Dropbox’s Strategic Bet
“Should I stay or should I go?” This question has no doubt made its way into the boardrooms of many companies wrestling with the options of continuing their dalliance with cloud services, or building up their own infrastructure. Netflix was raised as an example of one of these companies in the previous lecture, as we explored the tension that is inevitably created when you build your service on one of your largest competitor’s platforms. As it turns out, the answer to the question for Netflix was to “stay”, citing that AWS provided them with “the greatest scale and the broadest set of services and features” (https://media.netflix.com/en/company-blog/completing-the-netflix-cloud-migration).
Not all companies come to the same conclusion. Enter Dropbox, a cloud storage company that also ran its operations on Amazon’s platform. Quite unsurprisingly, Amazon too has its own cloud storage services, including Amazon drive, and the S3 system that I explored in my previous blog post. Here we see a similar conundrum – should we continue to build atop a platform owned by a competitor? In 2016, Dropbox revealed its answer – “No”. Over two and a half years, Dropbox coordinated an exit from Amazon, and entered an era where their services would run solely on their self-designed machines, running code in a language they created themselves (https://www.wired.com/2016/03/epic-story-dropboxs-exodus-amazon-cloud-empire/).
Yet, the reasons for this exodus seem not to stem from a fear of Amazon’s vested interest in their own cloud storage services, but rather economic sense. James Cowling, a Principal Engineer at Dropbox, had this to say on the matter – “You have to think of these large [tech] players as countries – friendly neighbours…”, “Amazon is many things, but I don’t think their primary thing is being a cloud storage provider like us”. (https://www.wired.com/2016/03/epic-story-dropboxs-exodus-amazon-cloud-empire/) Let us for now take this claim at face-value, and then examine the strategic bets that Dropbox has made in departing from cloud services, so often touted in class as the economically sensible option.
The first bet that Dropbox has made: its longevity. Dropbox is counting on steady growth and the type of scale necessary to justify a move off AWS. The term that is thrown around often is “economies of scale”, with the idea being that with scale, the per unit fixed cost decreases. As one can probably guess, “scale” is the key factor here. Cautionary tales litter the Valley, however, of companies that were not able to achieve this requisite scale. Zynga, for instance, had spent $100 million to construct its own data centers, only to return back to Amazon after it was unable to sustain the growth needed to justify such a move (http://www.datacenterknowledge.com/archives/2015/05/12/zynga-ditches-data-center-plans-aws/). Dropbox has to keep expanding in order to make sure it does not suffer the same fate.
The second bet: its ability to innovate faster than Amazon. Dropbox is following the footsteps of giants like Google who built their own boxes and servers after realizing that the solutions on the market were unable to suit their business needs. In March 2016, the VP of engineering at Dropbox, Aditya Agarwal, announced that their service would now run on “super servers” that they had built themselves (http://tech.firstpost.com/news-analysis/dropbox-built-their-own-super-servers-microsoft-and-amazon-had-better-watch-out-304277.html). That is, Dropbox is betting that their servers will be able to outperform Amazon’s offerings at a sufficient degree to justify the move.
Lastly, Dropbox has also made this bet: Cloud is the future. This may seem counter intuitive given that they are leaving Amazon’s cloud service. However, it is not so much that they are leaving cloud services as it is they are becoming one. They are now becoming more and more like a platform, rather than just the software running on the platform. Developers can even start building their own applications on the Dropbox platform (https://www.dropbox.com/developers). Far from departing from the cloud, Dropbox has expanded the offerings available on the cloud.
It is a gutsy move indeed for a company to invest in its own infrastructure. Many considerations come into play – business strategy, as well as a company’s own vision of the future. Dropbox has revealed where it stands on the matter, and has made its move. Time will tell if it is right.
Users who have LIKED this post:
3 comments on “Dropbox’s Strategic Bet”
Comments are closed.
Nice blog post Aaron, I really enjoyed reading it. Thank you for selecting an engaging topic and a cool example.
It’s fascinating to follow how some companies insource and outsource different parts of the business as they go through various stages in their growth. Some moves are strategic, other moves a practical, and finally, some are their only option at the time. I’ve had the opportunity to observe some such strategic changes in direction from the inside of a company (as an employee or in an advisor capacity) and both — buying vs going for it yourself — come with significant benefits and drawbacks. You bring up an interesting example and I look forward to adding that on my list to follow.
Really appreciate the different viewpoint from the ones that seem most popular in our lectures / discussions so far. I had no idea that drop box was building their own servers. Given the economies of scale achieved by the larger players in the WSC area it seems like it is becoming harder and harder to turn their rates down and instead build a custom system.
It seems to be a race between economies of scale and the cost of computing. As these companies get larger and build larger WSC facilities they are often able to achieve levels of efficiency new players cannot hope to reach without enormous amounts of money. This enormous barrier to entry makes it seem likely there will be few large scale providers in this space. However, in “The Datacenter as a Computer” LA Barroso claims that warehouse scale computing is not the niche area it first seems. The authors argue that “the attractice economics of low-end server class computing platforms puts clusters of hudners of nodes within the reach of a relatively broad range of corporations and research institutions”. Essentially they argue that because these systems use non specialized components and the cost of computing trends down, while enormous suppliers like amazaon incur enormous CAPex costs systems to provide these services for smaller companies are not as unattainable as one might think.
Great points addressed in this post! I can certainly see an economic advantage for Dropbox to create infrastructure rather than relying on AWS. With the proper utilization and deployment size, Dropbox can have a lower total cost of ownership of the datacenter by building their own infrastructure. As you addressed in your post, they are betting on cloud and themselves. By building their own servers, they are clearly betting on their future. AWS is great for all purpose cloud computing with generalized infrastructure, but Dropbox building their own servers is a great advantage for them to have infrastructure specific for their services and applications. Not only can this decrease total cost, but it could also decrease latency and power consumption since their servers will be specially built to handle their workloads.