The traditional marketing funnel is a leaky bucket. Users enter at the top, some convert, most drop off, and you start over. Loops are different. A loop takes the output of one cycle and feeds it back as the input of the next, creating compounding growth or compounding retention.
The question every product-led team eventually asks is: should we build retention loops or growth loops first? The answer depends on where your product is, and getting it wrong wastes months of engineering and product effort.
Defining the two loop types
Retention loops
A retention loop brings existing users back to your product repeatedly. The output of each session creates a reason for the next session.
Structure: User takes action > Product generates value > Value creates reason to return > User returns and takes action again.
Examples:
- Notion: User creates a document > Shares it with teammates > Teammates comment > Original user gets notified > Returns to respond
- Linear: Developer closes an issue > New issues are assigned > Notification sent > Developer returns to work on next issue
- Figma: Designer creates a component > Other team members use the component > Component needs updating > Designer returns to iterate
The key characteristic: the product itself generates the trigger that brings the user back. No email required. No push notification needed. The workflow creates the return visit.
Growth loops
A growth loop uses existing users to acquire new users. The output of existing user activity becomes the input for new user acquisition.
Structure: User takes action > Action exposes product to non-users > Non-users sign up > New users take action > Cycle repeats.
Examples:
- Slack: Team adopts Slack > Invites external collaborators via Slack Connect > External collaborators experience Slack > Bring it back to their own company
- Calendly: User sends a scheduling link > Recipient sees Calendly branding > Recipient signs up for their own scheduling > Sends links to their contacts
- Loom: User records a video > Shares link with colleague > Colleague watches and sees Loom > Signs up to record their own video
The key characteristic: user activity naturally exposes the product to non-users without requiring the user to explicitly "refer" anyone.
The "loops not funnels" mental model
Funnels are linear: awareness > consideration > decision > retention. Each stage has leakage, and you need to keep filling the top.
Loops are circular: each completed cycle generates fuel for the next cycle. This creates compounding effects:
- Retention loops compound engagement: More content/data in the product makes it harder to leave and more valuable to return to.
- Growth loops compound acquisition: More users generate more exposure, which generates more users.
The critical insight is that loops only work when the underlying action is strong enough. A growth loop built on a product nobody wants to use will not compound. A retention loop built on artificial triggers (badges, streaks, gamification) without real value will decay.
When to build retention loops first
Build retention loops first when:
1. You are pre-product-market fit
If users are not sticking around, acquiring more users just fills a leaky bucket faster. A growth loop on top of poor retention is a waste. Fix retention first.
Signal: Your retention cohort curves do not flatten. Week 8 retention is below 15% for a weekly-use product, or month 3 retention is below 20% for a monthly-use product.
2. Churn is your primary constraint
If your monthly churn rate is above 5% for SMB or above 2% for enterprise, every new user you acquire is offset by users leaving. The math is simple: reducing churn from 5% to 3% has a bigger impact on steady-state user count than doubling acquisition.
3. You have a small user base
Growth loops need a minimum number of users to generate meaningful output. If you have 100 users and your growth loop converts at 5%, you get 5 new users per cycle. That is not a loop; that is a trickle. With a small base, focus on making every existing user deeply engaged.
4. Your product has high switching costs once adopted
Products like databases, project management tools, and CRMs become more valuable with more data and more users. The retention loop here is the accumulation of data and workflows. Once a team has 6 months of project history in your tool, the cost of switching is enormous. Invest in making the product stickier.
When to build growth loops first
Build growth loops first when:
1. Retention is already strong
If your month-3 retention is above 40% and cohort curves are flat, your product is sticky. Users who activate stay. Now the constraint is getting more users in the door.
2. Your product has natural viral mechanics
Some products create viral exposure by design. Collaborative tools, communication platforms, and content-sharing products naturally expose non-users to the product. If your product has this property, a growth loop is low-effort and high-impact.
Examples of natural viral mechanics:
- Users share output with non-users (documents, links, videos, reports)
- Users invite collaborators to complete a workflow
- Users embed product output on external sites (badges, widgets, charts)
3. You have a content flywheel opportunity
If your product generates content that can be indexed by search engines (user-generated pages, public profiles, templates), you have the foundation for a content-driven growth loop: users create content > content gets indexed > new users find content via search > new users sign up > create more content.
