I love S-1s. I know I am weird but S-1s are loaded with great nuggets of insights about business models, company performance and have fascinating stories embedded in the sometimes turgid prose.
Dropbox has been one of my favorite startups for many years. I have taught a Stanford Business School case on Dropbox in my HBS class on Launching Technology Ventures for many years regarding their scaling their sales and marketing operation. We also teach an early stage case on Dropbox to all first year students at HBS. And a few of my former students joined the company early and became key executives over the years. So, when the Dropbox S-1 came out yesterday, I was excited to read through it and see what I could learn. To be clear, I am not an investor in the company and don’t own any shares (as far as I know!)
In short, Dropbox is a cash machine and possesses a magical business model. Anyone who concludes otherwise doesn’t understand the difference between cash flow and GAAP and the power of negative churn. Let’s jump in.
Is The Business Model Working? Yes!
Dropbox has a magical business model and the data in the S-1 proves it. The company reports that “we generate over 90% of our revenue from self-serve channels”. Think about that for a minute. The free product is so attractive that it drives massive adoption and the conversion from free to paid is so obvious and smooth (more usage leads to more storage leads to paid product) that the company has a customer acquisition engine that derives from a simply great product and a compelling value proposition. Forget sales and marketing, at Dropbox, the product itself is a massively effective and efficient customer acquisition machine.
Does it cost too much to service all these free customers? Happily, Dropbox is following a cost curve of declining storage and cloud costs. Gross margins have soared from 33% in 2015 to 54% in 2016 to 67% in 2017. If a CEO tells you she is going to increase gross margins from 33% to 67% in two years on a $1 billion revenue business, you would check her in to an insane asylum. Dropbox did it easily.
The other magical part of the Dropbox business model is that it’s so darn viral. Every time I share a Dropbox document to someone, they need a Dropbox account. Every time they set up a Dropbox account, it drives more document sharing and storage. And the cycle keeps going. 100 million new accounts in 2017 is another stunning number, proving that the viral machine continues to work well despite what some may worry is a growth saturation point.
Cohorts and Negative Churn
The other compelling part of the Dropbox business model is the power of negative churn (i.e., revenue from your existing customers, on a net basis, are growing not shrinking) and the behavior of its cohorts. We see this similar behavior in one of our enterprise storage portfolio companies, Nasuni and our database software company, MongoDB. Every year, the company starts with a customer base that grows 20–30% per year on a net basis. Thus, if you theoretically fired the entire sales force and shut down all marketing activities, these companies would still grow 20–30% per year.
Dropbox’s cohorts are behaving in a similar fashion. There is a chart buried on page 63 of the S-1 that tells the story. Cohorts of customers are growing rapidly months after signing up for the service. The January 2015 cohort grew to 2x the monthly subscription amount paid two years in. The January 2016 cohort grew to 2x the monthly subscription paid after 20 months. And the January 2017 cohort grew to 2x the monthly subscription amount paid after 10 months. In a word, wow. When cohorts are getting better over time, that dramatically, something very right is going on.
But What About the Losses?
Now I know many of you are wondering — if the business model is so magical, Jeff, what about the losses? The company reports a net operating loss of well over $100 million in 2017. How can you reconcile these incongruous data points?
Many business reporters don’t seem to understand that GAAP net income is a meaningless measurement for a subscription-based business model. Dropbox users sign up for an annual fee and pay in cash up front and then the company recognizes that revenue over the entire year. Further, there are massive non cash charges that hit the income statement, such as stock-based compensation. What you really need to examine is free cash flow.
Page 67 tells the free cash flow story. The company’s free cash flow has gone from negative $64m in 2015 to $137m in 2016 to $305m in 2017. Think about that for a minute. Here you have a company growing at over 30% year over year in top line revenue that is able to achieve a positive swing of nearly $400m in annual free cash flow over two years. And they did this while improving gross margin by 34% points. Again, wow.
A few other things struck me as interesting. After raising so much in private capital, most founders can expect to own 5–10% of their company in aggregate. According to the S-1, CEO/founder Drew Houston owns 25% and his co-founder, Arash Ferdowsi owns 10% for a combined 35%. Drew and Arash built a company from zero to $1 billion, raised a ton of money, and still own over a third of it. If Dropbox ends up being worth over $10 billion, these guys will each become billionaires on paper while still retaining complete control over their company. Amazing.
Another fascinating thing to me is that it has taken 11 years since inception to get to this point. I’d be shocked if Drew wasn’t still CEO of the company for many years after the IPO. Thus, building real businesses takes a long time, even if your business model is magical and you hit every major platform trend just right (i.e., massive secular shifts to cloud, storage, big data and mobile in the case of Dropbox).
Final fun tidbit: the company dumped $11 million in cash and stock into the Dropbox Charitable Foundation at the end of 2017. An interesting use of free cash flow and a nod to interesting initiatives ahead.
When I teach the Dropbox case at HBS, I end the class with a checklist of the business model elements of the company: strong virality, strong network effects, recurring revenue model, high gross margins, negative churn, low customer acquisition costs. I guess that’s how you get from zero to $10 billion in market cap in ~ 10 years.