On Monday and Tuesday of this week, Cloudera held its annual gathering for industry analysts. The setting was lovely Santa Monica, though the packed schedule didn’t exactly accommodate time at the beach just outside. Over the course of two days, a room full of analysts covering the company were walked through the past, present and future of Cloudera the business.
The timing of the event was probably less than ideal from the company’s perspective, given that the market dropped a hammer on it last week – a subject we’ll come back to momentarily. But the show must go on, and to Cloudera’s credit, just as with the reduced guidance that precipitated the drop, the company was candid about what it perceived the issues to be as well as plans for their mitigation.
In the wake of a couple of days immersed in the company’s business, product strategy and future roadmap, here are five important questions facing Cloudera.
- Does the recent financial hit matter?
The flip answer to this is, of course. Any time something like a third of your market value is lopped off in a twenty-four hour period, it matters. But the answer is less clear cut when timeframe is considered. The short term impacts from external perceptions to internal morale are obvious, but what does this drop tell us about Cloudera’s long term future? The trailing financial performance was acceptable if not over the top; revenue was up 42% on the year, exceeding expectations. For the year ahead, however, Cloudera expects the growth rate to be around half of that. Some of that is undoubtedly sales execution, as the search for a new global sales leader suggests. But it’s also clear that Cloudera is a company in the midst of a product transition, a position explicitly acknowledged by multiple senior executives.
Technology companies, perhaps more than any other industry, operate in a harsh, Darwinian world ruled by the phrase, “adapt or die.” As obvious as the need can be, perverse market incentives that reward short term thinking at the expense of the future can make it exceptionally difficult for public technology companies to take one step back in order to take two forward. Such is the price of being a public entity, and the capital that status affords.
If you conclude that such a transition is necessary, however, taking a one time major hit is clearly preferable to inaction and likely the better approach to amortizing potentially bad financial news over a period of multiple quarters. This is particularly true given that the company has only been public for a year, and thus the market has little in the way of a track record to gauge the correlation between short term impacts and long term results.
In which case, the argument can be made that while the short term implications of the financial hit are bad, this may not have significant implications for the company moving forward – it may in fact be a positive if it gives the company the breathing room it needs to retool itself on the fly. Which begs an obvious question: retool itself into what, exactly?
What is Cloudera now?
Few remember it today, but the early history of Cloudera involved an attempt to market and sell a cloud-based product. As with online-CRM tools circa 1999, however, the market simply wasn’t ready to embrace that at scale, or certainly not at volumes necessary to grow a traditional enterprise business.
What most people remember about Cloudera, and indeed what many still perceive the company to be today, however, is the Hadoop vendor. Which makes sense, because that was the core of the initial offering, and the company employed not just multiple committers to the project but one of the project’s original authors.
But as important as Hadoop was and still is, in certain cases, the market continued to evolve. Just as Google – whose HDFS and MapReduce papers were the original basis for the Hadoop project – moved on to other approaches and techniques it found more efficient, more accessible or both, so too have vendors in the Hadoop space. Cloudera hasn’t been just “the Hadoop company” for years at this point. That’s like saying that IBM is the adding machines company.
If that’s clear, what’s significantly less so is exactly what Cloudera is today. Is it a data warehousing company, an analytical database company, an operational database company, a data science company, or a machine learning company? The answer, depending on which Clouderan you talk to, is yes.
Which is both a strength and a weakness. On the one hand, Cloudera has the ability to attack multiple markets with TAMs measured in the double digit billions. But on the other, it’s like saying the Germany of World War I was lucky because they were effectively surrounded and thus could attack in any direction they chose.
Cloudera’s broad portfolio ensures it has a wide variety of competitors, which we’ll come back to, but it also means that marketing has an enormous challenge to tackle. It must reconcile not only multiple, distinct product categories with different cultures and buyers, it has to wrap them up neatly enough for the market and potential customers to understand – quickly – what Cloudera is and what it does.
At present, the individual product marketing messages are much more tightly constructed than the larger, macro narrative, but presumably that’s an area of focus moving forward.
Who is Cloudera selling to?
This is arguably the most interesting question to emerge from the show. Historically, Cloudera has taken great pains to articulate and call attention to its open source pedigree and ongoing community presence. This was important enough for the company that it would tolerate the downsides, which included becoming engulfed in public disputes over contribution levels and similar. It has always been unequivocally an enterprise software company, but it was an enterprise company that understood and invested in community outreach and evangelism efforts.
At this year’s analyst event, however, the community and open source efforts were much lower profile, essentially functioning as the substrate. Cloudera went out of its way, in fact, to emphasize its laser focus not just on enterprise buyers, but very large enterprise buyers. Due to a variety of factors ranging from the nature of the technology to Cloudera’s go to market model, its cost of sales are high, so a focus on the types of clients who will actually be profitable is understandable and logical.
