In Part 1 of this discussion on whether Dropbox should be in the cloud business, I shared my thoughts on Dropbox’s overall advantage in the areas of “core technology” and “product differentiation.” It’s also worth reiterating that “cloud business” here refers to infrastructure and/or platform-as-a-service, not software-as-a-service.
In this post, let’s look at the other three dimensions of building a successful cloud business -- enterprise talent, business motivation, and political will.
As I’ve noted in my post on Facebook, the tough part about building a cloud business has more to do with people than technology -- without the talent, the technology is not going to build and sell itself. In plain terms, a cloud business needs four types of technical roles: (1) engineers who build the products; (2) engineers who operate the infrastructure; (3) enterprise sales who are technical; (4) customer support and service. (Specific job titles differ depending on the company and could be misleading, so I avoid using them here.) Certainly, there are exceptional, edge case engineering talents who can fill multiple of these roles, but that rarely happens at scale.
So how does Dropbox fit on the talent acquisition front?
Dropbox has always had a stellar reputation among engineers who want to work on big, hairy problems. And this reputation is very much supported by its epic journey away from AWS as told by Wired a few years ago, and an active engineering blog that’s constantly showcasing the cool stuff Dropbox is building. One snippet of the Wired profile is quintessential to Dropbox’s reputation as an engineer’s wonderland, where James Cowling, one of the key engineers who worked on Dropbox’s private cloud, turned down a job from Google to work on its global-scale distributed database, Spanner. His "I wanted to build something" sentiment is a common refrain among ambitious, talented engineers, looking for “extreme engineering” opportunities for that addictive rush and sense of impact. To be clear, Spanner is as ambitious of an engineering feat as they come. The Spanner project and paper have inspired the imagination of many engineers and quite a few startups looking to mimic and improve upon it. I used to work at one of these startups, an open-source Spanner-inspired database company, and I’ve seen and felt the rush first hand. The only strike against Spanner for Cowling was that it already existed.
Thus, Dropbox has been and will be in a strong position to attract the kind of engineering talent to fill role types (1) and (2). As we head into lean times due to COVID-19, many startups will fail and acquiring engineering talent will become cheaper. Surely many of them, particularly the risk averse ones will head to big tech companies, and I’m not passing any judgment here on that preference. For the more risk-tolerant ones, seeking a balance of stability and ambition during what will be a volatile economic time in the foreseeable future, Dropbox is a great option. It’s big enough to provide a stable salary, but not so big and bureaucratic yet that one’s impact is limited, especially if the company takes on the ambitious challenge of building a differentiated cloud platform. (Dropbox headcount at the end of 2019 is 2,801, according to its FY2019 10K filing.)
As for role types (3) and (4) -- enterprise sales and customer support -- Dropbox will have some extra difficulties, because they will be challenging to recruit and catalyze an organizational cultural shift.
The recruiting side will be difficult, because Dropbox will be competing head-to-head with all the biggest players in the space with no obvious advantage; its unique reputation among engineers doesn’t really carry over here. In general, there will be an intense demand for technical sales people or engineers who can communicate and storytell; both of which are in short supply. It’ll be competitive all around, not just for Dropbox.
As for the organizational cultural shift, it’s a common problem for non-enterprise companies transitioning into the enterprise business world. Success stories are few and far between. GCP is going through the pain of this transition right now, which is hampering its ability to catch up to AWS and Azure. Facebook will go through the same pain if it ever decides to enter the cloud market. For what it’s worth, Dropbox is already directionally set on becoming an enterprise company, with or without a standalone cloud unit, so it’s going through the same transition pain that Google is with GCP, albeit on a smaller scale.
Dropbox has now been a public company for almost two years. At the time of this writing, it’s public market valuation of around $7.5 billion USD, 25% less than its highest private market valuation of $10 billion USD.
Its revenue growth is also slowing down. Below is a chart I plotted of the year over year revenue growth rate between Dropbox and Atlassian, a peer workplace SaaS company whose go-to-market adoption is similarly self-serve and bottom-up, and whose annual revenue level is in the same range of $1-2 billion USD. The latter characteristic is important when comparing growth rate, because your base has to be similar for the growth rate comparison to have any analytical value and not be deceiving. That’s the reason why I did not choose Slack or Zoom. I also did not choose Box, because it’s always been an enterprise-focused, high-touch, top-down sales company, thus materially different from Dropbox even though their core products can appear similar.
No company comparison is apples to apples of course, and I welcome your suggestions of a more appropriate company to place next to Dropbox. But it’s slowing revenue growth, whether in comparison or in isolation, is rather evident.
