Viral Coefficient
The viral coefficient is the average number of new users each existing user generates through invitations or shares. A coefficient above 1 means your user base grows on its own.
Key takeaways
- The coefficient is invitations sent per user multiplied by their acceptance rate.
- Raising invitation volume usually lowers acceptance, so the product can stay flat.
- The denominator choice changes the number more than most optimisation work does.
- Averages hide power users; check the median inviter before trusting the figure.
- A higher coefficient can coincide with lower revenue if invited users are worth less.
In depth
The viral coefficient is one multiplication: the average number of invitations a user sends, times the share of those invitations that turn into active users. Send four invitations at a fifteen percent acceptance rate and the coefficient is 0.6, meaning every ten users produce six more. Both inputs are averages across a defined population, so the denominator matters as much as the numerator: measuring it over all signups gives a different answer than measuring it over users who reached the point where sharing is offered.
The two inputs pull in opposite directions. Prompting for more invitations usually lowers the acceptance rate, because the extra contacts are weaker ties, so pushing invitation volume can leave the coefficient flat while annoying users and burning goodwill. Acceptance responds to who is invited and what the recipient sees, not to how many were asked. The coefficient also decays with saturation: the earliest cohorts invite their most receptive contacts first, so a number measured at launch overstates what later cohorts will produce.
Practitioners compute it per cohort and per segment rather than as one company-wide figure, because a coefficient of 0.2 overall can hide a segment sitting near 0.8. With a scorecard funnel the arithmetic is unusually clean: a completed quiz is an unambiguous acceptance event, so invitations sent from a result page and quiz completions attributable to those invitations give both inputs directly. Segmenting by result tier then shows which score bands actually recruit peers, which is where invitation prompts belong.
The coefficient says nothing about value. A user acquired through an invitation may be worth a fraction of one acquired through search, so a rising coefficient can accompany falling revenue. It also assumes invitations are the only path between users, which ignores word of mouth that leaves no trackable link and therefore never appears in the numerator. And because the number is an average, one power user with hundreds of contacts can carry a coefficient that no median user comes close to producing.
Example in practice
How to measure it
Record three counts for one cohort: users in the cohort, invitations they sent, and invited people who became active. Divide invitations by users to get invitations per user, divide activated invitees by invitations to get acceptance, and multiply. Keeping the two factors visible separately is the point; a coefficient reported alone cannot tell you which half moved.
Alongside the coefficient, track the share of users who send at least one invitation and the median number sent by those who do. Those two describe the distribution the average conceals. Then compare the retention of invited users against users from other sources: if invited users retain worse, a rising coefficient is producing volume you will lose again, and the number should be read as a cost signal rather than a growth signal.
Common mistakes
The frequent error is computing the coefficient over the wrong population. Dividing by everyone who ever signed up, including dormant accounts that never saw a share prompt, drags the number toward zero and makes every improvement look negligible. Define the population as users who reached the sharing step, report that alongside how many users reach it, and you separate a weak invitation mechanic from a weak path to the mechanic.
The second mistake is quoting a single lifetime coefficient. Because early adopters exhaust their most receptive contacts first, a figure averaged across all history flatters the present and hides the decay. Recompute it for each monthly cohort and watch the trend line instead of the level. If recent cohorts sit well below older ones, the mechanic has not broken; the reachable audience has simply thinned out and needs a new entry point.
Frequently asked questions
How do you calculate the viral coefficient?
Multiply the average number of invitations each user sends by the conversion rate of those invitations. For example, 5 invites at a 10% conversion rate gives a viral coefficient of 0.5.
Can a quiz funnel improve my viral coefficient?
Yes. A scorecard quiz that delivers a personalized, shareable result encourages finishers to invite peers. Each share becomes a trackable invitation event you can attribute and optimize over time.
What is a good viral coefficient?
Anything above zero is doing useful work, and anything at or above one is rare enough that most teams should not plan for it. As a rule of thumb, a coefficient in the low tenths already changes acquisition economics meaningfully because it discounts every paid signup. Chasing a number above one usually distorts the product long before it is reached.
Is the viral coefficient the same as the K-factor?
They are the same arithmetic, used with different discipline. The viral coefficient is usually quoted without a time bound, while K-factor is stated per cycle, such as per week or per fourteen days. That bound matters, because the same number produces wildly different growth depending on how often the cycle repeats. When a figure has no window attached, ask what period it covers.
How do I raise the viral coefficient?
Work on the acceptance rate before the invitation count, since acceptance responds to what the recipient sees while invitation count mostly responds to nagging. Concretely: make the invited person's first screen show the value directly rather than a signup wall, invite through the channel the sender already uses, and remove any step that requires the recipient to guess what they were sent.
Why does my viral coefficient keep falling?
Usually saturation rather than a broken mechanic. Each cohort invites its most receptive contacts first, so later cohorts are working a thinner pool and the same prompt yields less. A second common cause is a widening top of funnel: if you start buying colder traffic, the average user has fewer relevant contacts to invite, and the coefficient dilutes without anything else changing.
Can I calculate a viral coefficient without invitation tracking?
Only approximately. You can estimate it by comparing signups attributed to no paid or organic source against signups from known sources, but that residual bucket also holds direct traffic and offline word of mouth, so it overstates the coefficient. Treat the estimate as an upper bound and add proper link tracking before you make budget decisions from the number.
Should I segment the viral coefficient?
Yes, and usually by acquisition source and by user type. Users who arrived through a colleague behave differently from users who arrived through an ad, and one segment often carries most of the sharing. Segmenting shows you where to place invitation prompts and where they are wasted effort. A single company-wide figure is fine for a board slide and useless for a decision.