Built-In Growth: Referral and Viral Loops

The cheapest user you will ever get is the one another user brings you. When a product grows partly by its own use, every paid customer quietly subsidises the next, and your blended cost to acquire falls as you scale instead of rising. That is the prize. The catch is that most features founders call viral are not, and the arithmetic that decides it, the viral coefficient, is unforgiving and easy to fool yourself about.

The cheapest user is one another user brings

Referred users are not just cheaper; they tend to be better. A widely cited study by Schmitt, Skiera and Van den Bulte in the Journal of Marketing tracked referred and non-referred customers at a bank over years and found the referred ones churned less and were worth more over their lifetime. That makes intuitive sense: a friend pre-qualifies the fit before they ever sign up, and arrives with built-in trust no advertisement can buy. So word of mouth improves two levers at once: it lowers acquisition cost and it raises retention, which is why it sits at the heart of built-in growth rather than off to the side as a nice extra.

The unforgiving math of the viral coefficient

The viral coefficient, usually written k, is how many new users each existing user brings: invitations sent times the share who convert. Below 1, growth does not run away on its own, but the loop still amplifies everything else you do. The cumulative effect of a loop with coefficient k is roughly a multiplier of one divided by one minus k, so a small k still quietly multiplies your paid acquisition for free.

At k of 1 or above, a cohort more than replaces itself and growth becomes self-sustaining, the rare true virality of a handful of products in a generation. Almost no business should plan on it. The realistic and still transformative goal is a k somewhere below one that turns paid acquisition from a treadmill into a lever.

Three kinds of loop, and why most fail

Built-in growth comes in three flavours. Incentivised referral rewards a user for inviting a friend, the model Dropbox made famous by giving both sides extra storage, which worked because the reward was the product itself. Casual virality rides ordinary use that is inherently visible: sharing a document, sending a payment, posting a creation exposes the product to non-users. Network effects are different in kind: the product gets better for everyone as more people join, so users recruit others out of self-interest, not for a coupon.

Most referral features fail for one reason: they are bolted on. A loop only works when it rides a moment the user already wanted to act on, sharing something they are proud of, inviting the person they actually need on the other side. An incentive can accelerate a loop that already exists; it cannot manufacture one a delighted user would not start on their own. If people are not already telling friends without being paid, a referral bonus mostly buys you fraud and freebie-hunters, not growth.

How we read it

When we look at a consumer deal claiming organic or viral growth, we want it measured, not asserted. What share of new users arrive without paid attribution, and is that share rising or falling as the company scales? What is the actual k, and on which action does the loop fire? Are referred users retaining better than paid ones, or worse, which would suggest the incentive is attracting the wrong people? A real loop shows up as a blended cost that falls with size. In a market the size of Taiwan, that discipline matters more, not less: a smaller addressable base means a loop has fewer people to spread through, so the honest k is often lower than the founder hopes, and a measured loop that cuts blended cost is worth more than a viral story that does not survive the numbers. As Andrew Chen has written at length about growth loops, the durable version is built into the product's core use, not stapled to it, and it connects directly to lever two in the four levers and to the channel mix in where consumers come from.

The takeaway

Built-in growth is the closest thing consumer has to free distribution, and it rewards the same thing every other lever does: a product good enough that people genuinely want to bring others in. Chase the loop that rides real delight, measure it honestly as a coefficient, and treat any number below one as an amplifier rather than an escape from paying for growth. You cannot buy word of mouth, but you can build the product that earns it.