This summary explores everything known about viral loops from the Web 2.0 era to today's mobile age. It takes a deep dive into how viral loops work, how they are measured, their impact on product growth, and how they have evolved over time. It offers particular insight into the viral strategies of the early web era, the fundamental shift in viral marketing after mobile, and the viral strategies newly emerging in the age of artificial intelligence (AI).

1. The Golden Age of Viral Loops: The Web 2.0 Era 🚀

The Web 2.0 era (2005–2010) was a special period when viral products were systematically designed to reach millions of people. The first versions of things we now take for granted—social networks, user-generated platforms, collaboration tools, messaging apps—were built during this time. It was an era when the industry systematically learned how to create, measure, and optimize viral loops.

Most of the people who made viral products succeed back then became billionaires, or ended up as FAANG executives or venture capitalists. Then the mobile era arrived, and Web 2.0's viral knowledge began to fade. The author believes this lost knowledge is still highly relevant to today's Product-Led Growth, sharing flows in generative AI apps, and referral programs, and aims to compile all of it in this piece.

The key topics covered include:

• Simple viral loops that succeed but die quickly
• How to calculate and optimize the Viral Factor
• How Retention drives viral growth
• Why new users invite more people than existing users do
• The difference between word-of-mouth and deliberately engineered virality
• How mobile killed the classic viral loop
• What kinds of viral loops still work in the modern era

2. Understanding Viral Growth as an Equation 📈

Many people think of viral growth as simply a great video getting shared a lot and driving traffic, but the viral loops the author describes are those deliberately designed into the product—invitation features, tags, referral links, and so on. These kinds of viral loops share the following characteristics:

• They can be measured and tracked.
• They can be improved through product changes.
• There is a mathematical principle that applies to any form of product-led virality.

2.1. Measuring the Viral Factor

The Viral Factor is a simple ratio. For example, if 100 users enter a product and invite 150 new users who then sign up, the viral factor is 1.5. If they get 50 to sign up, it's 0.5. If the viral factor is below 1, the viral loop will eventually stop working.

Viral Factor = (Number of new users who signed up via the viral loop from a given cohort during a given period) / (Number of users in that cohort during that period)

2.2. Analyzing a Content-Sharing Loop 🎬

Take a viral loop where users create and share something—like a generative AI or photo filter app. A user creates content and shares it; new users see the link, encounter the content, and some decide to sign up and create their own content.

To track this systematically, you need to encode a sharer ID—something like sharer_id—into the share link. For example: productdotcom/vid/[video ID]?sharer_id=[sharer]

When a new user signs up via this link, you store the sharer ID alongside the new signup's information. As this data accumulates, you can calculate the viral factor by measuring how many new users a given cohort (say, 100 people who signed up three months ago) has brought in. Users without a sharer_id are treated as "Gen 1 users" or "on-ramp users" and excluded from the viral factor calculation. This ratio tends to stabilize over time.

2.3. Optimizing the Viral Factor ⚙️

Once you can measure the viral factor, the next question becomes: "How do I get this number above 1?" You can increase the viral factor through product changes—prompting new users to invite others on their first visit, making share links easy to copy and paste, and so on.

This measurable ratio can be put on a dashboard for ongoing monitoring, and A/B tests can reveal which changes improve share rates and conversion rates. This idea applies powerfully to Product-Led Growth as well, since the concept can be applied to any process where one user generates another.

2.4. Alternative Viral Factor Calculations and the Danger of Spam Loops 🚫

The viral factor is commonly described as invites × conversion rate, but this doesn't cover every kind of viral loop. What matters more is knowing the ratio between two cohorts.

Focusing solely on increasing invite counts and conversion rates can lead to spam loops. In the early Web 2.0 days, social networks grew by blasting friends with emails—linking address books and sending invite emails to hundreds of people at once. Conversion rates were high at first, but email providers eventually classified this as spam and effectiveness declined. That said, this approach is what allowed large products like Facebook and LinkedIn to get off the ground.

2.5. Why Chain Letters and Spam Loops Eventually Fail 📉

Like a chain letter, a viral loop can ultimately fail due to market saturation. Even if you invite 200+ people every cycle, after a few iterations you've reached everyone on the planet and are resending invitations to the same people. Response rates naturally fall.

• Invitations are useless to people who already signed up.
• No matter how many times you send to disinterested people, it won't work.

When response rates fall and the product has no retention, you end up losing most users after an initial spike.

The author presents several metrics that indicate whether a product can stick around and grow sustainably. Without meeting these, no matter how high your viral factor, the product will eventually fade.

