Real K-factor math. Find out if your product is actually viral.
You'd think K-factor needs a growth analyst — Astra computes it weekly from PostHog signups and tells you exactly which mechanic is driving (or not).
Astra calculates your referral coefficient (K-factor) weekly by mining PostHog signup events for referral source attribution. K = (invitations sent per user) × (conversion rate per invitation). She pulls the last 90 days of signups, classifies each by source (direct referral link, shared content, viral feature, paid acquisition), computes invitations-sent per active user, and tracks each invitation through to signup. Segmented K-factor matters more than blended — she shows K per cohort: net-new users, paid users, power users. Most B2B SaaS sit at K=0.1-0.3 (referral helps but doesn't drive growth). K=0.5+ is meaningful CAC reduction. K>1.0 = exponential growth without paid acquisition. The weekly Lark report shows current K, 4-week trend, K by source, and the top 2 levers to push it (typically: better share UX, stronger sender incentive, or removing friction in the recipient signup flow).
Set up properly: signup_source (direct/referral/viral/paid), referrer_user_id on signup. Backfill last 90 days from URL params + email links.
Per active user (logged in last 30d): how many shares, invites, public links sent. Average across cohort. This is the i in K = i × c.
Of all invitations/shares sent, what % resulted in a signup within 30 days. This is the c in K = i × c.
K for paid users vs free users, power users vs casual, new cohorts vs old. Different segments have very different K — averaging them hides the signal.
Weekly Lark: current K (overall + segmented), 4-week trend, K by referral source, top 2 levers to push it (share UX, sender incentive, recipient friction).
Weekly Lark report: K-factor (overall + segmented), 4-week trend, K by source, top 2 levers to push it.
Weekly Friday 4pm. Quarterly deep-dive on K trends.
We'll spin up your workspace, hand the prompt to Astra, and you see the answer in 60 seconds. Free.
Try this with AstraClear answers about wallet credit, usage, subscriptions, and how Tycoon charges for work.
K=0.1-0.2 is most B2B SaaS (referral helps but doesn't drive growth). K=0.3-0.5 is good (meaningfully reduces CAC). K=0.5-1.0 is excellent. K>1.0 = exponential growth. For consumer products K targets are higher (0.5+ to be considered viral). Astra benchmarks you against your category.
Then K is being driven by word-of-mouth alone, which is hard to measure precisely. She'll estimate it via 'how did you hear about us' signup field + organic search referral patterns. The recommendation is usually to ship a viral feature (see /ship-a-viral-feature) since adding share surfaces typically lifts K by 0.1-0.3.
K-factor is product-driven virality (anyone who uses the product creates new signups). Referral program ROI is incentive-driven (give $20, get $20). They're complementary metrics. K measures product strength; referral ROI measures program strength. Both can be improved independently.
Blended K hides the signal. Power users almost always have K=0.5-1.0 (they love it, they share). Free users almost always have K<0.1 (they're not invested). The real lever isn't always 'improve sharing' — it's often 'get more users to power-user behavior so the high-K cohort grows.'