What is K-Factor (Viral Coefficient)?
Also known as: Viral coefficient, K-value
What is K-factor?
K-factor is the number of new users each existing user brings in. The term comes from epidemiology, where it describes how many people one infected person infects. In growth marketing, it measures whether a product spreads on its own or needs paid distribution to survive.
The metric is the single best predictor of compounding growth. A K above 1 means the user base grows without ad spend. A K below 1 means every cohort decays unless paid acquisition refills it.
K-factor sits next to retention rate and user acquisition cost as one of the three numbers that decide whether a consumer app survives.
K-factor formula
The formula has two inputs.
K = i × cWhere:
- i = average invitations sent per user
- c = conversion rate of those invitations
A user who invites 10 friends with a 20 percent conversion rate has a K of 2. A user who invites 4 friends with a 25 percent conversion rate has a K of 1.
The formula looks simple. Measuring it in production is not. Invitations include direct sends, screenshots, link copies, and word-of-mouth that never touches a tracked channel. Most teams underreport invitations by 30 to 50 percent, per Andrew Chen's writing on viral growth loops.
Conversion rate also drifts. A friend who signs up but never opens the app does not really count. Most growth teams gate the conversion event on a meaningful action, not the install.
What K = 1 means
K equal to 1 is the break-even point of viral growth. Each user produces exactly one new user before churning. The base stays flat without paid spend.
K above 1 is the dream. Each user produces more than one replacement. The product compounds. Hotmail, early Facebook, and TikTok all crossed 1 at launch.
K below 1 is the reality. Most products land between 0.1 and 0.5. Growth slows as the cohort ages because invitations decay faster than they convert.
There is a second variable that matters as much as K. Cycle time. The days between a user joining and the user inviting others. a16z's growth content library frames it as the loop length. A K of 1.2 with a 3-day cycle doubles the audience every 18 days. The same K with a 30-day cycle needs 6 months for the same lift.
Real K-factors of viral apps
Three early-stage growth stories with the numbers behind each.
| Product | Era | K-factor | Mechanic |
|---|---|---|---|
| Dropbox | 2008-2010 | ~0.7 | Referral bonus. 500 MB free space per friend, both sides. |
| Slack | 2014-2015 | ~1.2 inside teams | Team invitations gated the product. One signup pulled in 5 to 50 coworkers. |
| 2009-2012 | ~1.5 | Phone-number import. Every contact already on the app appeared in chat. Zero friction. | |
| PayPal | 1999-2000 | ~1.0 | $10 sign-up bonus plus $10 per referral. Paid K. |
Sources: Sequoia's growth case studies and Andrew Chen's referral research.
Two patterns repeat. Each product wired the share mechanic into the core action of the app, not into a bolt-on referral page. And each one removed friction from the invite step. PayPal paid users to invite. WhatsApp read the address book. Dropbox bribed both sides with storage.
A bolt-on "invite a friend" page rarely clears K of 0.1.
How to engineer a higher K-factor
Five levers move K predictably.
- Wire sharing into the product, not next to it. Slack required team invites to be useful. Figma made shareable URLs the default state of a file. The share mechanic is the product, not a feature.
- Reduce invitation friction. A "copy link" button beats an email form. A pre-filled SMS beats a copy link. Address-book import beats SMS.
- Reward both sides. One-sided referrals convert at half the rate of two-sided ones, per Dropbox's published case data. Give the sender something. Give the receiver something. Costs go up, conversion goes up faster.
- Shorten cycle time. If the loop runs in days instead of weeks, K compounds 5 to 10 times faster. Push users toward the share moment in the first session.
- Pick a category where sharing is natural. Productivity tools that need collaboration ship with K above 1 by default. Single-player utilities almost never clear 0.3.
The biggest mistake is treating K as a marketing metric. K is a product metric. The PM owns it, not the growth lead.
K-factor in 2026
Pure viral growth is rare in 2026. App store gatekeeping, ad-tracking limits, and platform fatigue all dampen organic spread. Most consumer apps run a hybrid model: paid acquisition fills the top of the funnel, viral mechanics multiply each paid user.
The math of the hybrid is straightforward. A $25 paid CAC with a K of 0.4 becomes a blended CAC of $17.85. Each paid user brings 0.4 free users. The free users themselves bring 0.16. The series converges to 1 / (1 - K) free users per paid user, before churn.
Sequoia's recent growth playbook frames this as the new default. Pure viral is the outlier. Hybrid is the rule. Teams that ignore K and rely only on paid acquisition pay 30 to 50 percent more per active user than competitors who layer in even a modest viral loop.
The other shift is that platforms now reward engagement signals heavily. TikTok's For You Page treats shares and saves as ranking inputs. A piece of content with a K above 0.3 in the first hour gets pushed into wider distribution by the algorithm, which boosts K further. Algorithmic amplification adds a multiplier on top of the base K, which is new since 2020. See viral marketing for the format-level mechanics.
Real-world example with numbers
A fintech startup launches a budgeting app with a referral mechanic. Both sides get $5 in their account on a successful referral.
The first cohort numbers, 30 days after launch:
- 1,000 paid signups at $18 CAC, total spend $18,000
- Average invitations sent per user: 3.2
- Invitation-to-signup conversion: 14 percent
- K-factor: 3.2 × 0.14 = 0.448
- Free users from the paid cohort: 448
- Second-generation free users: 448 × 0.448 = 200
- Third-generation: 90
- Steady-state free users from $18,000 spend: ~810
Blended CAC drops from $18 to $9.95. Without the referral loop, the team would have spent another $14,500 in paid budget to acquire the same 810 users.
The K of 0.448 is unremarkable on paper. The compounding effect on CAC is not. That gap is why K-factor stays one of the most cited metrics in daily active users and monthly active users reporting at growth-stage companies.
A K below 0.1 is not worth instrumenting. A K above 0.3 changes the unit economics of the entire business.
Related terms
Frequently asked questions
What is a good K-factor?
Anything above 1 is exceptional. Most consumer apps land between 0.15 and 0.5, per Andrew Chen's analysis of growth-stage startups. A K of 0.5 still cuts paid acquisition cost in half because every paid user brings half a free user. Aim for 0.4 or higher in consumer. B2B rarely clears 0.2.
How is K-factor different from retention rate?
K-factor measures referral, retention measures stickiness. K asks how many friends a user invites and converts. Retention asks how often a user comes back. A product can have high K and low retention, like a viral quiz, or high retention and low K, like enterprise software. Both numbers matter for compounding growth.
Does the K-factor formula include cycle time?
The base formula does not. The full viral growth equation does. Cycle time is the days between a user signing up and them inviting others. A K of 1.2 with a 3-day cycle doubles the user base every 18 days. The same K with a 30-day cycle takes 6 months. Speed compounds.
Can paid acquisition replace a low K-factor?
Yes, at a cost. Sequoia's growth playbook frames this as a hybrid model. Paid acquisition fills the top of the funnel. Viral mechanics multiply each paid user by 1 plus K. A $20 CAC with K of 0.5 effectively becomes $13.30 per acquired user. Most modern consumer apps run this hybrid.
How do you measure K-factor in practice?
Pick a cohort. Count invitations sent per user over a fixed window, usually 30 days. Count how many of those invites convert to active accounts. Multiply the two numbers. Tools like Amplitude, Mixpanel, and Heap track invite events natively. Most teams underestimate K by ignoring indirect referrals like screenshots and link copies.