Retention Loops, Viral Loops, and Reactivation: The B2B Growth Marketer's Playbook

Most B2B retention work is a stack of feature launches, lifecycle emails, and in-app nudges that never bend the curve. The team is busy. The slack channel is on fire. The Notion doc is 40 pages long. And week-12 retention is still 18%.

The reason is uncomfortable: there's no loop. The team is shipping inputs to a flywheel that doesn't exist. You can run the prettiest "we miss you" email in the industry, and it won't matter, because the product isn't producing the next reason to come back.

This is a Growth Marketer's playbook for finding the actual loop in event data, doing the habit-moment math, and running reactivation that fits how B2B buyers actually behave. No fluff, no vanity metrics, no "delight" framing.

How to Read a Retention Curve

Before you ship a single retention experiment, look at a cohort table. If your product manager hasn't given you one, build it yourself in Amplitude or Mixpanel and stare at it for an hour.

Here's what a healthy B2B SaaS cohort table looks like:

Notice the curve flattens between week 8 and week 12, then settles around 38-40% and stays there. That flat tail means you have a product. Real users are getting real value on a repeating cadence. The width of that tail is your business.

Compare to a leaky bucket:

The curve never flattens. Every week, more cohorts leak. By week 20 you've lost 94% of every cohort. That's not a retention problem you fix with a lifecycle email. It's a value-delivery problem disguised as one.

The flat tail is the signal. If your tail isn't flat, no campaign will save you. Stop the email work and go find out why people stop coming back.

One more thing on cohort tables: month-1 retention is a vanity number for B2B. People stick around in week 1 out of obligation (the trial they signed up for, the demo they booked, the boss who told them to evaluate it). Month 3 is where the truth is. If you only ever look at activation and W1, you're missing the actual signal.

The 3 Loop Types, with B2B Examples

A loop is a closed sequence of user actions where each step's output becomes the next step's input. It's not a funnel, which is linear and ends. A loop produces more of itself.

There are three types that matter for B2B SaaS.

Engagement Loops: Usage Creates More Reasons to Use

The strongest engagement loops are ones where the act of using the product generates new triggers to use it again. Linear is the canonical B2B example.

Linear's loop, in event data:

1. User ships an issue (writes a ticket, assigns it, ships a PR).
2. Linear notifies the assigned teammate via Slack and email.
3. Teammate opens Linear, leaves a comment or status change.
4. Original user gets notified. They open Linear to reply.
5. The conversation produces a new issue, which restarts the loop.

The loop isn't in Linear's marketing copy. It's in their event log. Every issue created is a future trigger. Every comment is a re-engagement event. The product is structurally designed so that working in it produces more work in it.

If your product doesn't have this, no email cadence will fake it.

Viral Loops: Users Bring Users

Viral loops in B2B are rare and most often misdiagnosed. The B2B example that actually works is Loom.

Loom's loop:

1. User records a video.
2. User sends the link to a teammate, customer, or external collaborator.
3. Recipient clicks the link, watches the video.
4. To react, comment, or reply with their own video, the recipient signs up.
5. New user records their first video and sends it to someone.

The k-factor math here is straightforward. K-factor = (invites sent per user) × (conversion rate of those invites). If a Loom user sends 3 video links per week and 8% of recipients sign up, k = 3 × 0.08 = 0.24. That means every 100 active users produce 24 new signups per week from sharing alone. Over 12 weeks of compounding (with retention factored in), that's substantial organic growth without paid acquisition.

For comparison, most B2B tools that claim "viral" have a k-factor under 0.05. They're not viral; they're slightly above zero.

Monetization Loops: Usage Creates Revenue That Funds More Reach

Notion is the textbook B2B monetization loop.

1. One IC adopts Notion for personal docs.
2. They invite their immediate team to collaborate on a project.
3. Team usage triggers admin review (billing, security).
4. Admin upgrades to a paid team plan and adds more seats.
5. New seats invite more teammates from other departments.
6. Cross-department adoption triggers an enterprise conversation.

Each completed loop pulls more revenue and more seats into the product. Notion didn't grow on ad spend; they grew because the product structure converts every individual user into an entry point for a team plan, and every team plan into an entry point for the org.

If your monetization model is "free trial, then a sales call," you don't have a monetization loop. You have a funnel with a paywall.

How to Find the Loop in YOUR Product

The loop in your marketing copy is rarely the loop in your event data. Stop reading your homepage. Open Amplitude.

