Depending on the study and the category, somewhere between a fifth and half of purchase decisions trace back to a recommendation from someone the buyer trusts — a friend, a colleague, a review from a real person rather than an ad. That range is wide because word of mouth is notoriously hard to measure precisely, but the direction is consistent across every study on the topic: people trust people more than they trust brands. The problem is that most companies treat word of mouth as something that either happens or doesn't, rather than something with a mechanism you can actually design for.
A paid ad carries an implicit disclaimer: this company paid to say this about itself. A recommendation from a friend carries the opposite signal — this person has nothing to gain and chose to bring it up anyway. That's why referred customers tend to have higher retention and lifetime value than customers acquired through paid channels: they arrive with trust already built in, rather than trust the marketing has to manufacture.
Wharton professor Jonah Berger's research on virality (published in his book Contagious) breaks down what makes things get talked about into six ingredients, usually remembered by the acronym STEPPS:
Most products can't hit all six. The useful exercise is picking the two or three that are realistic for your category and building your messaging and product experience around those specifically, rather than trying to be generically "shareable."
Dropbox's early referral program is the most cited case for a reason: giving both the referrer and the new user 500MB of extra storage (worth roughly $1.25 in server cost at the time) turned into a reported 3,900% growth in sign-ups over 15 months. The mechanism worked because the reward was aligned with the product itself — more storage made the product better, not just cheaper.
Here's how the math works for a subscription business considering a similar program. Say your average customer is worth $50/month, and referred customers convert at 25% (higher than the 2–4% typical of cold traffic, because trust is already established). If you offer $25 credit to both parties for a successful referral:
The program pays for itself immediately if the reward is smaller than one month of customer value — which is exactly why most well-designed referral incentives are sized that way.
Consumer referral programs get most of the attention, but the underlying mechanics apply just as much to B2B, with a different delivery method. A software company might not run a "refer a friend for credit" program at all — instead, word of mouth shows up as a customer agreeing to a reference call during a prospect's sales process, a case study with their name and logo attached, or an unprompted mention in an industry Slack community or subreddit. These carry more weight than a consumer referral link because the recommender is putting their professional reputation behind the endorsement, not just spending a few seconds forwarding a link. The practical takeaway: B2B companies should build a habit of asking happy customers for a specific, low-friction favor (a two-line testimonial, a 15-minute reference call) at the moment satisfaction is highest, rather than waiting for word of mouth to happen unprompted.
Net Promoter Score (the "how likely are you to recommend us, 0–10" survey) is a proxy for referral intent, not referral behavior — someone can score you a 9 and never actually refer anyone. The more concrete metric is the viral coefficient, or K-factor:
K = i × c
i = average number of invites sent per existing customer
c = conversion rate of those invites into new customersA K-factor above 1 means each customer brings in more than one additional customer on average — true viral growth, which is rare and usually temporary. Most healthy referral programs run a K-factor between 0.15 and 0.5: a meaningful, compounding contribution to growth, not a replacement for other acquisition channels.
Not all word of mouth originates organically. Many brands seed it deliberately — sending products to micro-influencers (typically 10,000–100,000 followers) whose audiences trust their opinion more than a celebrity's reach would suggest. Micro-influencer campaigns often report engagement rates several times higher than macro-influencer or celebrity placements, precisely because the audience is smaller and more personally invested in that creator's recommendations. The distinction matters for budgeting: seeded word of mouth has a direct cost per placement, while organic word of mouth has zero marginal cost but zero guarantee it happens at all. Most mature word-of-mouth strategies run both — seeding to create initial momentum, and product/experience design (per STEPPS) to keep it compounding without ongoing spend.
Yes, arguably more than in consumer categories — B2B buyers weigh peer recommendations heavily because purchase mistakes are costly and visible internally. The mechanism looks different though: case studies, peer references during sales calls, and community reputation (G2 reviews, industry Slack groups) do the work that casual conversation does in consumer markets.
Product credit generally outperforms cash for subscription businesses because it deepens the relationship with your product rather than ending the transaction. Cash works better for one-time-purchase categories where there's no ongoing account to credit. Physical gifts add cost and fulfillment complexity that rarely earns back the extra goodwill.
You mostly can't track it precisely — that's the honest answer. The closest proxies are a "how did you hear about us?" field at signup, branded search volume growth over time, and NPS trends. Treat these as directional indicators, not exact attribution.