• Cohort retention curves flatten out (stickiness)
• Active users / signups > 25% (validates market size)
• Power user curve shows a smile shape (strong core user base)
• Viral factor > 0.5 (enough to amplify other channels)
• DAU/MAU > 50% (part of daily habits)
• Comparison against market (network effects)
• D1/D7/D30 exceed 60/30/15 (daily usage frequency)
• Revenue or activity per user grows over time (deeper engagement/habit formation)
• 60%+ organic acquisition at real scale (zero CAC is better)
• For subscriptions, 65%+ annual retention (retaining paying users)
• 4x+ annual growth rate at peak metrics

These metrics are strong signals that even when you generate lots of viral users, the product has enough stickiness to keep them. Without stickiness, any user spike will be temporary. This is why many Web 2.0 apps acquired millions of users via viral loops but ultimately failed to become successful businesses—they lacked retention.

2. The Two Types of Viral Loop ✌️

Viral products generally fall into two categories:

Category 1: Simple, Highly Viral Apps

Apps that focus on one very simple feature that is highly shareable or drives lots of invitations. Generative AI creative tools, early Instagram, and early YouTube are examples—products that often succeed overnight.

Category 2: Deep, Sticky Apps with Sharing Features

Products with complex functionality, high retention, and some sharing capability. Figma, Slack, and early Facebook fall into this category—they tend to grow slowly and steadily.

These two types are very different.

Category 1 apps have simple, high-conversion viral loops where the only thing users need to do is create viral content in the app. Early Instagram's focus on photo filters and Facebook sharing is the classic example. These apps are easy to build and are often the most successful during the early days of a viral platform.

Category 2 apps are complex to build but highly sticky, so once you acquire a user, they tend to stay for a long time. The viral factor accumulates gradually as the sharing feature is exposed to users across many sessions over time. It doesn't produce explosive growth like Category 1, but its consistency is its strength.

Current AI apps share the Category 1 viral loop characteristics of early Instagram and YouTube. They benefit from simplicity and fast growth, but also share the weakness of low retention that can leave them as temporary spikes. Is history repeating itself? 🤔

3.1. Step-by-Step Breakdown of a Viral Content-Creation Loop 📝

A simple content-creation loop works like this:

• You discover something great online.
• You watch it. (e.g., 50% view rate)
• You click the link to the content-creation tool. (e.g., 10% CTR)
• You use the tool to create your own content. It's great! (e.g., 20% creation rate)
• You share it on social media (or in a chat). (e.g., 50% share rate)
• More people see it. (X people view it)
• … and the loop repeats.

The basic math of this loop is that it goes viral when 0.5 × 0.1 × 0.2 × 0.5 × X > 1. Since the product of those first four numbers is 0.005, at least 200 people need to see the content for it to go viral. These figures are highly sensitive—small changes can dramatically shift the viral factor.

The multiplier effect of the viral factor looks like this:

• 0.1: 1.11×
• 0.25: 1.33×
• 0.5: 2×
• 0.75: 4×
• 0.9: 10×

Mathematically, the infinite expansion across first, second, and third generations is calculated as 1 / (1 - v) (where v is the viral factor). So a viral factor of 0.5 means 100 users produce 200 total users, with 100 of those being "free" users obtained through virality.

In practice, getting the viral factor above 0.5 is critically important. Below that threshold, it barely matters. A clear strategy, then, is to make the app extremely shareable and simple, with almost nothing to do beyond create and share.

If it works, it's extraordinary—but these apps often end up as temporary spikes. They trend on social media and then traffic craters the next day. If you managed to convert enough users to subscriptions in the meantime, you'll at least have some recurring revenue (ARR) to show for it. 😉

3.2. Why Viral Performance Degrades Over Time ⏳

The performance of these viral loops naturally tends to degrade over time. Here's why:

• Novelty effect: When a new product class (e.g., generative AI content tools) appears, users initially respond at very high rates. More people watch, try it, and are receptive to content. Over time, the novelty fades and users require a higher bar to engage with the same behavior. Think of 2D image generation AI—it was stunning at first, but now it's unremarkable.
• Market saturation: The more successful a viral loop, the more it exhausts its total addressable audience. Early on, 10 invitations might reach 10 valid targets; once the market is saturated, many invitees have already signed up, and the viral factor can fall by half. Later-arriving users also tend to be lower quality compared to early adopters.
• Incumbent platform regulation: Every viral loop runs on top of an existing platform—email, Facebook, YouTube, TikTok, etc. If your content floods the platform with certain watermarks or linkbacks and the platform dislikes it, the platform can start restricting those behaviors. Conversion rates drop sharply, the viral factor falls below 1, and overall growth slows. Platforms often regulate this aggressively because they're building competing solutions or simply don't want to be taken over. (Think of the Zynga vs. Facebook platform wars of the Web 2.0 era.)

