A growth loop is a closed system where the output of one cycle becomes the input of the next — users bring more users.
The funnel model — awareness → interest → consideration → purchase — describes a linear pass-through of prospects. It's useful for measuring conversion but limited as a growth engine: every new customer requires roughly the same top-of-funnel investment. Growth loops describe systems in which output feeds input: a new user's action produces content, invites, links, data, or revenue that generates more new users. When a loop's coefficient (new users produced per new user acquired) approaches or exceeds 1, growth compounds without proportional marketing spend. Most durable consumer and PLG companies run on 2-4 identifiable loops layered together.
(1) Viral loop — a user's use of the product exposes new prospects. Personal invites (WhatsApp, Slack), product-generated share artifacts (Loom videos, Calendly links), or network-inherent virality (Zoom meetings including non-users). (2) Content loop — user activity generates content indexed by search engines that pulls new prospects. Reddit, Stack Overflow, Quora, Pinterest are archetypal. (3) Paid loop — new customers generate revenue that funds more paid acquisition. Only compounds when LTV substantially exceeds CAC and payback is short enough to reinvest. (4) Sales loop — customers refer other customers into a sales-led motion; account expansion funds new AE hiring. Common in B2B.
Start from the compounding step, not the top: what action, taken by a new user, produces a plausible next new user? Slack's compounding step is 'user invites teammate to join their workspace'; Loom's is 'user shares a video with a colleague who watches on Loom's site.' Trace the loop backward from there: what happens before the compounding step (onboarding, first value moment), and what removes friction from it (default-on invites, share buttons at the point of highest emotional payoff). Instrument the loop end-to-end so you can measure the coefficient. Not every product supports a viral loop — force-fitting one usually produces annoying invite prompts and no growth.
Coefficient (K) = new users generated per user in one cycle. Cycle time = how long one cycle takes. A K of 0.5 with a 3-day cycle produces very different growth from K=0.5 with a 30-day cycle. Loops with K > 1 are 'viral' in the technical sense (each user produces more than one new user), but K=0.3-0.7 loops are more common and still enormously valuable when they subsidize paid growth. Track both numbers per loop and per cohort; loops decay as market saturates and copycats appear.
Single-loop dependence is fragile. Airbnb layered: viral (guests becoming hosts), content (SEO on city-and-neighborhood pages generated from listings), and paid (retargeting funded by booking margin). Notion layered: viral (share to non-users), content (public pages indexed by Google), community (templates and tutorials created by users). When a mature company plateaus, it's usually because one loop's coefficient decayed and a new loop needs to be built to replace it — not because 'marketing needs to work harder.'
(1) Confusing a growth channel with a loop — 'we get users from paid search' is a channel; a loop would explain why the return from that channel compounds. (2) Optimizing loop steps in isolation — improving the share rate but degrading the sign-up rate downstream produces no net gain. Instrument end-to-end. (3) Ignoring negative loops — churn, complaints, or negative reviews are also loops (bad experience → bad reviews → fewer new users). Measure and manage. (4) Assuming loops persist forever — network saturation, algorithm changes (Facebook 2013, iOS 14.5), and copycats decay coefficients. Refresh loops every 12-24 months.
Investor directory · Fundraising library · Articles A–Z · Company funding database