What You Don't Know About Dropbox's Growth Strategy To $10 Billion

Dropbox is one of the celebrities of the growth world. Like Airbnb and Hotmail, it merits a mention in nearly every “growth 101” book or article ever written. And well it should - Sean Ellis identified many of the most foundational growth principles during his tenure as head of growth at Dropbox¹.

But as much as we’ve heard about their viral video, their epic referral program, and their invite-only launch, I started to wonder if perhaps we were only seeing part of the picture. Sure, those tactics worked out quite well for Dropbox, but what were the guiding principles that lead them there?

So today let’s go down the rabbit hole and explore the lesser-known details of Dropbox’s rise to success.

Interlude: Why should I care about Dropbox’s growth?

You’ve probably heard of Dropbox and know they’ve done generally well, but what makes their growth story so special?

Within what was already a very competitive cloud storage industry², Dropbox has become one of the biggest, with a $10 billion valuation during the most recent funding round in 2014³. Focused primarily on the consumer market (though expanding to enterprise recently), Dropbox has 400 million users⁴ just 9 years since their founding in 2007.

Borrow and improve ideas

In early 2007, Drew Houston ran into a problem. Opening his computer at a bus station to get some work done in transit, he suddenly remembered that the files he needed were sitting on his desk in a thumb drive⁵. This wasn’t a problem he’d always experienced. Thinking back to his student days at MIT, he remembered fondly the intranet system that let him instantly access his files and even his desktop icons from any of the thousands of computers across campus⁶. Why didn’t something like this exist for the rest of the world?

During that bus ride, Drew wrote the first lines of code that eventually became Dropbox.

Looking for a solution to his file management problem, Drew didn’t find some eureka, why-did-no-one-else-think-of-it solution. On the contrary, the raw pieces needed for Dropbox already were heavily used by some people. As he wrote in his application to YC, “Hackers have access to these tools, but normal people don’t.”⁷

The thought that a good idea has to be never before heard of and new to the world is absolutely ridiculous. On the contrary, the most creative new ideas are generally common knowledge in one field, but completely new and innovative in another. When Drew started Dropbox, he knew the world didn’t need yet another cloud storage service. But it did need one that didn’t make people “constantly think and do things”⁸.

Growth Principle #1: The best startups aren’t new solutions. They’re standing on the shoulders of current solutions. (tweet this)

Launch powerfully (but know it’s only the beginning)

Dropbox had one of the best executed initial launches I’ve seen. They came out of the gate absolutely flying.⁹

Before they even had a product, Dropbox put up a landing page to start collecting email addresses¹⁰. Email lists are one of the oldest internet marketing tactics in the book, and that’s because they perform reliably well. When I have an idea for something, I like to put up a simple Typeform page that I can send my friends to. It doesn’t take long, and it helps get initial traction.

To get early feedback (and build buzz for their landing page), Drew put together a screenshare ¹¹ of how the product would work (long before they had a working product) and posted it to Hacker News¹². The 71 comments from potential early adopters provided a gold mine of user research.

Once they were ready to launch in private beta, Dropbox posted a short video to Digg and Reddit. This is the famous viral video that drove 70k waiting list signups in the first day alone. So how exactly do you build a “viral video”? For Dropbox, it was a matter of knowing their audience intimately. Much like Reddit today, both Digg (which was bigger at the time) and Reddit had a whole subculture of inside jokes and references.

Targeting their video exclusively to this crowd, Drew filled the video with easter eggs that only heavy internet users would notice¹³. After posting it to the website, they got 12,000 Diggs and 1506 Reddit upvotes which rocketed them quickly to both front pages and drove massive traffic (back in 2008, Digg could drive so much traffic that we used to talk about how “Digg-proof” our servers were)^{14 15 16 17}. Of course, even with all this preparation virality is never sure - success in virality is some combination of preparation, repetition, and randomness.

With their waiting list overflowing, Dropbox decided to launch an invite-only beta. The primary reason for this decision was product-driven: they weren’t sure how much use their servers could handle. But just as limiting initial users was important for their hardware, it also made more people want to join due to the psychology of restricted access¹⁸.

We hear a lot about this launch today, and it was undoubtedly wildly successful. But while it started Dropbox on the right track, the actual impact on long-term growth was actually fairly small. By month 7, Dropbox reached 1M users - a fantastic milestone, but still less than 1% of their 400M today¹⁹.

So how exactly did Dropbox drive the 99.75% of users who didn’t come from viral videos or invite-only betas?

Growth Principle #2: Optimize for launch momentum, but remember that long-term success depends on the months and years following. (tweet this)

Look for audiences that aren’t being served

While the cloud storage space was crowded in general, no free products were available for Linux at the time²⁰. Though small compared to the more common operating systems, Linux had a substantial following, particularly in the early adopter crowd. As early as Drew’s first post to Hacker News, Linux users requested Dropbox for Linux, and Dropbox listened²¹.