3.3. The Weaknesses of Overly Simple Apps 🤔

Despite these weaknesses, excessively simple and highly viral Category 1 products can succeed. YouTube and Instagram are the examples—the entire app could be explained in 3–4 screens of UI. Many features were added in subsequent years, but the simple core remained. Thanks to the network effects generated by enormous quantities of content, a small app with a deep content base can remain endlessly compelling. This is the magic of network products like messaging apps, user-generated content platforms, and social networks.

4. Viral Loops in the Modern Era: Retention Is King 👑

What passes for "viral marketing" on social media these days typically includes one-off tactics like:

• Startup rage-baiting and shitposting 😡
• Cinematic-quality launch trailers 🎬
• TikTok video clipping 🤳
• Out-of-home ads designed for social media amplification 📢
• Influencer marketing (astroturfing) 🌟
• Founders becoming influencers themselves 🗣️

These are fun, but the author predicts they won't sustain their effectiveness as long as classic viral techniques like referral programs, share links, and invitations. The reason is that these one-off tactics fail to continuously scale the ratio of new users to daily active users. In other words, as DAU grows, new user counts need to grow proportionally—but these tactics only generate a one-time traffic burst. No matter how good the retention, DAU grows only linearly; exponential growth is out of reach.

These techniques are effective for temporary spikes, but they're not suitable as ongoing user acquisition sources. Releasing a great launch trailer once a year is fine, but doing it every month or every week diminishes returns. The same is true of rage-baiting—once everyone does it, nothing stands out.

That said, the author acknowledges these strategies are still useful—for reasons explained below.

4.1. The Death of Web 2.0 Virality 📉 and the Arrival of Mobile

Web 2.0 saw enormous viral innovations—email invitations, content sharing, Facebook apps—but ultimately mobile killed them. The Web 2.0 golden age worked because receiving an email or notification from a friend was a new experience that generated high response rates, allowing viral factors to exceed 1. That's why Facebook, LinkedIn, YouTube, Spotify, Pinterest, and many other products were able to launch via clever viral loops.

In the early days, apps like BirthdayAlarm (2001) and Plaxo (2002) grew rapidly by integrating address books and inviting hundreds of friends via email. They were very simple, but they helped give birth to the social networking app category by combining social profiles with feeds. What mattered was that these apps didn't stay mere viral apps—they had genuine utility and retention.

Over time, consumers grew accustomed to these techniques, response rates fell, and spam filters kicked in. And critically, the world shifted to mobile. When email virality dominated, users could import their address book and send hundreds of email invitations at once—viral factors of 2× or more were possible.

Mobile was completely different. Apple made mobile contacts accessible, but required users to invite contacts one at a time. Who actually does that? Attempts to send invitation SMS messages from servers using services like Twilio led to SMS spam—and millions of dollars in fines eventually ended those experiments. The platform shift from email to SMS virality, combined with declining novelty of invitations, caused response rates (and viral factors) to plummet.

At this point, the era of the excessively simple, highly viral app came to a close. And achieving a viral factor above 1 in a single session remains virtually impossible to this day. 😥

4.2. Retention Is the True King of Virality 👑

Modern apps don't rely on blasting invitations the way previous generations did. Instead, they operate on a few key principles:

1. Diverse top-of-funnel channels
2. Excellent retention that drives virality

First, you need diverse, organic user acquisition channels. Social media, video launches, press coverage, SEO, and paid marketing can all funnel users into the product. If these channels continuously drive users into the product, even one-off spikes are fine.

Take the Uber app as an example: roughly 50% of first rides came from paid marketing, 10–20% from the referral program, and the rest from word of mouth, SEO, and so on. The paid-ad users cost $10–$20 each, but they were worth it. The key is that some kind of user acquisition source must keep working.

Second, the product needs to generate many user sessions—that is, it needs strong retention. The viral factor is sometimes simplified as invites × conversion rate, but this assumes all virality happens in one session. In a product with high retention, users can be guided to share, invite, and refer across many sessions.

Total Viral Factor = (Viral factor from session 1) + (Viral factor from session 2) + …

This can be thought of as the sum of every point on the retention curve, where each session can generate a small amount of viral factor. The viral factor for each session is determined by the proportion of users who interact with viral features and the resulting share, invite, and conversion rates.

The author offers a rule of thumb: roughly half the viral factor is generated in the first session, and the rest across all subsequent sessions. In the first session, users are in an account "setup" mode—easy to prompt to invite colleagues or friends, with high intent. In the second or third session, users are in a different mindset, expecting value from the product, making it harder to route them into viral flows—but the opportunity is still there, repeatedly.

In practice, many apps run multiple loops simultaneously. Dropbox had several:

• Sharing folders with colleagues
• Inviting people to shared folders
• A referral program
• Other Dropbox apps (each with their own viral loops)

These loops operate at different levels, but the key point is that users can be persuaded to participate in all three across many sessions. At Uber, beyond the referral program, various loops operated—inviting friends to ride together, or exposing friends to Uber via the "Share ETA" feature.

