Word-of-mouth in numbers: ROI, conversion, trust

Who invented structured word-of-mouth? Tupperware parties. True story.

In the 1950s, Brownie Wise built a retail empire on a disarmingly simple principle: a woman invites her friends over and shows them products she actually uses. No media, no ad buying, a chain of trust, and a commission.

Seventy-five years later, we can do better. But before doing better, we need to measure, and that is where things get complicated: word-of-mouth drags around a stockpile of statistics copied from blog to blog, with no traceable source.

Here is what remains after sorting: the figures I was able to trace back to their source, and the ones I discarded.

Does word-of-mouth really drive sales?

Yes, and the order of magnitude is higher than what most media plans assume. Word-of-mouth is the primary factor behind 20 to 50 % of all purchase decisions, and it generates more than double the sales of paid advertising in categories as diverse as skincare and mobile phones (McKinsey).

You have probably come across a more dramatic version: “word-of-mouth generates five times more sales than advertising.” It circulates everywhere, attributed to McKinsey, and appears in zero McKinsey publications. A number distorted by continuous repetition.

The documented version is more than enough to justify a budget, with one nuance: it measures the incremental impact of recommendations on sales, not a ROAS comparable to your Meta campaigns. Rebuilding a shared metric is the topic of How to manage word-of-mouth like paid.

Who actually generates these recommendations?

A minority of people, with an outsized effect. McKinsey estimates that 8 to 10 % of consumers are “influentials”: they generate three times more recommendation messages than average, and each of these messages has four times more impact on the recipient’s decision (same source).

Translated to a customer base of 50 000, that is 4 000 to 5 000 people whose opinions carry structurally more weight. None of them will reach out to your marketing department on their own, and none have an account followed by 80 000 people, which explains why they are absent from every influencer plan: see “Alternatively Influential”: influence without an audience.

Why is a referred customer worth more than they seem?

Because they arrive with trust already built in, and they pass it on.

First, on trust: 88 % of consumers state they trust recommendations from people they know more than any other advertising format (Nielsen, Trust in Advertising 2021). 50 % more trust a recommendation than the lowest-ranked formats in the same study: online banners, mobile ads, SMS, and search engine ads.

Next, on value. A Wharton study on a retail banking referral program shows that referred customers have about 18 % higher retention probability than others, with a higher lifetime value, and that they refer 31 to 57 % more than customers acquired through other channels (Schmitt, Skiera & Van den Bulte, Wharton).

This is the point that most CAC calculations miss: they compare two customers with different costs, when in fact they are two customers with different values. An identical CAC across both channels is therefore not an equivalent return.

One final clue, right before your eyes for years: customer reviews. The purchase probability for a product with five reviews is 270 % higher than that of the same product with no reviews, ranging from 190 % for low-priced products to 380 % for higher-priced items (Spiegel Research Center, Northwestern University). Word-of-mouth in its most passive form.

What reward actually triggers a recommendation?

Cash, paid fast. 68 % of consumers prefer cashback over points systems, compared to 32 % who prefer points (PayPal / Reach3 Insights, study conducted in July 2025 among 1 068 US adults).

The same study adds three valuable figures: 81 % of consumers say rewards influence their purchasing behavior, 59 % say they make them more loyal to a brand, and 49 % say they make them more inclined to recommend it. A well-designed reward produces referrals, on top of retention.

For Gen Z, 55 % admit they would switch brands for a cheaper or better deal (PayPal survey covered by UK media). Loyalty out of habit is holding up less and less.

The form of the reward matters just as much as its amount. A voucher is a discount on a future purchase with you: it is not a reward, it is an invitation to spend again. Cardlytics makes this case through immediate gratification, without quantifying it (Cardlytics), and the full reasoning is in Cash vs points: why instant rewards win.

The phrasing itself carries real weight. Wharton research shows that the way the incentive is framed can increase both the number of referrals sent and the conversion rate of the resulting new customers by 86 % (Knowledge@Wharton). Same reward, same customer base, two completely different outcomes.

