settled

How Famous Founders Grew Fast on Almost No Budget

From Reddit's fake accounts to Gmail's invite-only launch and ChatGPT's no-waitlist release: the low-cost growth tactics behind some of tech's biggest products, and what still works today.

#growth#startups#marketing#case studies

Most “growth hacking” advice is recycled. The better source is the founders themselves: interviews, podcasts, old mailing list threads, and the occasional candid blog post about what they actually did when they had no money.

I went digging through a pile of these stories. Some I’d half-remembered, some a LinkedIn post sent me down a rabbit hole for. Here is what holds up, grouped by the mechanism rather than the company, because the mechanisms are what you can reuse.

1. Make the place look alive before it is

Reddit is the cleanest example. Co-founder Steve Huffman has said the team created fake accounts to set the tone and make the site feel populated. Alexis Ohanian has put the early numbers starkly: in the first weeks, the vast majority of submissions came from the two founders posting under invented usernames. Accounts say the founders kept supplying a large share of links for months.

The lesson isn’t “fake your users.” Nobody posts in an empty room, so the founders solved the empty-room problem by hand. The honest, repeatable version is seeding a new community with real content, real conversations and real people you personally recruit.

A note on ethics: fake identities worked for Reddit in 2005. Today they carry real reputational and platform-policy risk. Treat this as a story about cold starts, not a template.

2. Borrow someone else’s distribution

YouTube built embeddable video in from the start. Co-founder Steve Chen has described designing it so the player could live on any site that allowed HTML, from classifieds to auction pages. The big wave came from MySpace, where users wanted video on their profiles and had no easy way to add it. YouTube filled the gap, and MySpace’s own engineers spent months trying to block it.

Microsoft did the same thing at the deal level. When it licensed MS-DOS to IBM, it kept the right to license it to other manufacturers. When the IBM PC was cloned across the industry, MS-DOS rode along on every clone.

LinkedIn used address-book import to turn each new member’s contacts into invitations. It was effective, and also the kind of tactic that later became controversial once spammy copies appeared.

The pattern: find a place where your audience already gathers, and make it trivially easy for your product to show up there.

3. Turn every user into a distribution channel

Three of the best-documented stories are about building the spreading into the product.

Hotmail is the classic. The idea didn’t come from the founders. Investor Tim Draper pushed for an invitation to be attached to every outgoing email, and he has said the original wording was a playful “PS I love you.” The shipped version was plainer, but the mechanism was the same: every message a user sent doubled as an advert with a built-in endorsement from the sender. Hotmail launched in July 1996 and reportedly reached around 12 million users in 18 months. Draper and Steve Jurvetson wrote that it spent under $500K on marketing in that time, and that banner-style promotion on other sites brought in comparatively few signups.

PayPal paid people to spread it. Max Levchin has described giving $10 to the user who invited a friend and $10 to the friend on signup. One retrospective says the first version was $20 each way, later cut as the cost bit. Peter Thiel has estimated the program cost something like $60–70 million in the early years. That is hardly “low cost,” but it replaced advertising that Thiel judged too ineffective, and some of the money returned as invited users began sending payments inside the network.

Dropbox copied the idea and made the reward the product itself: extra storage for both sides, shown during onboarding. The headline figure, about 100,000 to 4 million registered users in 15 months, gets repeated everywhere. Two caveats are worth keeping. The end date varies by source, and referral wasn’t the only engine: reportedly about a third of users were already coming from word of mouth before the program existed. The referral program amplified a product people already liked to talk about.

4. Do things that don’t scale

Airbnb is the textbook case. When New York listings weren’t booking, the founders noticed the photos were poor. Rather than send hosts photography tips, they rented a camera, visited hosts and shot the apartments themselves. Secondary sources report that listings with professional photos booked around 2.5 times more. Earlier, they’d funded the company partly by selling election-themed cereal boxes, which is how Paul Graham first noticed them.

Airbnb also built an integration that cross-posted listings to Craigslist and reached out to people posting there. It helped growth, but it was unauthorized and broke Craigslist’s terms, and it was later dropped. Interesting history, not a template.

Tinder got its early traction offline. Whitney Wolfe went to sorority chapters, presented the app and had members install it on the spot, with fraternities following. Tinder’s technical co-founder Joe Munoz described this to Bloomberg, and users reportedly went from under 5,000 to about 15,000 during her campus tour. Campus reps later ran parties where the entry fee was showing you’d downloaded the app. The common thread is density: a few packed campuses beat a thin national spread, because a dating app is useless without people nearby.

