Open source software is accelerating innovation in technology like never before. Over the last decade, non-proprietary, community-developed software has been adopted into virtually every aspect of computing, becoming the foundation for everything from Big Data and cloud computing to mobile devices and Bitcoin.
Yet despite the amazing successes of the open-source community and especially those who use the technology (think Google and Facebook), building a business around the concept has proven to be a much tougher nut to crack. Monetizing a technology that, by its very nature is free for anyone to use, is difficult to accomplish – and all the more so when facing stiff competition from rivals who sell what is effectively the same product. There’s probably no better example of this challenge than the intensely competitive world of the NoSQL database, where a whole gaggle of vendors are scrapping to win customers from each other, as well as larger, proprietary software-wielding rivals like Oracle.
The problem isn’t so much getting people to adopt open-source technology, but rather to pay for it, writes Jeff Kelly, principal analyst at Wikibon in a recent blog post. A case in point is the MongoDB NoSQL database, ranked as the fifth most popular database overall according to DB-Engine. It has an enormous user base already, but a quick glance at the numbers suggests that MongoDB Inc. is struggling to turn its popularity into profit.
That observation is based on a Wikibon estimate that MongoDB pulled in revenues of just $46 million for 2014, meaning the company’s estimated $1.6 billion value is a heady 35-times lagging 12-month revenue. Less clear are MongoDB’s operating costs but with more than 400 employees on LinkedIn, it’s likely they are significantly greater than the current annual revenues it’s generating.
Somewhat controversially, Kelly said MongoDB’s example raises a big doubt as to whether it, and other open-source NoSQL startups, such as Basho Technologies Inc., Couchbase Inc., and DataStax Inc., will ever be able to turn a profit in a market Wikibon estimates will be worth just $1.825 billion by 2017.
“Make no mistake, building a successful open source NoSQL company is going to be a long, hard slog,” Kelly wrote.
Kelly isn’t alone in his dour assessment of the open-source model’s long-term prospects. Last year, Peter Levine of the venture capital firm Andreessen Horowitz wrote a long-winded essay in which he argued there’ll never be another Red Hat Inc., a company that’s built a billion-dollar business by selling support services atop its free software platform, and which is widely viewed as the poster child for the open-source model. Levine points out that Red Hat isn’t even that successful anyway, because its open-source support model generates a fraction of the revenue of other licensing models used by companies like Microsoft and Oracle.
Levine argues that while Red Hat is making money, it’s hugely disadvantaged in comparison to its rivals, saying “it’s nearly impossible to properly invest in product development, support, or sales the way that companies like Microsoft or Oracle or Amazon can,” even with the support of the open-source community behind it. What’s more, there’s a risk that if your open-source software becomes too popular, proprietary software vendors will simply co-opt the code base.
“I experienced this first-hand as CEO at XenSource, where every major software and hardware company leveraged our code base with nearly zero revenue coming back to us,” Levine wrote. “We had made the product so easy to use and so important, that we had out-engineered ourselves. Great for the open source community, not so great for us.”
Open-source opens doors
It may well be that there is never another “Red Hat”, but that doesn’t mean the NoSQL vendors won’t be able to succeed. Indeed, most of them already recognize that Red Hat was something of an anomaly, and that they’re unlikely to mirror its success by selling support services alone.
“The way Red Hat become successful, that model doesn’t really work today,” admitted Kelly Stirman, director of products at MongoDB. “People have too much confidence in open-source software now that you can’t just say, ‘Hey we’ve certified it for certain systems,’ and people feel like they need to buy it. That model is no longer sufficient to drive commercialization of the product.”
Hortonworks Inc. CEO Rob Bearden echoed Stirman’s comments in an earlier interview with SiliconANGLE, saying there are numerous parallel business models that can be followed to monetize open-source software, besides just copying Red Hat’s strategy. He pointed to Facebook and Twitter as two examples of companies that have done exactly that, and indicated Hortonworks intends to use parallel examples within the Hadoop space.
