In one sense the timing of Hortonworks IPO filing couldn’t have been better. It came on Monday, just a week after Forrester analyst Mike Gualtieri said that Hadoop in the enterprise was no longer optional.
“The jury is in. Hadoop has been found not guilty of being an over-hyped open source platform. Hadoop has proven real enterprise value in any number of use cases including data lakes, traditional and advanced analytics,” he wrote in his Forrester blog.
He went on to say that the dazed and confused CIOs who haven’t had Hadoop on their agendas thus far will get with the program in 2015.
So if you’re one of the three independent commercial vendors providing software and support for the open source big data muncher, it’s highly likely that your business is going to take off. Especially because there’s not an analyst, or even a competitor, we can find who thinks Hadoop is a winner-takes-all-market.
And it’s a big market.
A Big Market (That's Only Growing)
Tom Reilly, CEO of Cloudera, a provider of hybrid open software solutions around Apache Hadoop, points out that the market sizing number included in Hortonworks’ filing is $50.2 billion (Allied Market Research), growing at a CAGR of 58 percent. He adds that 451 Research has published similar growth rates for the Hadoop market.
“Regardless of the source, the data-related segments (data management and data storage) of enterprise software comprise a multi-billion dollar market rapidly growing, driven by analytical workload expansion,” he says.
And when Reilly says “rapidly growing” he certainly knows from where he speaks. A report issued by Deloitte earlier this week says that Cloudera is growing at a rate of 4439%, well ahead of other start-ups like Box which gets a lot of attention.
Cloudera, by the way, is the only Hadoop distro provider that made the list.
John Schroeder, CEO of MapR, which sells a proprietary Hadoop solution, sees the Hortonworks filing as a move that speaks positively of the Hadoop market overall.
“It’s proof positive that the market is maturing and good for investment,” he says.
Needless to say, both Reilly and Schroeder believe that their solutions around Hadoop offer customers more value than what Hortonworks (or each other) have to offer, but that’s good for the market in general because it represents choice. And the three primary Hadoop vendors have very different approaches.
Hortonworks provides a completely open source distro of enterprise grade Apache Hadoop which is free for the taking. They make their money by providing support and services around it. Cloudera provides its own open source distro of Apache Hadoop, as well as proprietary software solutions that extends it and promises to make it more secure, easier to query and so on. They also provide education, training and other services around it. MapR sells proprietary software built on open source Hadoop and other open source projects.
Which is best? No vendor in the Hadoop ecosystem is willing to pick sides just yet. And perhaps there is no need to.
Ben Werther, CEO of Platfora, the big data analytics company, says that his company works with all three Hadoop distros, that there is no clear winner yet, and that at this point that’s not a problem. “The potential for the Hadoop market is vast. Five years ago, very few people even knew what Hadoop was,” he says.
From an overall usage standpoint, the opportunity for big data vendors may be bigger than we think because this goes beyond greenfields and into the traditional data management space. A space which is experiencing a substitution effect led by both Hadoop and NoSQL providers, according to George Mathew, COO of Alteryx, a provider of big data analytics.
Matt Pfeil, co-founder of DataStax, the company behind NoSQL Apache Cassandra (and not a MapR competitor) always points out that the opportunity for third platform companies -- like his -- is to unseat the incumbents (in DataStax’s case, Oracle). Of the Hortonworks IPO he says, “It's definitely a sign that traditional technologies are evolving into new ones.”
And this is why the DataStax team, for example, isn’t all that interested in trading barbs with other NoSQL startups.
Why fight over a crumb when there’s a whole pie to eat?
Pulling For the Next Generation
This is the general feeling we’re now getting from the big data community as a whole. So much so, in fact, that the “Hadoop wars” seem to have quieted down and Hortonworks competitors are pulling for a big debut.
“The market opportunity is enormous and we expect there to be multiple winners in next generation data management,” says Reilly.
Though not speaking of the Hortonworks filing, Diane Bryant, senior vice president and general manager of Data Center Group at Intel, made a notable statement while referring to her company’s hefty investment in Cloudera earlier this year.
Commenting on the potential of the Hadoop market as a whole, she said, “In 2000 Intel saw Linux coming and invested heavily in Red Hat; in 2005 we saw virtualization happening and invested in VMware; in 2008 we started investing heavily in hyper-scale computing. We think big data and Hadoop will dwarf all of them.”
It Takes Time to Grow From Scratch
While there were more than a few market watchers who were a bit taken aback by Hortonworks’ financials because they reveal that the company is burning a lot more cash than it’s bringing in, those who are closest to, or actually in the market weren’t shocked by its balance sheet.
“Honestly, I expected revenues to be slightly higher, but remember that Hortonworks hasn't been selling product that long,” says Nick Heudecker, a research director who focuses on information management and big data at Gartner. ”It takes time (and money) to develop a sales pipeline, recruit and train staff. Those aren't easy or fast activities,” he adds.
He also points out that the Hadoop as a commercial endeavor is really only six years old and that Hortonworks is much younger than that.
While MapR’s Schroeder says that Hortonworks has to be “pretty brave” to go to market with “those financials,” he also admits that his company and Hortonworks have different needs for capital.
Hortonworks’s “product” is providing services and support around the constantly evolving and improving open source Apache Hadoop. It’s a more labor and man hour intense proposition than that of a software company that builds something once and resells it. (Think about buying plots of grass for your front lawn versus hiring a lawn care company to plant heirloom seeds and care for your garden.)
The sentiment in the market, overall, is that now is the time to plant seeds and nurture young growth rather than to pick fruit. It’s for this reason that Hortonworks wants to raise capital.
And given that no one we spoke to said a single bad thing about HDP, their open source Hadoop distro, or its ability to provide services with the market being as big as it is, this might be exactly the right time to file for an IPO.