5. Make access feel scarce

Three of the best-known launches used exclusivity as the engine:

  • Gmail launched by invitation only in 2004. Invites became a status object, reportedly even reselling online, which generated press no ad budget could have bought.
  • Facebook started at Harvard only and expanded campus by campus, keeping demand ahead of access. Reportedly most Harvard undergraduates joined within weeks of launch.
  • Slack called its early release a “preview” rather than a “beta.” Same product, different signal: beta says “might break,” preview says “you’re early.”

Scarcity works when the product is genuinely good and the waiting list is a feature, not a barrier. It backfires when people just get annoyed.

6. Remove every bit of friction

ChatGPT launched on 30 November 2022 as a free “research preview.” No waitlist, no pricing page, no onboarding. It’s widely reported to have reached roughly 100 million users within about two months, faster than any consumer app before it. The underlying technology had existed in research form for a while. What changed was that anyone could try it in seconds, then screenshot something surprising and share it. Paid plans came later, after demand was obvious.

WhatsApp went the other way on the same principle: no ads, no gimmicks, growth driven by the fact that your contacts were already on it. Co-founder Jan Koum has spoken about wanting the network to grow on its own merits.

7. Give it away to own the next generation

In the early 1980s, Apple donated computers to roughly 9,000 California schools, helped by a state tax-credit law Steve Jobs lobbied for. The net cost to Apple was a fraction of the retail value. Kids learned on Apple machines, and so did their parents’ buying habits.

It’s slow, it needs deep pockets, and it’s still one of the neatest examples of a giveaway that works as marketing, lock-in and goodwill at once.

8. Ship it, and let usage decide

My favourite from this list, and one that came to me through an a16z post. In February 1993, Marc Andreessen proposed an <img> tag for inline images on the web’s developer mailing list. Tim Berners-Lee and others preferred different approaches that kept images more flexible and more in the user’s control. Andreessen argued browsers would add inline images regardless, so the sooner there was one way to do it, the better, and shipped his version in Mosaic.

Mosaic’s popularity did the rest. The tag you use today exists because someone shipped it in the product people actually used, not because a committee agreed.

For anyone building tools: a working version in real hands beats a perfect specification that nobody can try.

9. Earn the press, then own the audience

Zerodha’s Nithin Kamath shared a story on LinkedIn that makes a good bookend for this list. In the early days he took sales calls under the name “Sachin”, because customers didn’t expect to be talking to the CEO. One day a journalist from The Economic Times called, he answered as Sachin, and had to switch identities on the spot. When the story ran, he says monthly account openings went from a few hundred to almost 500, and he decided he had found a growth hack.

He kept talking to journalists for the next five or six years. Then traditional financial media lost reach to social platforms, and he started posting on Twitter and LinkedIn himself. His verdict is that there is no comparison: telling your own story directly to your audience has been far more valuable than the press ever was. His one condition is that you have to tell it well, or you become a meme.

Two cautions. These are his own numbers, and a spike after one article shows timing, not proof of cause. And the lesson isn’t “ignore the press”. Press is borrowed attention that proves the model works. The compounding asset is the audience you own.

10. A framework instead of a tactic: Kunal Shah’s Delta 4

Everything above is a how. Kunal Shah’s Delta 4 theory is a whether. His idea is that a product spreads on its own when it is at least about four points better than the current way on a 1–10 scale. Below that, people stay with what they know. Above it, switching feels obvious, and users tell others without being asked.

It’s a useful filter before you spend a rupee or a dollar on promotion: if your product isn’t clearly better, no growth tactic will fix that.

What I’d actually take from this

  1. Start with a cold-start plan. Who are the first 100 people and what will they find when they arrive? Go dense in one place before going wide.
  2. Piggyback on existing behaviour. Embeds, imports, integrations and bundles are all ways of showing up where people already are. Build the spreading into the product, as Hotmail’s sign-off did, rather than bolting on a campaign.
  3. Do the unscalable thing first. Photograph the listings, visit the sorority, onboard the first customers by hand, then automate what works.
  4. Use scarcity honestly. Invite-only is a pacing tool, not a trick.
  5. Strip out friction. Every field, step and decision before the “wow” moment costs you users.
  6. Check your delta. If the improvement isn’t obvious, fix the product first.
  7. Use the press, but build your own channel. Borrowed attention fades; an audience you can reach directly keeps compounding.

Sources and further reading

I’ve linked the main sources I used. Several of these are retrospectives or case studies rather than primary documents, so treat specific numbers as indicative.