“We only released our first Hadoop offering 2 ½ years ago,” Bearden said. “We’ve since generated over $120M of subscription run rate during that period of time. We’ve built a business very big and very fast based on this model.”
The basic software is simply a way to generate demand for the more complex commerical offerings that vendors do charge money for, MongoDB’s Stirman explained. By feeding free software into the market, the company is able to grow interest in the MongoDB NoSQL database and refine the product further, based on customer feedback. Once you have a refined, mature and wildly popular product, then you can go about monetizing it, Stirman explained.
“If you look at our business, we have an open-source product that’s wildly popular, and we have a commercial offering that’s differentiated, and then we have a SaaS business that allows people to more effectively manage the infrastructure they’re using,” said Stirman. “People are willing to pay for those things; it’s called an open-core strategy. The idea is the same: to build your business around the open core.”
That simple premise is being pursued by other vendors too, including DataStax, the company behind the rival Cassandra NoSQL database. Matt Pfeil, chief customer officer and co-founder of DataStax, likens open-source to a kind of “freemium” business model: first you get people using the product, talking about what they like and don’t like and what they need, and then you provide the add-ons they need, at a cost.
“It’s on us as a team to prioritize those pain points and solve them with offerings from DataStax,” said Pfiel. “In the early days, that was heavy on services, because those were the immediate needs for enterprises to be successful with Cassandra. As time progressed, software has been a requirement for success, and we’ve built out a comprehensive selection of those offerings.”
The “freemium” model
The “freemium” model has been hugely successful in some markets, but Wikibon’s Kelly questions whether it will work with enterprise-grade technology such as NoSQL. The rule of thumb in the open source world is that just one out of 1,000 users becomes a paying customer for a commercial vendor, said Kelly, echoing a metric cited by MongoDB solution architect Henrik Ingo in the Open Life blog. That begs the question: is that ratio enough for open-source NoSQL vendors to flourish, or do they do something to move the needle?
Bob Wiederhold, CEO of Couchbase, said it really isn’t necessary to improve that ratio in his niche. His company is focused squarely on the big boys – enterprise customers that use Couchbase for their most critical applications.
“We expect the vast majority of our business to come from enterprise deployments of Couchbase under mission-critical Web, mobile, and IoT applications,” Wiederhold said. “They operate at significant scale and can never go down. For these applications, it is unusual for a business not to purchase the paid enterprise version of our software.”
Wiederhold holds to the logic that such customers would never try to skimp and save by using free software when it’s so critical to their businesses. The risk of relying on unpredictable community support or internal expertise to address problems that may arise is just too large for most enterprises to tolerate. Paid enterprise subscriptions guarantee peace of mind.
Another company that believes big business will always pay big bucks for the best feature set is Basho, which offers a number of premium, enterprise-grade features atop its Riak database that are designed to deliver high availability and scalability across multiple data centers. It’s these premium features, rather than just pure OSS support, that helps Basho to attact a bigger percentage of paying customers, said CEO Adam Wray.
“If you look at all the open-source community, there are around 3,000-plus companies that use our software,” Wray said. “The pay ratio for MongoDB might be a 1,000 to one, but for us its more like ten to one.”
Wray takes a similar view to Couchbase’s Wiederhold, saying that serious customers are more than willing to open their wallets if it means getting their hands on reliable technology that delivers value to their business.
“As soon as they start to scale out, they realize they’re not only going to need the licensed software, but they also want our services and support from a tier one, two, three perspective, because they’re dealing with function workloads,” Wray said. “Workloads can now be in multiple data centers, and your data tier better be able to handle that and keep it accurate.”