Given the accelerating importance of self-service, volume oriented data offerings from the likes of Amazon, Google and Microsoft however – vendors who are at once important partners and active competitors – it would behoove Cloudera to not over-rotate towards its large targets, but rather complement that focus with a broad, bottom up strategy designed to elevate awareness and generate more opportunities longer term. Many of today’s large enterprise data accounts on the hyperscale providers have roots in simple, kick the tires proof of concepts spun up by a single practitioner or small team.
Who are Cloudera’s competitors?
As noted above, Cloudera is unusual in the breadth of its potential competition. For vendors of everything from traditional data warehouse products to several types of databases to acronymns like ETL, BI, and AI/ML, Cloudera has something in its catalog that overlaps, even if tangentially. Some of these competitors, as in the case of data warehouse vendors, are incumbents backed by decades of trust and inertia. Others, as in the case of hyperscale database vendors, are the literal definition of disruption and have both economies of scale and massive visibility working on their behalf.
With the exception of Amazon, there are few if any companies that can expect to compete head to head in every market that Cloudera currently has at least some presence in. Which means that part of Cloudera’s transition is not just its own product roadmap, but picking its battles moving forward, identifying areas it believes it can compete in on a sustainable basis moving forward.
The market has likely priced in its concern about the hyperscale competition in particular, but while it’s impossible for any vendor to compete with them on an equal footing, the opportunity for Cloudera to differentiate itself in a meaningful and defensible way exists. Which brings us to the last question.
What is Cloudera’s future?
Cloudera clearly has ideas about where growth is going to coming from. The enormous focus on its cloud efforts, from partnership to managed platforms like Altus, are both a return to its roots and an implicit acknowledgement that the dramatic uptick in appetite for as-a-service offerings in other database categories apply to Cloudera’s core areas of opportunity as well.
Similarly, analytics was an area receiving substantial attention. One of the most common complaints about Hadoop implementations was that while scale was not a problem, extracting meaningful value from data on the platform was. Cloudera has clearly taken these lessons to heart, and from investments in analytical database capabilities to partnerships with third party analytics providers, there’s room for growth.
And lastly, there’s Machine Learning. Aside from investments in projects in the area, Cloudera made the significant decision to acquire Fast Forward Labs late last year. Given the profile and reputation of its founder Hilary Mason, it would be easy to see this essentially as an acquihire. One of the reasons Fast Forward Labs acquired the types of clients it did, in fact, was because they were unable to hire people of Hilary’s caliber themselves. But this belies the uniqueness of the Fast Forward Labs model and approach, which is quite distinct and is clearly an asset against both pure ML body shops or product only offerings. As Cloudera co-founder Mike Olson publicly stated, the acquisition was about more than Hilary and team, as impressive as they are.
All of those areas are product-centric, however. What about the macro trends within the market? How might Cloudera leverage them to grow its business in a defensible manner?
Two trends that appear to be eminently exploitable, if under focused on at the event, are the foundational importance of data and the increasing appetite for multi-cloud strategies.
As one of the slides from the event event quoted the Economist, “data is the new oil“. This borderline cliché is vital for Cloudera to not just understand but deeply internalize. Particularly as the company ramps up its ML efforts, data is more likely than software to determine success from failure for its clients. As Andrew Ng wrote in 2016:
Among leading AI teams, many can likely replicate others’ software in, at most, 1–2 years. But it is exceedingly difficult to get access to someone else’s data. Thus data, rather than software, is the defensible barrier for many businesses.
The company needs a cleaner, better articulated linkage between its past as a data platform and its future as an ML one. The good news is that there are products on the way that embrace the value of data generally and telemetry specifically, though they’re not publicly announced yet and therefore cannot be discussed here.
Data is likely to be one of the best moats Cloudera can build around its business – far more so than software – not strictly because of the volumes of it customers are accumulating, but also because of the advent of multi-cloud strategies.
While the hyperscale providers in particular enjoy many advantages over Cloudera as they do other vendors, they also come with limitations in visibility and access. Cloudera and other third party providers will always be able to traverse the variety of on and off premise environments better than the hyperscale providers can, because the incentives for cloud providers are to limit each others access to their own environments while offering software that can only be obtained from the provider such as BigTable, CosmosDB, or Redshift. In practice, this means that Cloudera can offer a middleware-like promise of platform independence to its customers. Instead of the classic “write-one, run anywhere” it’s “store once, access anywhere.” As many advantages as the cloud platforms have, they’re not likely to ever be competitive with that particular feature.
For that to represent a meaningful and sustainable advantage for Cloudera long term, its macro strategy and market message must reflect the importance of a data-oriented, multi-cloud approach. At present the company isn’t there yet – as it acknowledged, it’s in the midst of a transition. The decisions it makes during this transition and what it chooses to prioritize, however, are likely to tell us a great deal more about the company’s future than last week’s market correction.
Disclosure: Cloudera paid for my travel and expenses to the event, but is not currently a RedMonk customer. Amazon, Google, IBM and Microsoft are RedMonk customers.