Types of growth loops

Once you see funnels vs loops, the next question is which loop your product can actually run. The output a happy user naturally produces — invites, content, revenue, references — dictates what is available.

Network effects mean the product gets more valuable as more people use it. They are not identical to growth loops, but they supercharge loops: each new user improves retention and invite conversion for the next user. Without a loop, network effects are a latent asset. With a loop that does not deepen the network (pure paid spam), you rent growth without building a moat. See competition and moats.

Walking through the shapes

Content / UGC + SEO. Users create content → content ranks or gets shared → new users arrive → they create more content. Pinterest, Stack Overflow, Notion templates, YouTube, and many PLG docs sites run variants. The product investment is creation tools, public surfaces and indexable pages — not just ads.

Paid. Revenue (or gross margin) funds acquisition → new paying users → more revenue. Works when LTV > CAC with enough payback speed. The loop breaks when paid channels inflate CAC faster than LTV rises. Engineers meet this in attribution pipelines, paywall experiments and margin-aware feature costs (especially AI usage).

Sales. Customers succeed → case studies / references / expansion → new pipeline → more customers. Common in enterprise. The product work is admin-grade reliability, proof artifacts and usage that creates champions. Slower cycle time than viral loops; high ACV can still win.

Collaboration / cross-side. Inviting a teammate is both activation and acquisition (Slack, Figma, Calendly). Marketplaces need liquidity loops: more supply improves demand conversion, which attracts supply. Those loops are fragile early and strong later — classic cold-start.

Loops interact with moats. A content loop that builds an indexed corpus is harder to copy than a paid loop anyone can rent with ad budget. A collaboration loop that fills a workspace with teammate history raises switching costs. Prefer the loop that both grows and accumulates an asset (data, network, content, integrations).

Instrument each edge with a named rate. Content: create → publish/index → impressions → signup → create. Paid: pay → margin → ad spend → signup → pay. Same discipline as K-factor, different events.

Common mistakes

• One loop to rule them all. Mature products run a system of loops. Optimize the constraint loop, not the vanity one.
• Channel conflict. Paying for users who would have come through organic, then declaring paid "more efficient" because attribution is last-click.
• Treating network effects as a growth plan. They amplify a loop. They do not replace one.

Check yourself

A single-player notes app wants K > 1 via "invite 10 friends to unlock." What loop should it try instead?

Content, paid or a real referral — not a fake gate. Collaboration virality does not fit a single-player job.

Why is a content loop often a better engineering bet than a paid loop?

It accumulates an indexed corpus — an asset a competitor cannot rent next quarter.

Key takeaways

• Content, paid, sales and collaboration loops fit different product shapes.
• Network effects amplify loops; they are not a substitute for one.
• Prefer loops that accumulate an asset, not only temporary traffic.

Further learning

Next, design the explicit viral case: referral programs.