In September 2008, Dropbox exited beta and became available to the public. They released a Linux client the same day²².

Going out of their way to help this smaller audience paid off with thousands of downloads. When Dropbox reached 4 million users in February 2010, 5.6% of their users had installed Dropbox for Linux.²³

Similarly, only a third of their first 4 million users were in the United States - thinking internationally resulted in millions of additional downloads.²⁴

Growth Principle #3: Always go out of your way to serve your users. (tweet this)

Build a technical infrastructure that will grow with you

We very often focus on the customer side of growth - how can we get more people using the product - but that’s only half of the story. Particularly with a data heavy product like Dropbox, getting too much growth too fast could very easily have broken the product. Their invite-only beta helped manage server load initially, but after public launch they lost any control over when or how many people joined.

For the first few years, Dropbox only had one to three people working to scale the backend from 4,000 to 40,000,000 users²⁵. Adding that many zeros while keeping everything stable is no easy feat, but Dropbox focused on one of their guiding principles: the product should just work²⁶.

To make their small team scale as effectively as possible, they focused on minimizing complexity. Using Python for nearly all coding saved “thousands of lines of C code”. Standardizing hardware meant parts were interchangeable and maintenance was faster. Even getting small details right - like logging all events in UTC time (and only translating to time zones on user’s computers) - saved hundreds of hours by making troubleshooting database issues much easier²⁷.

Remember when Twitter used to go down every other day? It wasn’t a big deal (we all loved the fail whale), but if Dropbox went down it would be a very big deal. Worse, if they somehow lost files. So reliability was a very big deal as Dropbox scaled. To achieve this, Dropbox opted for popular database stacks knowing that if one of the tools broke a lot of much bigger companies would put a lot of effort into fixing it fast. Even more importantly, they spent a lot of time intentionally breaking things so they could find a fix before it broke in real life. As early head of server engineering wrote, “if something can fail, make sure it does”²⁸.

It’s important to achieve product-market fit before you floor the growth pedal, but it’s equally critical that your product doesn’t shake to pieces when you’d doing 60 down the interstate.

Growth Principle #4: Hockey stick growth is thrilling, but make sure your technical backend doesn’t collapse like thin ice. (tweet this)

Make sure you can afford to grow

In the freemium business model, free users are essentially a marketing cost. Just as it’s critical that the technical backend is ready to support fast growth, Dropbox knew they’d also need to be able to afford the cash costs of millions of free users.

To make the economics of freemium balance out, Dropbox constantly looked for small ways to save without hurting the user experience. Normally, these will be little gains here and there, but occasionally big wins are possible.

In one particularly amazing instance, Dropbox found a single feature was generating roughly half their hosting costs. The unlimited “undo” feature let both free and paid users recover deleted files forever. Storing all those files obviously ate a ton of space (and growing fast), yet very few users actually took advantage of the feature. By being very open about the change, Dropbox was able to limit the feature to paid accounts only without much fallout from free customers.²⁹

Cutting back is never as fun as building, but staying lean is critical to startup success.

Growth Principle #5: Be careful that you’re not hemorrhaging cash on features that your users don’t even care about. (tweet this)

Try everything methodically

After finding product-market fit almost from the first day of launch, Dropbox seems to have spent it’s early years (1) building growth, (2) scaling the cloud architecture, and (3) making the product even easier to use. On the marketing side, they systematically tried a variety of tactics.

Paid media is generally the fastest scaling strategy, so they thought about this channel as early as Drew’s YC application³⁰. But after spending a few thousand, costs were high ($233-$388) and profits were low ($99 gross per paying yearly user)³¹. And even if they could optimize enough to make the economics work, the volume of people actively looking for “cloud storage” wasn’t very large. If you’re on the cutting edge of a new market, you can’t wait for customers to find you³².

Turning to more creative ideas, Dropbox ran a puzzle/challenge contest giving away free storage space ³³. This drove massive engagement among their current users (nearly 500,000 entrants the first year), boosted their social media presence (likes/tweets), and drove a fair amount of PR buzz (mostly in the tech press)³⁴. Like the vast majority of marketing ideas, this appears to have landed somewhere in the frustrating grey zone between “runaway success” and “obvious failure”. It worked well enough that they ran a second contest about a year after the first³⁵, but it hasn’t been seen since 2012.

Similarly, Dropbox’s use of video drove a certain degree of success, but eventually got left behind. Way back in 2007, online video was still a very new thing, and Drew’s screencast on Hacker News likely helped his application to YC³⁶. As we know, the viral Digg/Reddit video during beta drove 70k email signups in the first 24 hours alone³⁷. And to top it off, Dropbox’s iconically simple homepage featured an explainer video for years (and racked up an estimated 30 million views)³⁸. Their willingness to experiment with a new medium certainly helped Dropbox’s growth, but it wasn’t quite the alchemist’s machine they needed.

