Most growth programs are built as funnels: pour traffic in the top, hope customers come out the bottom, repeat. Funnels are linear, predictable, and eventually expensive. Loops are different — each new customer feeds the next batch of acquisition, lowering blended cost over time. The companies that compound for a decade have at least one well-designed loop running under the hood. Most can be built without a single new engineer.
Section 01
Understand the difference between a funnel and a loop
A funnel takes an input (ad spend, email blast, sales effort) and produces an output (a customer). A loop takes an output (a new customer) and turns part of that output back into an input (referrals, content, network effects, integrations). Funnels scale linearly with spend; loops scale geometrically with adoption. Knowing which one you're building is the first move.
Section 02
Identify the natural exhaust of your product
Every product produces something on the way to delivering its value — a shared link, a tagged email, an analytics widget, an embedded badge. That exhaust is the raw material for a loop. Slack's 'invite your team' is exhaust. Calendly's branded scheduling link is exhaust. Spend a week mapping every artifact your product produces that touches a non-customer.
Section 03
Design for the second-order user, not just the first
The person who experiences your product through a customer's exhaust — a recipient of a Calendly link, a viewer of a shared dashboard — is the seed of your next acquisition. Optimize that surface ruthlessly: clear branding, a low-friction next step, and a reason to click. Most companies treat this surface as an afterthought, which is why their loops never form.
Section 04
Make the referral mechanic native to the workflow
Bolted-on 'refer a friend' programs rarely work because they ask the customer to leave the product to do extra work. Embedded referral mechanics — share a project, invite a collaborator, send a result — work because the share is already the next thing the customer wanted to do. Find that moment and make the share feel inevitable.
Section 05
Measure the loop, not just the campaign
Loops have specific metrics: branching factor (how many new visits each customer generates), cycle time (how long from new customer to next new customer), and conversion rate of loop traffic. Track those three numbers monthly. If branching factor is below one, the loop is broken and no amount of campaign spend will fix it.
Section 06
Layer content loops underneath product loops
Even product-led companies benefit from a parallel content loop: customers ask questions, the company answers them publicly, those answers attract new customers who ask new questions. Done over years, this becomes a moat that competitors can't shortcut. It also requires real editorial discipline — most companies start a blog and then quit by month four.
Section 07
Be patient — loops compound slowly, then suddenly
Funnels show results in weeks. Loops show results in quarters. Most companies kill their loops before they compound, mistaking the slow start for failure. Pick one loop, instrument it, give it two quarters of disciplined investment, and only then judge whether it's working.
The takeaway
Funnels scale with spend. Loops scale with adoption. Find the natural exhaust of your product, design for the second-order user, and let the loop compound over quarters — not weeks.