One NoSQL vendor was even more confident, insisting that Kelly’s one-in-1,000 ratio doesn’t even apply to its business model. Despite offering a free-to-download developer version of its database, MarkLogic Corp. claims that it’s not an open-source business at all. David Ponzini, senior vice president of corporate business development, told SiliconANGLE the company isn’t worried about arriving at a sustainable business model because it already has one
“I’ve heard [the one-in-1,000] statistic about open-source. It would be concerning for us if we were an open-source company, but MarkLogic is not an open-source company,” Ponzini said. “We are a viable enterprise company, which means we deliver a database platform customers see value in purchasing the moment they start using it.” The company’s licensing terms stipulate that the free developer version of its product is not for production use.
How many NoSQL vendors do we need, anyway?
Enterprise customers might be willing to pay to get their hands on premium features, but with all of the NoSQL vendors being so fixated on a relatively small enterprise market niche, it seems inevitable that there won’t be enough space for all of them. Couchbase’s Wiederhold agreed, saying a number of vendors will fall by the wayside over the next couple of years. That’s always the case in a fast-growing market like NoSQL, where there’s initially a large number of open source projects and venture-backed startups, but as the industry matures, Wiederhold said leaders emerge and distance themselves from the rest.
“That has already happened in the NoSQL market where Couchbase, MongoDB and DataStax are the clear leaders,” he said. “The gap is already too large and the leaders are now all making substantial investments in product development and sales and support channels, making it increasingly difficult for smaller players to survive. I suspect the rest of the pack will be acquired or will go away over the next 12 to 24 months.”
But Basho’s Wray said things aren’t quite so clear-cut, pointing out that his company has already tied up more than 200 paying customers in the enterprise. In a nod to one of his rivals, he argued that there’s even room for smaller niche players, and that’s where a vertical specialist like MarkLogic with its XML-cum-NoSQL database might succeed.
“There’s always going to be room for niche players,” Wray said. “But the reality for those who are going to get the lion’s share is that they need to have a justifiable base to work from, with large clients who can say, ‘Yes, I’ve made that bet and you should too.’”
If so, then there’s a reason to be optimistic about MongoDB’s fortunes. While the company is bleeding cash right now, its strategy of popularizing its software first has at least made it the most widely-used of all NoSQL databases.
“This is where we have an enormous advantage over others in the market, including the Hadoop distributions, because people are running MongoDB at large scale for mission-critical applications, and we meet all the security and operational requirements for fortune 100 companies to use it,” said MongoDB’s Stirman. “Thirty-nine of the Fortune 100 are paying customers in production, and that number is growing quickly.”
Show me the money
The NoSQL vendors talk a good game, but at the end of the day it all boils down to how much cold, hard cash they can make, and for all the talk of enterprise adoption, it’s still not clear if anyone is doing so.
MongoDB, which took a battering in Kelly’s initial analysis, was evasive when talking about actual profits, declining to say how much money it is losing. Couchbase and DataStax also refused to be drawn on the matter, insisting that as private companies they could not talk about finances.
Basho was more forthcoming about the current state of its business, though. CEO Wray readily admitted that the company is still operating at a loss, but insisted that that is a strategic choice. He said it’s feasible for Basho to go cash-flow-positive as early as next year if it chooses to do so, but that it is making a conscious decision to invest for the long term. Basho continues to sign bigger and bigger deals, invest in R&D and reach out the the market instead of targeting immediate profits.
“We’re not profitable, we’re running in the negative but nowhere near like our competitors are,” Wray said. “We’ve heard our competitors are burning as much as $10 million a month, but in our case we’re burning significantly less. I’m talking factors-of-ten less.”
MarkLogic’s Ponzini was equally optimistic when talking about his company’s prospects. He told SiliconANGLE the company’s annual revenues are well north of $100 million and growing in excess of fifty percent on the foundation of the $71 million it has raised, though he neglected to discuss details Nevertheless, he insisted MarkLogic is well positioned to turn a profit in the future.
“We are not worried about creating a sustainable business model, because we already have one,” Ponzini said.