Open Data Science Presents Business Opportunity to New Continuum CEO
Continuum Analytics turned some heads last month when it announced the appointment of IT industry veteran Scott Collison to be its new CEO. With stints running departments of Microsoft, VMware, and Salesforce.com, Collison looks to grow Continuum’s data science platform business through a combination of enterprise savvy and open-source scale.
In a recent interview with Datanami, Collison explained how his experience in the enterprise software world prepared him for his latest gig as the head of Continuum Analytics, a growing 140-person Texas outfit that’s quickly become a driving force in the Python data science community.
“When I was approached by the folks at Continuum, what really appealed to me was the intersection of open source and platform,” says Collison, who has also worked at tech startups, SourceForge, as a college professor, and at a management consulting firm. “It’s such a powerful thing to have both of those elements in on place. I haven’t really seen a company that has those qualities since I was at Microsoft.”
Collison sees echoes of Microsoft‘s popular .NET framework, which he was involved with as the director of Microsoft’s Platform Strategy Group in the early 2000s, in Continuum’s strategy. “These are vehicles for application development that are, to a greater or lesser extent, general purpose in nature,” he says. “That’s what I find quite exciting about Continuum and the Anaconda open source project.”
Anaconda, of course, is Continuum’s flagship data science product for analysts and data scientists. First released in 2012, Anaconda functions as a distributed data analysis platform for analyzing data with Python and the array of popular Python-based tools, such as Numpy and SciPy. Continuum co-founder Travis Oliphant, who gave up the CEO role and is now president and chief data scientist, was the primary developer of the NumPy package and was a founding contributor of SciPy.
As interest in Python as a language for data science increased, downloads of Anaconda exploded. In 2015, the company had recorded 3 million downloads of the free Anaconda distribution since it first became available. That number rose to 11 million in 2016.
“Even by Microsoft standards, that’s a lot of downloads,” Collison says. “When you start seeing a company throw up numbers like that, and they’re accessing a market that’s very hot in data analytics, and you’re talking about a million Python developers, tens of millions of data scientists. and over 100 million Excel power users–that starts looking like a really good market to me.”
By handing the reins of the company over to an experienced software veteran in Collison, Oliphant and Peter Wang, Continuum’s other co-founder, can focus their efforts on creating data science tools in demand by the Python community.
“I want to be super clear. I’m not a guy coming into transform a struggling company. That is not the story,” Collison says. “The story is, this company is growing so rapidly and the two founders are such domain experts that they need to have an additional person to come in and help manage rapid growth. It’s a great problem to have.”
It will be fun to keep an eye on Continuum, which is hosting the annual AnacondaCON data science conference next week in its hometown of Austin, Texas. (Your humble Datanami editor will also be in attendance at the conference, and will be moderating a panel on AI on Thursday).
The whole concept of a data science platform is starting to take off, and Continuum will have plenty of competition from other startups looking to exploit the market opportunity too, including Data Science, Databricks, H2O, Domino Data Lab, and others, not to mention established analytic firms SAS, Mathworks, and IBM who are looking to fend off startups and maintain their legacies.
“We size the market as a $140 billion addressable market opportunity. That’s a very big number!” Collison says. “It’s growing because we see that right now there’s a preset of people in data science who are fluent in R or SAS. What is happening is that Python and Anaconda and these packages that our guys have written like Numpy are democratizing data science, and features that we have like Fusion democratize it even further.”
Fusion is an add-on to Anaconda that extends the data science capabilities to Microsoft Excel. It’s only available in Anaconda Enterprise, which also includes an enterprise data science notebook, and which starts at $60,000 annually for 10 users. Other money-making devices in the company’s product repertoire (it also provides technical services) include Anaconda Workgroup, which brings scalability and performance improvements over the core Anaconda distribution at an annual cost of $30,000 for 10 users, and Anaconda Pro, which adds technical support for $10,000 yearly for 10 users.
While Continuum has generated a ton of goodwill with its free Anaconda package, the company—which has raised $29.5 million across three rounds of funding–is also a profit-seeking entity. And so one of Collison’s other goals will be to turn more of those 11 million downloads into happy paying customers.
“One of the metrics we’re going to start tracking pretty assiduously is how many of those do we convert to commercial,” the CEO says. “We did very well commercially in 2016. We doubled our revenue. We absolutely killed it. That was on the strength of open source adoption via these downloads.”
The overall market for data science tools is growing, and so is interest in Python specifically. When you combine that with the huge pool of open source talent that Continuum can tap into, it’s a business wave that Collison is only too happy to ride.
“We as company will continue to add value to the open source platform, and so will the community,” he says “That gives you a fundamental ability to punch above your weight as a rapidly growing mid-size company.”
The other business dynamic that Collison hopes to exploit in the data science space is customer unhappiness with SAS maintenance fees, which he called Continuum’s “bread and butter,” echoing the views expressed by CEOs of other data science platform providers.
He says: “I’m happy to steal SAS’s lunch money all day long.”