Dropbox needs growth drivers. Part of that effort is its new “Spaces” product announced in September 2019, a work collaboration software that connects other work collaboration software, like Slack and Zoom. It’s a meaningful effort, as the company moves into the enterprise and as other similar consumer products like iCloud are eating away at its base, but the Spaces's traction level is too early to tell. Furthermore, Spaces as it is currently conceived falls short of Dropbox’s own narrative of becoming a “platform”. The wording “our platform” was used 177 times in its S1 filing. Yet, the original Dropbox file sync was a single-purpose application, and Spaces only appears to be a common “glue” of other workplace tools. None of these products are what you would consider a “platform” -- a place on which other applications can be built.
Building a cloud business would not only make Dropbox the platform tech company it’s been telling the world it already is, but more importantly, give it the long-term growth and revenue driver it needs to live up to its expectation and valuation. This undertaking is not for the faint of heart and requires technical capacities, solid business rationales, and strong commitment from the top. The business imperative is there, and so are the technical capabilities. But is there political will to commit?
Judging from its recent announcement of a $600 million USD stock buyback program, Dropbox management seems to prioritize short-term stock price over long-term growth and ambition. Boosting its own stock price via buybacks may well be related to executive compensations, which is often connected to the company reaching a certain stock price over some timeline. But I don’t want to be totally cynical about this decision. If company leadership thinks the company is undervalued by Wall Street, and they can boost the company’s share price and public perception, which may help with employee morale and even improve retention, then it’s a valid managerial move to make.
The unfortunate tradeoff is that chunk of cash not being spent on product innovation, R&D, and potentially going into a huge and growing industry like cloud, where there are still opportunities to enter and differentiate. As co-CEO of Atlassian, Scott Farquhar, shared during the Morgan Stanley’s NextGen Software CEO Summit held virtually in late March, R&D dollars may yield higher ROI than sales dollars during an extreme crisis situation like the coronavirus pandemic. Of course, Dropbox’s buyback was authorized before COVID-19 became a global pandemic; no one can predict the future.
Further complicating the “political will” question is Dropbox’s frequent change in senior executives over the last year. Rank-and-file employees typically need a quarter to become productive and start hitting their strides. C-suite or VP-level executives need longer to get situated, and even longer still to take on ambitious new initiatives if they choose to. For what is worth, Dropbox’s infrastructure lead who was a key member of the cloud migration effort away from AWS is still with the company according to his LinkedIn, so some valuable institutional knowledge is retained.
At the end of the day, building a cloud computing platform as a business is expensive, complex, and needs long-term vision and commitment. When the idea for building Alibaba Cloud was percolating in 2009, the decisive factor between a go or no-go decision was Jack Ma committing 1 billion RMB a year for 10 years (in Chinese), which he followed through. That’s about $146 million USD per year, using the 2009 exchange rate of 6.83. Alibaba Cloud is now a legitimate player in the global cloud computing industry.
As I shared in Part 1 and in this post, Dropbox has some valuable advantages on its side to build a differentiated cloud business and drive the whole company to the next level. However, those advantages won’t be around forever. Whether the company will put those advantages to work and commit for the long-term is an open question.
在关于Dropbox是否应该步入云业务的第一部分讨论中，我分享了我对Dropbox在“核心技术”和“产品差异化”领域有些整体优势的看法。再次声明，这里谈到的“云业务”是指基础设施或平台服务（IaaS or PaaS），而不是软件服务（SaaS）。
作为一个公司，Dropbox在工程师圈子里还是享有很高的声誉的。几年前在Wired杂志里记载的离开AWS故事，以及Dropbox活跃的工程师博客，都随时在支持扩大这一声誉。那篇Wired文章里有一个片段淋漓尽致的体现出为什么Dropbox是个工程师的天堂。在Dropbox搭建私有云项目中的一位关键工程师 James Cowling，在决定加入Dropbox之前拒绝了谷歌的一份工作，那份工作是做谷歌最厉害的全球规模分布式数据库，Spanner。他的“我想造点东西”的情绪是所有有雄心、有才华的工程师们的一个常见的副词。他们总在寻找“极致工程”的机会，因为从中实现的影响力、获得的成就感令人上瘾。Spanner其实已经是“极致工程”里最厉害的一个项目了，它的科技和论文激发了许多工程师和许多初创公司的想象力，使他们都想模仿并改进Spanner。我曾经在这种初创公司中的一家工作过，是一家受Spanner启发的开源分布式数据库公司，所以我亲眼目睹并感受过那种“瘾”。对Cowling来说，唯一不喜欢Spanner的地方就是它已经存在了。