In short, viral factor isn't determined by the number of invitations sent in one session—it's generated by the sum of all viral features engaged across all sessions. The more loops you have, and the more sessions good retention gives you to deploy them, the more viral your product becomes over time. And the more novel and exciting the product is to users—as with AI products—the easier it is for the whole system to work together.

4.3. Low-Retention Apps Must Be Spammy. Sticky Apps Don't Have To Be. 😇

When a product has high retention, there are many sessions in which to ask users to share or invite others. Even a small, unobtrusive viral sharing feature can be enough to push the viral factor above 1. Conversely, a product with low retention that averages only 2–3 sessions per user must use highly visible, spammy tactics to push users toward viral features at all. This is why sticky, high-retention apps can grow more viral over time.

The author recalls that early Facebook was far less spammy than competing social networking tools of its time. There was an email invitation feature, but it lived quietly in the right rail of the website—most users were never forced to invite friends. The author attributes this to Facebook being a highly sticky, well-built product from the start, with a viral factor that was built up gradually over time rather than through short-term spam tactics. Ultimately, this is part of why Facebook won out over competitors who drove user frustration through low retention and spammy behavior.

4.4. Is a Viral Factor Below 1 Still Useful? 🤔

Many people dream of exceeding a viral factor of 1.0, but the author believes those special moments only appear during very brief windows when a new platform or novel mechanism emerges and ultra-simple viral apps can thrive. Most of the time, viral factors are 0.2 or 0.3 or lower.

But that's still valuable! A viral factor of 0.2 means that for every 1,000 users who sign up, you get 200 users "for free"—which substantially discounts your CAC (customer acquisition cost). In this way, viral loops serve as a supporting mechanism that multiplies the effectiveness of your marketing spend. A high-retention product will continue to grow its registered user count in proportion to its active user base, even if the growth isn't fast.

This is where the concept of viral loop "speed" becomes important. Especially for high-retention, low-spam products, virality may not show up quickly. If the viral factor is built up slowly across many sessions, the rate at which the viral loop generates invitations is also slow. A product like a social network that's used daily means users can send invitations every day, letting the viral factor accumulate quickly and generate signups fast. But something like Dropbox's referral program—useful, but perhaps used once a month—can take years to generate viral signups, even if the product is very sticky (and Dropbox did indeed acquire hundreds of millions of users). The loop velocity is simply slow.

This kind of viral loop is slow but powerful, and can grow an already large user base even larger. This becomes especially important in later stages when you don't want to rely extensively on paid marketing. In consumer or prosumer markets where you need to acquire hundreds of millions of users, it's hard to buy your way to that scale through paid marketing alone. Instead, spending millions on marketing while benefiting from the discount effect of organic distribution (SEO, ASO, etc.) and viral loops is the right approach.

4.5. Is the Era of Rage-Baiting Posts and Memes Here to Stay? 🎉

So what does today's viral marketing landscape look like?

If you believe the framework described above, rage-baiting posts, memes, and cinematic videos only help create one-off user spikes. They're not repeatable—but that's okay. The reason is that these signups get amplified by the characteristics of today's AI tools.

The current generation of AI tools often have a "create and share" viral loop built in, which means any type of user that enters the funnel can be amplified. If users can create something genuinely new and original—music, video, or anything else—using an AI generative model, many of them will. And since many of those who interact with the created output will want to share it with friends, that share rate tends to be high as well.

This, the author believes, is why visually compelling AI tools tend to have enormous viral potential. They leverage the "create and share" loop, and we also happen to live in an era of visual social media—short-form video, clips embedded in posts, and so on. The characteristics of generative AI output align very well with what performs well on social platforms. That means broad distribution will follow.

Conclusion: Pursuing Sustainable Viral Growth with Retention and Multi-Loop Strategy ✨

The simple, direct viral loops that drove explosive growth in the Web 2.0 era have lost much of their effectiveness in the mobile age. User fatigue, platform regulation, and a new technological environment have all played a role. Achieving a viral factor above 1 in a single session is now nearly impossible.

Viral growth in the modern era is evolving away from reliance on one-off marketing tactics toward a model of robust retention and diverse viral loops that expand the user base over the long term. Top-of-funnel strategies—early rage-baiting campaigns or visually striking content—remain valid tools for driving initial user acquisition, but the core is a sticky product experience that naturally generates and amplifies viral loops across many sessions within the product.

AI tools show high viral potential as their "create and share" loops mesh well with the characteristics of visual social media platforms. But to avoid being a temporary trend and achieve sustainable growth, the path forward requires going beyond an initial user spike to secure long-term retention—and carefully designing a multi-loop viral strategy where users naturally draw others into the product the more they use it. In the end, never forget: retention is the true king of virality. 👑