Growth loops: how to find the loop your product already has

A growth loop is a cycle where users' actions bring in the next users, and most products already have one. Picture a founder whose sign-up report shows "direct" and "word of mouth" as the two biggest rows.

Rows like "direct" and "word of mouth" often hide the loop. Someone opened a shared file, saw a public page or read a review, then signed up. The tracking lost the trail. When paid acquisition costs more each quarter, a board member says "you need a growth loop." Before you design a new one, find the one already in your data. The discipline is the same as finding the bottleneck before you scale spend.

What a growth loop is

A growth loop is a closed system: users come in, take an action, and that action brings in more users. Reforge's 2018 article "Growth Loops are the New Funnels" puts it this way: "Loops are closed systems where the inputs through some process generates more of an output that can be reinvested in the input."

A funnel runs in one direction. Reforge sums up its logic as "Put more in at the top, get more out at the bottom." In a loop, the bottom feeds the top. Reforge keeps the funnel but gives it a smaller job: it "helps explain a specific step within a Growth Loop, but misses the larger picture of the loop itself."

Reforge's example is Pinterest's loop, as Reforge described it in 2018, not necessarily how Pinterest grows today. A user signs up or returns and activates on relevant content. They save or repin content, which gives Pinterest quality signals. Pinterest distributes that content to search engines. A searcher finds it and signs up or returns, which starts the next turn.

Loops do more than bring in new users. Reforge lists loops that serve "new users, returning users, defensibility, or efficiency." This guide stays with the first two, because those are the ones that show up in your sign-up and activity reports.

Why a loop beats a one-off push

A loop compounds, and a one-off push does not. Reforge shows the difference with two initiatives. Initiative A delivers "500 new engaged users this week, but nothing after." Initiative B delivers "20 new users in week one, 22 in week 2, etc (growing 10% WoW) for every week going forward."

We ran the cumulative totals from those two lines. By our arithmetic, B passes A's 500 users in week 14, at 559. By week 20, B has brought in 1,145 users while A still sits at 500.

The catch sits in Reforge's own wording: "for every week going forward." B only wins if the loop keeps turning at that rate. If the loop stops returning users after week 6, B never catches the push. Measure the turn instead of assuming it.

Tactics also wear out in a way loops tend not to. Reforge writes that "As they get copied, effectiveness decreases and always trends to zero". A loop built on your own product's action is harder to copy. A competitor would need your users and the output they create.

The four loop types you are most likely to have

Most products run some version of one of these four loops. Read the table as a checklist against your own product.

Reforge argues that "The fastest-growing products are better represented as a system of loops". You don't need the whole system on day one. You need to know which loop already carries weight.

A channel is not a loop unless something returns. Ads paid from the seed round are a channel. Ads paid from the margin of earlier customers are a paid loop, because revenue from one cohort buys the next.

How to find the loop you already have

Start with a test from Reforge: "ask five different people to whiteboard the answer to a seemingly simple question: How does your product grow?" One result Reforge lists: "Everyone has a different answer". The audit below replaces those opinions with data. Run it on the last 90 days of sign-ups.

1. Where did new users come from? Pull sign-ups by source for the last 90 days. Keep the rows you usually skip, such as "direct", "other" and "word of mouth".
2. Which sources did an existing user or customer create? Mark every source that exists because someone already used the product. That covers invites, shared files, public pages, reviews, referrals named in sales calls, and ad spend paid from revenue.
3. What action produced that output, and what share of users take it? Name the exact event, such as "sent an invite" or "published a page". Then count the share of new users who take it within 14 days of sign-up.
4. How long from one user's sign-up to the sign-up they caused? Match referred users back to the user who brought them. The median gap is your rough cycle time.

Add a "how did you hear about us" field at sign-up, with options that match your likely loop sources. Put referral parameters on every invite link, shared file and public page, so the source survives the click. Without them, question 4 has no answer.

Write your strongest candidate as four boxes, as in the worksheet above. If you can't fill the fourth box with real sign-ups, you have a channel. Keep looking, or treat that loop as something to build later.

Measure the loop with two numbers

Judge the loop on how much new input each turn returns and how many days one turn takes. Both appear in David Skok's 2009 post on viral marketing, which models invite loops. Applying them to content, paid and review loops is our generalization, not his.

