Why funnels can't compoundCompounding Growth Systems

The funnel is a useful diagnostic. It's a terrible strategy.

The growth funnel has dominated marketing thinking for two decades. It's intuitive. It draws cleanly on a whiteboard. It maps to a CRM.

It's also incomplete in a specific way: a funnel describes a single linear path from input to output. There's no concept of how the output reinvests as new input. There's no compounding. To get more at the bottom, you put more in at the top. The math is forever linear.

Brian Balfour has argued for years that the fastest-growing products in software, consumer, and marketplaces are better described as loops^[1] — systems of self-reinforcing cycles where the output of one cycle feeds the input of the next. New users produce signals (content, referrals, transactions, reviews, data) that pull in more users. The flywheel turns on its own momentum. The math compounds.

The strategic implication is uncomfortable for funnel-organized teams. The question "how does your product grow?" cannot be answered by pointing at a funnel diagram. The funnel describes a step within the system. The loop describes the system itself.

What a growth loop actually is

A growth loop has four components, every time, in this order:

1. Input. What enters the system. New users, new content, new transactions, new data, new investment.
2. Action. What the input does inside the product. A user signs up, posts content, makes a purchase, hits a milestone.
3. Output. What that action produces externally. An indexed piece of content. An invitation. A public review. A referrable artifact.
4. Reinvestment. How the output produces new input. The indexed content ranks in search. The invitation lands in an inbox. The review increases conversion for the next prospect.

If any of those four are missing, you have a funnel, not a loop. A funnel produces output that ends in revenue and stops. A loop produces output that ends in more input.

The three canonical loop categories

Balfour organizes loops into three core categories: viral, content, and paid.^[1] Most mature products run a portfolio of multiple loops at once.

Viral loops

Viral loops produce a referral or invitation as a direct output of the user's action. A user uploads a file to Dropbox and shares it with three collaborators. The collaborators sign up. A user joins Slack and invites their team. The team becomes new Slack users.

The two key levers are the viral coefficient (the number of new users each existing user generates) and the cycle time (how quickly the loop completes). A coefficient above 1.0 produces self-sustaining growth with no additional spend. Below 1.0 needs supplementary fuel from other loops.

Content loops

Content loops produce indexable or shareable content as a direct output. Yelp users write reviews. Those reviews rank in search. Search drives new users. New users write more reviews. Pinterest users pin images. Pins rank in image search. Search drives traffic. Traffic creates more pins.

The key levers are content velocity (how much new content the loop produces per unit time), content quality (whether the content earns SEO authority or social distribution), and distribution capture (whether the system retains the new visitor long enough to produce the next piece).

Paid loops

Paid loops use revenue from existing users to fund acquisition of new ones. The two key levers are LTV:CAC ratio (does the loop produce enough revenue to fund forward) and payback period (how quickly the loop replenishes capital).

Paid loops feel less like loops and more like funnels with reinvestment. That feeling is partly correct. But the loop framing still adds value — it forces explicit capital reinvestment planning, not just budget allocation against revenue targets.

Why funnel-organized teams underperform

Teams organized around the funnel make three predictable mistakes that loop-thinking corrects.

Siloed ownership. Acquisition owns the top. Product owns activation. Lifecycle owns retention. No one owns the loop. Each function meets its KPIs while the system's compounding rate stays flat.

Linear capital allocation. Funnel teams ask "how much should we spend on acquisition this quarter?" Loop teams ask "what's the compounding rate of our viral loop, and is paid acquisition the right way to fund growth?" The first question gets a budget. The second gets a strategy.

Missing second-order effects. A funnel team measures campaigns by direct conversions. A loop team measures the same campaign by its downstream contribution to the loops the product runs. A campaign that brings in 1,000 users with low conversion rates might still be the right campaign if those users produce high-quality content for the content loop.

How to identify your product's actual loops

For most products, the loops are already running. They're just not named or instrumented.

1. List the product's outputs. Not what your company produces. What users produce inside the product. Posts? Reviews? Workflows? Transactions? Invitations? Data points?
2. For each output, ask: does this directly drive new input? If yes, it's a loop. If no, it's not.
3. Measure the loops you find. Cycle time, conversion rate at each step, compounding rate.
4. Rank by leverage. Some loops have coefficients of 0.05. Some have 1.4. Concentrate optimization on the one or two highest-leverage loops, not on every loop equally.

Building new loops vs amplifying existing ones

The default impulse when learning about loops is to build new ones. This is almost always wrong.

Amplifying an existing weak loop produces faster results than building a new one from zero. Loops compound once cycle time is short and conversion rates are decent. A loop with a 0.6 coefficient and 14-day cycle time is producing real value — slowly. Pushing the coefficient to 0.8 (33% lift) and the cycle to 10 days (29% lift) roughly doubles the loop's annual output. That's an achievable optimization.

Building a new loop from zero takes 6–18 months for most products.

Amplification levers to look at first:

• Cycle time compression. What slows the loop down? Make it faster.
• Conversion rate at the weakest step. Most loops have one step dramatically less efficient than the others. Fix that one.
• Quality of input. Right users? Right content quality threshold?
• Compounding triggers. Are there moments in the loop where an additional output could be deliberately triggered? Post-purchase reviews. Post-onboarding referrals.

What loops mean for marketing organization

Adopting loops as the planning unit has organizational consequences. The team can no longer be structured around funnel stages.

The right structure is around loops, with each loop owned by a cross-functional team (PM + engineer + designer + data + marketer) controlling the full input-action-output-reinvestment chain.^[2] That's essentially the Reforge prescription for growth teams: pod structure, full-stack ownership, loop-level KPIs.

If your current team is organized as acquisition + activation + retention + revenue, you have a funnel-structured team. Loop-structured teams produce loop-shaped results.

How RGM applies loops in client work

Every RGM growth strategy engagement starts with a loop audit. We identify which loops are running, measure them honestly, rank by leverage, and concentrate the next 6–12 months of investment on amplifying the 1–2 highest-leverage loops.

We avoid recommending new loops in the first quarter. That's a multi-quarter project. The amplification work usually produces 60–80% of achievable lift in the first 90 days.^[5]

The output of a loop audit is a single-page diagram, a measurement plan that instruments each step, and a quarterly OKR set tied to loop coefficients and cycle times — not to funnel stages.