Three questions to ask the data:

1. What action do retained users repeat?

Run a cohort of users who hit week 12. List the events they fired in weeks 1-12, ordered by frequency. The top 1-3 events are your candidate loop actions. Not the events you wish they'd fire. The ones they actually do.

In Amplitude, this is a User Lookup combined with an Event Frequency report filtered to your retained cohort. In Mixpanel, it's a Cohort report joined with an Insights breakdown by event.

2. What triggers that action?

For each candidate loop action, look at the event 1-2 steps upstream. Is it a notification? A teammate's action? A scheduled task? An external integration?

If the trigger is "they remembered to log in," you don't have a loop. You have a journal that requires self-discipline. That's a fragile retention story.

3. What does completing the action produce that triggers the next one?

The output of step 2 has to feed back into step 1 (or into another user's step 1). If it doesn't, the loop is broken.

Common failure mode: the action produces value for the user but doesn't generate a new trigger. A user runs a report, gets the answer, closes the tab. That's a one-shot. If your product is full of one-shots, retention will leak no matter how good each individual interaction feels.

Name the failure mode. I call it the false habit moment: the action feels valuable but doesn't generate the next trigger. Once you can see false habit moments in your event data, you stop building campaigns around them.

The Habit Moment Math

A habit forms at the intersection of frequency, value, and triggers.

Frequency: how often does the user need to do the underlying job? Daily, weekly, monthly, quarterly? B2B SaaS sweet spot is weekly. Daily-habit B2B products are rare; they're mostly chat (Slack), dev tools (GitHub), or sales tools with active pipelines (CRMs). If you're trying to manufacture a daily habit on a weekly job, you're going to fail and make people resent your notifications.

Value: does completing the action produce a clear, immediate outcome the user can feel? Closed deal, shipped issue, sent invoice, scheduled meeting. If the value is delayed (analytics that need 30 days of data), you have a delayed-gratification problem and the habit won't form on its own.

Triggers: what reminds them to come back? External (notification, email, calendar), internal (problem they want to solve), or contextual (they were already in adjacent software). Strong loops use all three.

Worked example. A project management tool with these signals:

• Frequency of underlying job: weekly sprint planning, daily status, weekly retros.
• Value: clear, immediate (issue shipped, blocker raised, status updated).
• Triggers: external (Slack notifications on assignment), internal (sprint cadence), contextual (referenced in standup).

That's a healthy habit moment. The math compounds: a weekly cadence × clear value × three trigger types is what produces a flat-tail cohort.

Now contrast: a "team OKR tracker" used quarterly.

• Frequency: 1-2 times per quarter.
• Value: aspirational, not immediate.
• Triggers: only the calendar reminder.

That's not a habit moment. It's a memory test. Cohort retention will leak straight through, and no email will fix it. The fix is product (more frequent check-ins, clearer immediate value), not lifecycle.

Reactivation Campaigns That Work for B2B

Most reactivation programs are "we miss you" emails sent to dormant users on a 30-day timer. They convert at under 1% and feel like begging.

What works is event-based reactivation: the email goes out when something material has changed since the user left. The trigger is the event, not the calendar.

Four reactivation triggers that consistently outperform the calendar version, with rough conversion ranges from B2B SaaS teams I've seen:

1. The integration they asked for shipped. Pull the list of dormant users who hit your in-app "request integration" feature or asked support. When the integration ships, send each one a personal-feeling note: "Last August you asked about the Salesforce sync. It's live. Here's the 90-second setup." Reactivation rate: 8-15%, vs. 0.5-1% for time-based.

2. A teammate from their company signed up. Run a query against your Auth events. When a new user from a previously-dormant user's domain activates, email the dormant user: "[Teammate name] from [Company] just started using [Product] for [use case]. They've connected to your old workspace." This works because it's true and it's relevant. Reactivation rate: 6-12%.

3. A new feature solves the use case they bounced on. For dormant users who churned out of a specific workflow (you can identify them by the last event they fired before going inactive), wait for the next product release that addresses that workflow. Send: "You tried [Workflow] in October and it didn't have [Capability]. We just shipped it. Here's how it works in 60 seconds." Reactivation rate: 5-10%.

4. Their pricing tier changed in their favor. If you reduce price, increase a free-tier limit, or unbundle a feature, dormant users who churned for cost reasons are the highest-converting segment in your database. Reactivation rate: 12-25% if the original churn reason was price.

