is perceived = product marketing within the product

Conviction over experimentation

• You can't split-test your way to love; experiment-led PM with too many small tests produces slow, incremental gains
• Build conviction on what matters through qualitative and quantitative insight, then make bigger moves
• Wise's core strategic bet: lowest-cost, fastest, highest-quality platform will attract the world's volume — no experiment needed to validate that direction
• Teams still measure impact (pre-post analysis, holdout groups) but don't run significance tests before deciding whether to act
• Example: the currency converter was a founder-rejected idea the SEO team shipped anyway — now drives massive organic discovery

Team structure for a global-local product

• Wise sits between international banks (local tech stacks, deep local integrations) and pure tech companies (single global stack, no local infrastructure) — it has both
• Global product teams own overall KPIs and the shared code base; local/regional teams contribute directly via pull requests and own market-specific conversion metrics
• The domain model emerges from local feedback rather than being designed top-down
• Org structure evolved from fully autonomous KPI-driven teams to a squad/tribe model at ~30 teams
• Squads align to products (Wise Account, Business, Enterprise, regional); tribes provide light-touch strategy