What is the k-factor in viral growth?

The formula

k = invitations sent per user × conversion rate of an invitation

If the average user sends 8 invitations and 15% of recipients sign up, k is 1.2. Each generation of users produces a slightly larger next generation, and the base grows without any acquisition spend.

At k below 1, each generation is smaller than the last. The series converges — you get a finite multiplier on whatever you acquired elsewhere, not perpetual growth.

What k above 1 actually requires

Both terms have to be high simultaneously, and they tend to trade off. Making invitations easier raises the count and usually lowers the conversion rate, because low-effort invitations reach people with weaker interest. Making invitations targeted raises conversion and lowers the count.

The products that sustained k above 1 for long typically had the invitation embedded in the core action rather than bolted alongside it. Hotmail's footer on every outgoing message is the origin story of the whole idea: the invitation was the product being used, so the count was enormous and cost the sender nothing. PayPal's referral bonus paid both sides in cash — expensive, deliberately so, and it bought a network that then defended itself.

Cycle time is the underrated variable

K tells you the multiplier per generation. Cycle time tells you how long a generation takes. Growth rate depends on both, and cycle time is usually the cheaper one to improve.

A user who invites on day one is worth far more than an identical user who invites on day thirty, because the compounding starts sooner and every subsequent generation shifts forward with them. Practical levers: prompt at the moment of first value rather than at signup, reduce the steps between deciding to invite and the invitation being sent, and make the invited person's first experience fast enough that they reach their own invitation point quickly.

Why k decays

No loop holds its coefficient. Early users are enthusiasts with dense, relevant networks. Later users have thinner ones, and increasingly the people they would invite are already members. Saturation shows up as a falling conversion rate on invitations even though nothing about the product changed.

This is why viral growth curves bend. The right response is a different loop, not a harder push on the old one.

Measuring it honestly

Attribute properly — a signup that would have happened anyway does not belong in the numerator. Measure over a window long enough to capture delayed invitations. And segment: a blended k hides that one segment has a k of 2 and the rest have almost none, which is a much more actionable picture than the average.

Dropbox's referral loop is often quoted for its headline growth figure; the more transferable lesson is that the reward was storage, so both parties got a better product and the invitation strengthened retention rather than trading it for reach.