A mechanism where using the product itself generates more usage — unlike a classic funnel that stops once a user converts.
In practice
A classic funnel ends at conversion; a growth loop instead feeds its output back in as the next cycle's input. That's exactly this tool's own mechanism: every shared result (see Referral) potentially brings in a new visitor, who in turn completes their own test and shares it. Growth teams generally group loops into three families: content loops (generated content pulls in organic traffic), viral loops (sharing directly brings in new users), and paid loops (revenue generated funds the next round of acquisition).
The formula
Next cycle's input = this cycle's users × share who take the loop action × conversion of what that action produces
- Loop action
What a user does that creates the next user's entry point: shares a result (viral loop), publishes something search engines index (content loop), pays money that funds ads (paid loop). If nothing a user does creates an entry point for someone else, you have a funnel, not a loop.
- Conversion
How much of what the action produces becomes a new user: readers of a shared page who start their own, searchers who land on a user-generated page and sign up, ad impressions bought with revenue that convert.
- Cycle time
How long one turn takes — from a user entering to the users they generate entering. A loop with a lower ratio but a daily cycle outgrows a stronger loop that turns monthly.
The ratio output ÷ input is the loop's multiplier; for the viral loop it is exactly K. Above 1, the loop is self-sustaining; below 1 — the usual case — it multiplies whatever you feed it from outside, which is still the cheapest growth there is.
Worked example
This site's own loop, with the real mechanics
- Input: someone takes the Tour. Loop action: they share their result page — a link with a preview image showing their score.
- Conversion: a reader of that page clicks "take your own Tour" (the link carries a reference to the result that brought them) and completes fifteen questions. They are now an input, and the cycle turns.
- Suppose 20% of Tours are shared, a shared page is read by 12 people, and 6% of readers start their own: 0.2 × 12 × 0.06 = 0.144 new Tours per Tour. Multiplier 0.144; every 100 Tours from outside eventually yield about 117.
- Cycle time is hours, not months: a result is shared the day it is created, and read within days. That is why even a modest multiplier matters here — it compounds fast.
A loop is designed, not discovered: the preview image, the visitor's call to action and the reference on the link are each one term of the formula. Remove any of them and the loop is a funnel again.
Orders of magnitude
- Three families, three speeds. Viral loops (a user brings a user) turn in hours or days; content loops (user-generated pages that rank — public profiles, reviews, Q&A) take months to build and then run for years; paid loops (revenue funds acquisition) turn as fast as your payback period. Most durable companies run at least two.
- A loop multiplier above 1 is as rare as a viral coefficient above 1 — they are the same thing for the viral family. The realistic goal is a loop that turns a €500 CAC into an effective €350, not one that removes acquisition spend altogether.
- Loops decay: the network saturates, the content ages, the ad auction gets pricier. A loop that ran at 0.3 last year and 0.15 today is telling you where the next product work is.
How to improve it
- Draw it. One box per step, from a user entering to the next user entering, with a number on each arrow. If you can't draw it, you don't have one yet.
- Instrument the arrow you can't see. Usually it's conversion: who arrived through the loop's output, and did they become an input? A reference on every shared link and first-touch attribution answer that.
- Work the weakest arrow, not the whole loop. Doubling the share who share when readers never convert doubles nothing; fixing what the reader sees first might.
- Shorten the cycle before raising the ratio: getting an invitee to their own first value the same day is worth more than a small gain in share rate.
- Add a second family when the first plateaus: a viral loop plus a content loop (public results, glossary pages) reach different people at different speeds.
In the Tour
Does your product have a sharing or referral mechanism?
- Yes, built into the product20 pts
- Only in communication, not the product7 pts
- No0 pts
The Tour has no "growth loop" question because a loop is not a stage — it is what the Referral stage becomes when it is built into the product. The first Referral question asks exactly that: is there a sharing or referral mechanism in the product (20), only in your communication (7), or none (0)? A loop that exists "in communication" — a newsletter asking people to recommend you — has no action inside the product to run on, and cannot turn.
Questions people ask
Growth loop vs. funnel — which one should I use?
Both, for different questions. The funnel (AARRR) tells you where people drop between arriving and paying; the loop tells you whether the people who stay create the next arrivals. A team with only a funnel keeps buying the top; a team with only a loop can't tell which stage is leaking. Draw the funnel inside the loop.
What are the main types of growth loop?
Viral (a user brings a user — invitations, shared artefacts, collaboration), content (what users create gets indexed and found — profiles, reviews, public documents), paid (revenue from customers funds the acquisition of the next ones), and sometimes sales (a customer becomes a reference that closes the next). The famous product-led companies usually combine a viral loop with a content loop.
Can a B2B tool have a growth loop?
Yes, and the most common one is collaboration: the product is more useful with a colleague in it, so users invite colleagues, who invite theirs. Document tools, design tools and messaging all grew on it. The weaker B2B version is the output loop — reports, dashboards or links that get sent outside the team and carry the product's name.