4. Your unit economics support it
Growth loops often require the product to be free for the loop participants. If your cost to serve a free user is near zero and your monetization comes from a subset of users upgrading, a growth loop is economically viable.
Decision framework
Use this flowchart to decide where to start:
Step 1: Check retention
- Is your month-3 retention above 30%? If no, build retention loops first. Full stop.
- If yes, proceed to step 2.
Step 2: Check viral potential
- Does your product naturally expose non-users to the product through normal usage? If yes, build a growth loop. The mechanics are already there; you just need to optimize them.
- If no, proceed to step 3.
Step 3: Check user base size
- Do you have more than 1,000 active users? If no, build retention loops. You do not have enough users to fuel a growth loop.
- If yes, proceed to step 4.
Step 4: Evaluate effort
- Can you build a growth loop with less than 2 weeks of engineering effort? If yes, do it. The learning will be valuable even if the first version does not scale.
- If no, build retention loops. They are more predictable and you can layer growth loops on later.
Design templates
Retention loop: the workflow loop
Best for: Project management, CRM, development tools.
- User completes a task or workflow step
- Product generates the next task or step automatically
- User receives notification or sees the next step in their queue
- User returns to complete the next step
Implementation: Create a workflow engine that chains actions. When one action completes, the next is queued. Notifications are tied to the queue, not to arbitrary schedules.
Retention loop: the data accumulation loop
Best for: Analytics, databases, knowledge management.
- User adds data to the product
- Product generates insights or connections from the accumulated data
- Insights are surfaced to the user (dashboard, digest, alert)
- User returns to consume insights and adds more data
Implementation: Build automated insight generation that improves with more data. Show users what they are missing by not adding more data.
Growth loop: the collaboration invite loop
Best for: Collaboration tools, communication platforms.
- User creates shared content or workspace
- User invites collaborators (teammates, clients, partners)
- Collaborators sign up to participate
- Collaborators create their own content and invite others
Implementation: Make inviting collaborators the path of least resistance. The invite should be part of the core workflow, not a separate "invite friends" feature.
Growth loop: the content SEO loop
Best for: Community platforms, template marketplaces, knowledge bases.
- User creates public content (template, guide, profile)
- Content gets indexed by search engines
- New visitors discover content via search
- Visitors sign up to create or customize content
- Their content gets indexed, expanding the surface area
Implementation: Make user-generated content indexable by default. Add structured data. Create category and tag pages that aggregate content.
How to measure loop health
Retention loop metrics
- Return rate: Percentage of users who return within the expected usage interval (daily, weekly, monthly)
- Loop completion rate: Percentage of users who complete the full loop cycle (action > trigger > return > action)
- Time between loops: Average time between loop cycles. Shorter is better, but it should be consistent.
- Cohort retention curves: Plot retention by weekly or monthly cohorts. Healthy retention loops produce flattening curves.
Growth loop metrics
- Loop cycle time: How long does one complete loop take from user action to new user signup?
- Branching factor: How many new users does each existing user bring in per cycle? A branching factor above 1.0 means exponential growth. Below 1.0 means the loop is supplementary, not primary.
- Loop-sourced signups: Percentage of total signups that come from the growth loop vs. paid acquisition or organic search.
- Quality of loop-sourced users: Do users acquired through the growth loop activate and retain at the same rate as users from other channels?
The real answer: you need both
Retention loops and growth loops are not an either/or decision in the long run. They are a sequencing question. The most successful PLG companies build retention loops first to ensure users stick, then layer growth loops on top to compound acquisition.
Figma is the canonical example. They first built retention loops through real-time collaboration and component libraries that made the product indispensable for design teams. Then they built growth loops through shareable prototypes, community templates, and multiplayer editing that exposed the product to new users.
Slack started with retention loops (channels, integrations, searchable history that made it the team's communication backbone) before growth loops (Slack Connect, shared channels) accelerated cross-organization adoption.
The pattern is consistent: retention first, growth second. A product that retains well and grows slowly will eventually win against a product that grows fast but cannot retain.
Start by measuring your retention cohort curves. If they are not flattening, that is your first priority. Once they flatten, look for natural viral mechanics in your product and build a growth loop around them.