Growth Loops & Acquisition

Core thesis

Funnels stay linear. Loops compound. A growth loop is a system where each cohort of users generates the next cohort — outputs become inputs. AI amplifies loops through relevance, personalization, and timing, but loops must be deliberately designed. They are rarely discovered by accident.

Loop anatomy

• Viral loops: users share and bring in users. Variations: personal viral (invites, referrals), social viral (shared outputs, status), financial viral (monetary incentives).
• Content loops: content creation attracts users who create or consume more content. Variations: company-generated, user-generated, supply-generated.
• Paid loops: revenue funds acquisition that generates more revenue. Variations: ad loop, sales loop.

Loop metrics that matter

• Cycle time: how long from acquisition to generating the next user
• Conversion rate between each step in the loop
• Amplification factor: how many new users each existing user generates
• CAC, LTV, and payback period where relevant
• Trust and quality signals: loops degrade when volume outpaces relevance

AI viral loop rules

Personal viral loops: net positive but more competitive. Use AI for contextual invites, high-influence inviter detection, predictive invite timing, and invited-user onboarding. Watch for spam filters, invite fatigue, and AI-simulated network value reducing the need for real people. Social viral loops: net positive only with real differentiation. Use AI for wow moments, shareable proof assets, and advocate discovery. Watch for shorter surprise windows and fast feature commoditization. Financial viral loops: strong potential if incentives align with core value and unit economics hold. Use AI for personalized referral offers, incentive elasticity testing, fraud prevention, and dynamic rewards. Watch for AI compute costs plus referral incentives exceeding LTV.

Growth model sequence

Spark (temporary arbitrage, novelty, underpriced channel) → Loop (convert spark into compounding growth mechanics) → System (connect multiple loops into a reinforcing growth model). Most companies get stuck between spark and loop — they ride the arbitrage until it closes and never build the loop that outlasts it.

Three acquisition levers (in order)

• Optimize existing loops: more output per cycle, faster cycles. Lower risk but has a ceiling. Always do this first.
• Add new loops: harder, often requires product, channel, or model changes. Creates step-function growth. Do this second.
• Add linear activities: use as activation energy or fuel for existing loops. High-intent, low-volume activities that start the loop. Do this last unless you have no loop at all.

Bottleneck-shift rule for AI loops

AI unlocks one constraint and moves the bottleneck elsewhere. Content loops: creation was the bottleneck — now curation, discovery, trust, and distribution are. Ad loops: creative production was the bottleneck — now unique targeting data and experiment design are. Always ask: after AI removes the old constraint, what is the new system constraint? The teams that win are the ones that identify and solve the new bottleneck while competitors celebrate solving the old one.