The first number is the viral coefficient. Skok's method is to "multiply the number of invitations by the conversion rate". For any other loop, use the same logic: new users returned per 100 users who enter the loop. A content loop would count sign-ups from public pages per 100 users who publish.

Skok also writes that "The Viral Coefficient must be greater than 1 to have viral growth". That line is about growth from invites alone. A loop below 1 still helps. Every user it returns is a user you didn't buy, so it lowers the cost of the users you do buy.

The second number is cycle time. Skok's example: "after 20 days with a cycle time of two days, you will have 20,470 users, but if you halved that cycle time to one day, you would have over 20 million users". In his model, cycle time matters more than the coefficient and should be as short as possible.

For your own loop, cycle time is the median number of days from a user's sign-up to the sign-up they caused. Shorten it by moving the output action earlier. Ask for the invite during setup, not in week three.

Measure both numbers per weekly or monthly cohort, never as an all-time average. Reforge frames the core question as "How does one cohort of users lead to another cohort of users?" Loops that run on returning users also need people who stay. Check the retention curve of the same cohort before you trust the loop.

Fix the weakest step first

Fix the weakest step inside the loop before you buy more input. Each loop contains its own small funnel: input, activation, output action, return. The weakest of those steps caps the whole cycle. More top-of-funnel spend only pushes more users into the same leak.

The weak step is usually activation or the output action. Users who never reach the product's value don't invite, publish or review anyone. Check your activation rate before you redesign the invite flow.

On Masa, Kreto found the mandatory sign-up wall was the biggest drop-off point. Moving registration after the core experience was the primary driver of sign-up conversion rising from 32.7% to 63%. The same project added a review prompt after positive engagement, which is the action step of an app store rating loop. Fixes like these cross product and marketing, so Kreto's growth marketing work treats both as one funnel.

List every candidate fix for the weak step and rank them with ICE scoring. Then pick one loop to work on. Reforge observes that the fastest-growing products are typically powered by one or two major loops that change over time. Reforge warns that drawing a ton of loops typically means "you just have a ton of low-powered loops that aren't sustainable".

Common mistakes

• Drawing ten loops. A whiteboard full of arrows looks thorough. Each loop needs an owner and two numbers, so map them all, then commit to one.
• Copying a famous company's loop. Their loop runs on their users and their output. Yours has to run on an action your users already take.
• Counting a channel as a loop. If nothing from the output returns as new input, you have a channel. Spend that never comes back does not compound.
• Measuring output but not cycle time. A count of invites sent says little if invitees sign up three months later.
• Splitting the loop by funnel stage. Reforge describes marketing teams that bring in low quality users at the top of the funnel to hit their own goal. Retention further down pays for it. Give one person the whole loop.

FAQ

What is the difference between a growth loop and a funnel?

A funnel shows how users move toward one goal and where they drop out. A growth loop asks a further question: do users who reach the goal bring in the next cohort? You need both views. The funnel tells you which step leaks. The loop tells you whether fixing that leak changes how fast the product grows.

Is a growth loop the same as a flywheel?

Mostly, yes. Both describe growth where results feed back into the system. The difference is precision. A flywheel is usually a picture of momentum. A growth loop names each step, the action that creates output, and two numbers you can track every month. If your flywheel diagram has no measurable step, rewrite it as a loop.

Can a B2B SaaS product have a growth loop?

Yes, and many already run one. Collaborators invited into a workspace, shared reports that outsiders open, customer case content that prospects find, and partner referrals can all bring in new accounts. These loops often turn slowly. Track them by account rather than by user, and attribute each new account to the source that created it.

How long before a growth loop shows up in the numbers?

Cycle time decides that. A loop that turns in days shows up within a few weeks. A loop that turns once a quarter needs at least one full turn before you can judge it. Pick one cohort and follow it from sign-up to the sign-ups it caused. Decide whether the loop works only after that turn completes.

Where to go from here

You now have a loop map, a step to fix and two numbers to track: new input per turn and cycle time. Often the weak step sits between product and marketing, like a sign-up wall that stops users before they ever invite. Kreto works on exactly that kind of step as an outsourced growth team. If that is where your loop breaks, tell us what you found.