Win-back windows by ARR tier:

• Self-serve under $1k ARR: 30-90 days. After 90, they've found a substitute.
• Mid-market $1k-$25k ARR: 60-180 days. They have buying cycles, contracts, and replacement projects.
• Enterprise over $25k ARR: 90-365 days. Renewal calendars are long. The right time to come back is at their next budget cycle.

The "miss you" email is event-blind, and it tells a dormant user nothing has changed. Of course they don't come back.

Why Most Viral Loops Fail in B2B

The single biggest reason a B2B viral loop fails is the single-player to multi-player flip never happens. An IC loves the tool, uses it daily, and never invites anyone. Year-one retention can look great. Year-two ARPU is flat. The tool is loved by individuals and never spreads.

Three reasons the flip stalls:

1. The artifact isn't shareable. If using the product produces a thing that lives only inside the product (a private dashboard, a personal note, an internal-only report), there's nothing to send to a colleague. Loom works because the artifact is a link. Notion works because the artifact is a public-by-default page. If your artifact is a screenshot the user has to manually export and email, you've broken the loop.

2. The collaboration moment requires admin friction. If the second seat needs IT approval, a paid upgrade, or a credit card, the IC won't push. They'll keep using it solo. Free seat #2 changes this dramatically. Slack's freemium model worked for this reason: the IC who loved it could pull a teammate in without asking permission first.

3. The handoff isn't designed. B2B work is full of natural handoffs: marketing to sales, sales to CS, design to dev, manager to IC. The product should make completing one side of the handoff trigger the other. If a marketer can hand a lead to sales inside the product, sales has a reason to be in the product. If they can't, you have two single-player products.

Name the failure mode. I call it the single-player trap: the product produces value for one user but doesn't structurally produce the second seat. You won't fix it with a "share this with a teammate" email. You fix it in product by changing what the artifact is and where the handoff lives.

Measurement: Cohort Tables Beat Vanity Dashboards

The dashboard your CEO looks at probably shows MAU, growth, NPS, and a churn percentage. None of those will tell you if your loop is working.

What to actually look at, weekly, in Amplitude or Mixpanel:

1. Weekly cohort retention by acquisition source. Are paid-traffic cohorts retaining at 38% W12 and organic at 52%? Then your paid sources are bringing in users who don't fit the loop. Stop scaling them, or fix the targeting.

2. Retention by activation event. Group users by whether they hit your aha moment in week 1. Activated cohorts should retain 2-3x better than non-activated. If they don't, your "activation event" isn't the actual aha moment. Find a better one. Run the same cohort table grouped by every candidate activation event and pick the one that produces the biggest retention spread.

3. Retention by team size. B2B retention almost always rises with team size. If your single-user cohorts retain at 22% W12 and your 5+-user cohorts retain at 61%, you've just found the case for shipping team-collaboration features. The numbers tell you the loop is multi-player and you need to help the flip happen sooner.

4. The north star metric that maps to the loop, not the OKR. If your loop is "ship issue → notify team → reply → ship more," your north star is weekly active issues per team, not MAU. If your loop is "record → share → recipient signs up," it's videos shared per active user × invite conversion rate. The metric should be the rate at which the loop completes. If the metric goes up, the business goes up. If the metric is "monthly logins," it's measuring presence, not the loop.

In Amplitude, the relevant report types are Retention Analysis (weekly granularity, custom activation event), Cohort comparison, and Pathfinder (to see what events come before and after your candidate loop action). In Mixpanel, the equivalents are Retention reports, Cohort breakdown, and Flows.

Don't trust a screenshot from a single week. Loop diagnostics need 8-12 weeks of cohort data minimum. If your team's been live for less than that, focus on activation, not retention.

Closing

The Growth Marketer's job in retention is to find and amplify the loop, not to paper over its absence with campaigns. If you're staffed and KPI'd around emails, in-app messages, and reactivation flows, you're working downstream of a problem that lives in product. The numbers will tell on you.

The good news: when the loop is real, even small amplification work compounds. A 4-point lift in W12 retention on a flat-tail cohort is worth more, in revenue, than every "we miss you" email you'll ever ship.

Find the loop. Do the math. Trust the cohort table. Skip the rest.

Learn More

• Activation Metrics That Actually Predict Revenue
• PLG vs Sales-Led: Picking the Right Growth Motion
• Lifecycle Email That Doesn't Suck
• North Star Metrics for B2B Growth Teams