The figure that spoils everything: under 5 %

Here is the downside. In standard referral programs we observe at Frak, fewer than 5 % of customers actually activate their link.

The preceding numbers describe the potential of word-of-mouth. This one describes what brands actually capture. The gap comes from the mechanics offered to customers, accurately summarized here:

“If your friend orders, you get a 10 € voucher valid on your next purchase, but be careful not to lose it, because you will need to enter this code, on a hidden page, during a full moon.”

Complicated, unrewarding, and above all, untracked. The complete anatomy of this failure is in Why your referral program is not working.

Then there is a measurement problem. Even when it works, word-of-mouth travels through private messaging and lands without a referrer: direct traffic accounts for around 37 % of visitors, filed under “direct” or “organic,” meaning nowhere. The details are in Dark social: why your word-of-mouth is invisible in Analytics.

Across the first brands we onboarded, once the link is trackable and the cash reward is paid on purchase, we observe an average 26 % decrease in customer acquisition cost, and up to 40 % of sales coming from recommendations. Our numbers, on our customer base, not an industry average.

What these numbers do not tell you

An article that piles up statistics without scrutinizing them is part of the problem. Here are the limitations, and what I removed from this page.

What I cut. Four very popular figures were dropped from this version: referred leads converting “3 to 5 times higher”, “+37 %” retention, “+16 %” LTV, and “+200 %” adoption. All are attributed to Deloitte or Wharton across dozens of websites, yet none link to a verifiable publication. The exact same issue as the “5× McKinsey” claim, and no reason to fall for it twice.

Self-selection bias. Customers who refer are often already your best customers: part of the outperformance is attributable to them, not the channel. The Wharton study isolates both effects cleanly, which explains why it reports 18 % where blogs claim 37 %.

Attribution. Measuring untracked word-of-mouth relies heavily on self-reporting and modeling. As long as your recommendations do not carry an identifier, your own figures will be estimates.

Age. The McKinsey study dates back to 2010, Wharton to 2011, before the collapse of trust in advertising. This trend tends to favor word-of-mouth, but it remains a limitation to acknowledge.

The most uncomfortable takeaway in all of this: generate your own data.

Noise is free. What costs money, and delivers returns, is a structured, tracked, and rewarded channel.

A referred lead arrives pre-qualified, because someone they trust handled the vetting for you. You do not pay for attention; you collect intent. You just need to be able to prove it on a dashboard.

Frequently asked questions

Is word-of-mouth more profitable than paid advertising?

In terms of generated sales, McKinsey finds that word-of-mouth produces more than double the sales of paid advertising in certain categories, and serves as the primary factor behind 20 to 50 % of purchase decisions. A direct comparison with ROAS remains imperfect, as the two metrics do not cover the exact same scope.

How do you measure the ROI of word-of-mouth?

By giving each customer a trackable link and tying the sales it generates to a predetermined cost. Without an identifier, recommendations arrive without a referrer and fall under direct traffic, which represents around 37 % of site visitors. With tracking, the channel becomes directly comparable to a paid campaign, line by line.

How many customers actually activate a referral program?

In traditional setups we observe at Frak, fewer than 5 % of customers activate their link. The gap between this and the channel’s potential comes down to friction: delayed rewards, codes to enter, poor visibility, and no tracking. Friction explains this baseline, and it can be fixed.

Should you reward customers with cash or vouchers?

68 % of consumers prefer cashback over points systems, compared to 32 % who prefer points, according to PayPal and Reach3 Insights. A voucher ties the reward to a repeat purchase, turning it into a discount. Cash is perceived as a net gain, which drives higher participation.

---

I am Virginie Maire, co-founder of Frak Labs, which turns your customers into a scalable, profitable, and authentic acquisition channel. A dedicated entrepreneur and mother of two, I have spent 20 years navigating media, social networks, influencer marketing, and e-commerce… and I am still having just as much fun!