"What's your K-factor?" is one of those questions founders learn to dread, because most answers are either "I don't know" or a number computed so loosely it means nothing. This guide fixes both problems: what K actually is, how to measure it honestly, and what to do when it is (inevitably) lower than you hoped.

The definition

The K-factor measures how many new users each existing user generates through viral mechanisms. The classic formula:

K = i × c

where:

i = the average number of invitations (or viral exposures) sent per user
c = the conversion rate of those invitations into new users

If the average user sends 5 invites and 10% convert, K = 0.5: every 100 users generate 50 more, who generate 25 more, and so on.

What different K values mean

K > 1 — true virality. Each generation is bigger than the last; growth is exponential without any acquisition spend. This is extremely rare and almost never sustained: even famous consumer apps typically hold K above 1 only for short windows.
K = 0.5 — every acquired user is worth 2 users in total (the geometric series 1 + 0.5 + 0.25 + … converges to 2). Your effective acquisition cost is halved.
K = 0.2 — a 25% amplification on all acquisition. Unspectacular, permanently valuable.
K ≈ 0 — your product has no viral surface. Common, and fixable.

For B2B SaaS, an illustrative healthy range is K between 0.15 and 0.4. Treat any claim of sustained K above 1 with suspicion — including your own.

The variable everyone forgets: cycle time

Two products with K = 0.5 can grow at wildly different speeds. The missing variable is viral cycle time (ct) — how long a full loop takes: user joins → sends invites → invitee joins.

The compounding over time t behaves approximately like:

Users(t) = U0 × (1 + K + K² + … + K^(t/ct))

Practically: a product with K = 0.4 and a 3-day cycle outgrows a product with K = 0.6 and a 30-day cycle within a quarter. When you invest in virality, halving cycle time is often cheaper and more powerful than raising the coefficient.

How to measure K without lying to yourself

Pick a cohort, not a blended average. Take all users who signed up in a given week. Blended numbers mix mature power users with fresh signups and inflate i.
Count exposures, not just formal invites. Shared links, collaborative artifacts, forwarded outputs and public pages are all viral surface. Instrument each.
Attribute conversions honestly. A new signup counts toward K only if you can trace it to an existing user's action — an invite token, a referral link, a shared-artifact landing. "Probably word of mouth" is not attribution.
Window it. Measure invites sent within N days of signup and conversions within N days of the invite. Unbounded windows make K drift upward forever.
Report K with its cycle time. "K = 0.3, ct = 6 days" is a real measurement. "K = 0.3" alone is half a number.

A worked illustration: of a 1,000-user cohort, 220 users share at least one artifact (22% share rate), sending an average of 4 exposures each = 880 exposures. Landing pages from those exposures convert at 9% → 79 new users. K = 79 / 1,000 = 0.079. Sobering, and typical for a first honest measurement.

Seven levers to improve K

1. Widen the sharing base (raise the % of users who share at all). In most products the share-rate is the weakest edge: 10–25% of users produce all viral exposure. Move sharing earlier in onboarding, and make the shareable artifact the default output of the core workflow.

2. Increase exposures per sharer — carefully. Batch invites (invite your team, not one colleague) and multi-recipient artifacts raise i. The guardrail: every exposure must carry value for the recipient, or you are converting your users' goodwill into spam.

3. Improve the recipient landing experience. The invitee should land in context: who invited them, what they were invited to, one obvious action. Illustratively, moving from a generic homepage to a contextual invite page can raise invite conversion from around 5% to around 15%.

4. Reduce signup friction on the viral path. Every extra field on the invitee's signup form taxes c. Viral signups justify a shorter form than organic signups — you already know their context.

5. Give both sides a reason. Double-sided value (the inviter gets collaboration or credit; the invitee gets immediate utility) beats one-sided referral bribes for durability.

6. Shorten the cycle. Prompt sharing at the aha moment rather than in week 3. Send invite reminders. Make the invited user's activation instant. Every day shaved off ct compounds.

7. Make the artifact publicly discoverable. A shared artifact that also lives on a public, indexable URL keeps recruiting long after the original share — a viral loop quietly becoming a content loop.

K-factor in B2B: think accounts, not just users

Consumer virality math assumes person-to-person spread. In B2B, two loops run at different altitudes and are worth measuring separately:

Intra-company K: users inviting colleagues into the same account. This drives seat expansion and retention more than new logos — enormously valuable, but do not count it as viral acquisition of new customers.
Inter-company K: exposure crossing organization boundaries — reports sent to clients, links shared with partners, artifacts posted publicly. This is the K that creates new accounts, and it is usually 5–10x smaller than intra-company K.

A B2B product can look impressively viral on blended numbers while acquiring zero new logos virally. Separate the two, and design distinct mechanics for each: team-invite flows for the first, outward-facing artifacts for the second. The outward loop is harder to build and worth far more per unit of K.

The mistakes that inflate or kill K

Incentive-stuffed invites attract users who came for the reward, not the product; they churn fast, and since K amplifies whatever you acquire, you amplify churn.
Optimizing K on a leaky product. Virality multiplies retention outcomes. Fix week-4 retention first; then virality multiplies something worth multiplying.
One-time measurement. K decays as early adopters (who over-share) give way to mainstream users. Re-measure by monthly cohort.
Confusing K with NPS. Willingness to recommend is not the same as instrumented, attributed referral. Only one of them compounds.

A realistic improvement roadmap

Do not attack all seven levers at once. A sensible quarter: first, fix attribution so the baseline K is trustworthy (two weeks of instrumentation work). Then run one experiment per month on the weakest edge — typically the recipient landing experience first (cheapest, fastest to test), then the share prompt's placement, then cycle-time reductions. An illustrative trajectory from teams that work this way: K moving from 0.08 to 0.20 over two quarters — which sounds modest until you notice it is a permanent 25% amplification on every future acquisition dollar, compounding forever at zero marginal cost.

From measurement to model

The real payoff of an honest K-factor is that it slots into a growth model. Combine K, cycle time, paid acquisition volume and retention curves, and you can simulate your user base 12 months out — and see, before spending a quarter of roadmap, whether raising the share rate from 22% to 30% matters more than cutting cycle time from 6 days to 3.

That simulation step is where spreadsheets usually give up and assumptions go untested. Growth Pilot's loop builder and Monte-Carlo simulator let you model exactly these edges — share rate, exposures, conversion, cycle time — and watch the distribution of outcomes before you commit the roadmap.