When we’re in the doldrums of the “transition to growth” stage, it’s easy to become frustrated and feel like hockey stick growth is a cruel joke that never happens in real life. But as long as the product is good, systematic testing of marketing tactics should eventually uncover the One Ring that will change everything.

Growth Principle #6: Finding growth tactics is like playing Battleship. A systematic approach leads to victory. (tweet this)

Do more of what’s already working

After experimenting with tactics like paid media and PR, Dropbox took a second look at where their current users were coming from. After all, they still reached one million users by month seven, so something was working. Looking at the lifecycle path most users followed, they realized that word of mouth was a huge driver³⁹.

Doubling down on this already-working channel, Dropbox quietly launched an official referral program in Fall 2009^{40 41}. Inspired by PayPal’s $5 signup incentive⁴², Dropbox offered free bonus space to both the referring user and the new user⁴³.

Giving a reward to both people was brilliant because it removed the psychological barriers of either “I feel bad making money off my friends” or “my friend gets something but I don’t”. And because the reward was baked into the product, Dropbox didn’t have to worry about the hassle of affiliate payouts (though they did experiment with a full affiliate program too)⁴⁴.

This referral program grew to drive 35% of daily signups⁴⁵. That’s a massive, compounding boost in growth.

But that wasn’t the only way they improved word of mouth. As powerful as direct referral programs are, virality built organically into the product can be just as powerful. One common feature of cloud storage is the ability to share folders with friends who often aren’t users yet. Many companies (such as MegaUpload) used the shared folder page to make a little ad revenue, but Dropbox focused on a pleasant user experience and healthy promotion of their app⁴⁶.

Viral product features like shared folders drove an estimated 20% of daily signups. Not quite as massive as the referral program, but still a tremendous boost. Thanks in large part to these efforts building on organic word of mouth, Dropbox grew from 100k users in September 2008 to 4M users in January 2010. That’s a fantastic first 15 months.⁴⁷

Even today, Dropbox is still experimenting in this area. Recently they launched a new product feature that lets you invite specific people to upload files to your account ⁴⁸. It’s a great alternative to massive email attachments, and it gets the Dropbox brand in front of lots of potential users.

Growth Principle #7: New marketing ideas are great, but how can you double-down on things that already are working? (tweet this)

Remove all friction (or at least as much as possible)

The referral program and viral product feature breakthroughs were great for driving massive exposure (April 2010 had a trailing average of 2.8 million direct referral invites sent every 30 days⁴⁹), but like any funnel many of those potential users were dropping off somewhere along the way.

To boost the effectiveness of all marketing efforts, Dropbox ran a series of surveys and a/b tests to improve two primary funnels: (1) the signup flow and (2) the referral flow.

First, they researched their users from both qualitative and quantitative perspectives. Heavy investment in analytics told them what users were doing, but surveys helped answer the why. Sean Ellis was interim head of growth for many of these early months⁵⁰, and his experience doing user research here is likely part of what lead him to start Qualaroo years later to simplify the user research process.

Leveraging learnings from the research stage, they then ran a/b tests to optimize the key funnels. Today, we run a/b tests all the time, but in ²⁰⁰⁸⁄₂₀₀₉ it was a different world. Even Optimizely didn’t exist until December 2009⁵¹. Even more challenging, many critical steps (including the signup flow) didn’t happen on the internet at all - users signed up for Dropbox offline after downloading the program. But despite these challenges, the Dropbox team persevered and ran a rigorous testing and optimizing schedule.

The great thing about a/b testing is that you can achieve outsized results with even small changes. If you currently get 100 visitors a day, boosting your website’s conversion rate from 5% to 5.5% is the equivalent of driving an extra 10 visitors. And while an extra 10 visitors is always just 10 visitors, the conversion rate boost grows with you; once you get to 200 visitors a day, the 0.5% extra becomes 20 visitors.

Growth Principle #8: Achieve even a small conversion rate boost and the magic of compounding interest becomes yours! (tweet this)

Looking Forward

Between their 400 million users⁵² and their $10 billion valuation as of January 2014⁵³, Dropbox has already achieve extraordinary success in their nine year history.

Dropbox’s long-standing tradition of “just working” combined with their focus on design gives them a strong foothold in the consumer cloud storage market. At the same time, the cloud storage industry in general is still quite competitive with nearly a dozen major players and dropping server costs will likely start margin-damaging price wars.⁵⁴

Far from resting on their laurels though, Dropbox is actively expanding past the consumer space into enterprise storage. Ramping up back in November 2014, they see lots of opportunity in bringing companies still relying on locally-administered shared drives into the modern age⁵⁵, and they’ve already brought Dropbox full circle by landing MIT as an enterprise client⁵⁶.

Building vertically in the consumer market, Dropbox has also shown interest in building applications on top of Dropbox either through acquiring apps like Mailbox⁵⁷ or building their own like Carousel⁵⁸.

It feels like startup energy is still alive and well at Dropbox.