Growth loops
Optimize for comments, then wire the output of one turn into the input of the next.
- comments > shares
- giveaway machine
- praise-led
- creator armies
A loop compounds because its output feeds its input. A stunt spikes and fades. A loop turns once and sets up the next turn.
The one law for social loops
Comments > shares. The feed rewards conversation, not distribution. Optimize every mechanic for comments. Then reply to all of them. Each reply is a notification that pulls that person back, and replying roughly doubles engagement on your own post.
LoopThe giveaway → pipeline machine
Make entering stupid-easy but flattering, and give away things people show off.
see detail →hide detail ↑
Four comment mechanics that work:
- WRONG answers only. People fetch a funny line and paste it.
- Tag a friend, you both win. It's a gift to give, not a bother.
- Guessing games. Zero risk of looking dumb.
- Team challenges. "Get 5 people from your company to comment, all 5 win." That pushes the post into their work Slack for you.
Give away things people show off - joke apparel, desk toys, experiences for couples. Free trials and gift cards don't travel. Gamify the brand, not the product.
The loop
↺ ship the prize with a handwritten note + "post a photo, tagged, get a $150 card" → winners post → new audience → back to the top
LoopPraise-led marketing
The cheapest way into someone else's audience isn't paying them - it's praising them.
see detail →hide detail ↑
Four steps:
- Build a real "Top [Role]s" platform. Public criteria, open nomination box.
- Research each winner properly before you contact them.
- Reach out decisively. "We're writing an article about you and your role." Not "considering."
- Let them shape their own section. They'll share it when it publishes.
Keep it honest and specific. Include smaller voices, because they're hungrier to share. Aim at audiences that already post - marketers and creators, not CFOs.
LoopCreator armies & community as distribution
- Pay per view, not per follower. Open the ambassador program to anyone. Post about us, get paid per impression, no minimums. Programs like this pull eight-figure monthly impressions.
- Cultivate future buyers. Student clubs and certifications teach people your product before they're hired. They arrive on day one already using it.
- Scarcity as belonging. Make the certification hard to earn. People put it on their profile, and that turns users into public advocates.
Advanced: the cross-platform flywheel
Never say "giveaway" on the social post itself. Announce the contest in your newsletter: "commenting on my post enters you."
The post stays organic. Newsletter readers flood the comments. The algorithm boosts it. New people find the newsletter, subscribe, and come back to comment. Two channels feeding each other.
Three loops with almost no marketing spend
- A footprint on every free app. Each app shipped on the free tier carries a small "Built with" footer. Every creator's audience becomes a discovery surface. Acquisition scales with usage, not budget.
- A public template directory. Clonable templates and prompt blueprints catch long-tail search and drop new users into something that already works.
- Small campaigns beat big ones. The $1M hackathon wasn't the winner. $5K "hackathons for good" in mission-driven niches returned more.
Steal: put your name on whatever your users publish. Turn your best outputs into a clonable library. Spend small in narrow communities instead of big in broad ones.
When the transaction itself becomes the recruiter
Groupon didn't bolt virality onto a coupon site. It built distribution into the purchase.
- Thresholds. A deal only unlocked once enough people bought in. Early buyers had a reason to recruit friends.
- A live counter. It showed how many more were needed. That created urgency and proof at the same time.
- Shareable prizes. The best deals were experiences people wanted to do with someone else.
Steal: make the purchase mechanic do the recruiting. Thresholds and counters beat "refer a friend" overlays, because the user actually wants the deal to spread.
Do this first
- 01Pick your best-performing piece of the last quarter. List everything it produced: comments, emails, users, data. That list is your loop input.
- 02Wire one output back into an input. Example: turn giveaway comments into a DM sequence, or turn tool usage into next month's content.
- 03Change one CTA from "share this" to a comment mechanic - a question with a wrong answer, a pick-one, a tag. Measure comments per post for two weeks.
Product-led growth
Ship a playable link instead of a demo form - the product is the marketing.
- playable link
- delight
- dogfood in public
The best distribution is a product people can use without talking to you. If your product is hard to explain, stop explaining. Let them touch it.
The default
"Try it yourself" > "Book a demo." If people can play with your product in a browser, skip the demo form and ship the link. A playable AI demo hit a million users in a matter of days, mostly from people posting clips of it.
MoveShip a public, playable link on day one
Replace the press release with a URL.
see detail →hide detail ↑
Let people generate the "aha" themselves. A face-swap that runs in the browser. A live game. A one-of-one demo for each journalist with their own face in it. When the product is the content, distribution takes care of itself.
MoveBuild delight where nobody expects it
Small, joyful, in-product moments that exist purely to be shared.
see detail →hide detail ↑
A hidden easter egg, a celebration animation when a user completes a core action, random funny release notes people anticipate. These cost little and manufacture word of mouth from inside the product.
MoveDogfood in public
Using the product in public is the marketing.
see detail →hide detail ↑
Assign someone to use your own product in public and post the output. A filmmaker who does nothing but make videos with your video model. A team that publishes the exact playbook and the receipts from using your own ad product to cut cost-per-lead 10x.
Old-school PLG loops still work
The Dropbox "space race" is evergreen. Students nominated their university to win extra access, then spammed classmates to vote. Any loop where a user's own self-interest recruits the next user belongs in the toolkit.
Activation is a time-to-value problem, not a messaging problem
The biggest conversion lever wasn't copy. It was cutting signup-to-working-app down to 60-120 seconds.
They deleted the middle steps: approval gates, pre-deploy code checks, config reviews. Activation tripled, even though first drafts shipped rough. A zero-credit "discussion mode" let people plan before spending anything, which killed the anxiety that stalls a first session.
Steal: time the path from arrival to first real outcome. Then delete gates. A slightly worse first result, much faster, wins. Make exploring free.
Do this first
- 01Time yourself going from your homepage to a real outcome in your product. Write down the number of minutes and the number of gates.
- 02Delete one gate this week - the signup wall, the demo form, the credit card, the onboarding survey. Pick the one blocking the aha.
- 03Replace your "Book a demo" button with something playable: a live sandbox, a prefilled example, a shareable link.