Growth Loop

What is Growth Loop?

A growth loop is a closed system where a product's output becomes new input, so each cycle feeds the next and growth compounds instead of resetting.

The term comes out of a four-post Reforge series co-written by Brian Balfour, Casey Winters, Kevin Kwok, and Andrew Chen. Their core claim: a funnel moves in one direction, you put users in the top and get revenue out the bottom, with no built-in way to reinvest that output back into new input. A loop closes that gap. As the post puts it, "loops are closed systems where the inputs through some process generates more of an output that can be reinvested in the input." A referral loop takes a signed-up user, has them invite a friend, and that friend becomes a new input to the same loop. A content loop takes a user's post, lets it rank in search, and turns each new visitor into another potential poster.

Elena Verna later folded the same idea into her "Five Laws of Growth," telling B2B and PLG teams to "build loops, not funnels" because loops are self-sustaining while funnels need constant top-of-funnel fuel. Reforge and Verna agree on the mechanism. Where they differ is emphasis: Reforge frames loops as a modeling tool for diagnosing where a product's growth actually comes from, Verna frames them as an operating priority for a growth team's roadmap. Produck treats both as true and uses the modeling lens more, because the loop only matters once you can point to the exact step where it stalls.

Why it matters for product-market fit

A growth loop is a hypothesis about what makes someone come back and bring someone else with them. You can't diagnose that hypothesis by staring at a dashboard. You diagnose it by listening to the users going through the loop right now and asking why it worked or why it stalled.

That's the connection to Produck's Listen, Diagnose, Decide, Ship loop. Listen surfaces the actual moment a loop breaks, a user who invited nobody or a piece of content nobody found. Diagnose figures out whether that's an Activation problem (they never reached the value that would make them want to invite someone) or a mechanism problem (the invite flow itself is broken). Decide picks the fix that reopens the loop fastest. Ship gets it in front of real users so you can listen again and see if the loop actually compounds this time.

This is also why a growth loop only pays off downstream of Retention. A loop that recruits users who churn in a week isn't compounding, it's just spending effort to refill a leaky funnel with extra steps. And a loop is not a synonym for growth. It's a mechanism you point at a North Star Metric so you can tell whether the loop is actually moving the number that matters or just generating activity that looks busy. Most teams that think they've lost product-market fit have actually lost track of which loop, if any, is compounding. That's the deeper argument in our piece on why the PMF stack is broken: you can't decide what to build next if you can't see which loop is producing your growth and which is just noise.

When it works, and when it doesn't

It works when

• The output of one cycle is a genuine input to the next, a new user or a piece of indexed content, not just a vanity action that looks like reinvestment.
• Retention is already solid enough that the people the loop brings in stick around long enough to trigger the loop again.

It falls short when

• The team builds the loop mechanism before anyone has confirmed users get real value from the product, so the loop just recruits people who bounce.
• Leadership tracks the loop as a slogan instead of a model, celebrating "we have a referral loop" without measuring the actual reinvestment rate.
• The loop depends on a channel the company doesn't control, a platform API or a partner's terms, and that channel changes without warning.
• Nobody owns diagnosing why the loop's cycle time is slowing down, so it gets slower for months before anyone notices in the metrics.

How to apply it

1. Draw your current loop on paper: name the exact input and the action that produces the output, and confirm the output really does become a new input rather than just a downstream metric.
2. Pull five to ten real user sessions or feedback threads from people who went through the loop recently and check where they dropped off.
3. Diagnose whether the drop-off is an activation problem or a mechanism problem, with retention feeding back into the loop either way.
4. Pick one lever to fix, the invite prompt or the content's discoverability, rather than redesigning the whole loop at once.
5. Ship the fix to a slice of users and measure the loop's actual cycle time and reinvestment rate, not just top-line signups.
6. Feed what you learn back into Listen so the next diagnosis starts from real behavior instead of a guess about why the loop stalled.