In a surprise announcement made late last week, MapR is replacing CEO Matt Mills with his predecessor, John Schroeder, who cofounded the company. The sudden change reflects a back to basics move for MapR.
Mills, a 20+ year Oracle sales veteran, had been brought into MapR in 2015, and named CEO a year later, to build the sales organization. With the change, Mills left the company.
MapR has always positioned itself as a rebel. While it now calls itself offering a converged platform, not Hadoop, it still competes in a market defined by Cloudera, Hortonworks, and cloud providers with their own big data services. For now, the market can best be characterized as a rising tide lifting all boats. MapR, like its prime rivals, has reported subscription growth year over year in the 50% range. For MapR's publicly-held rivals, their stocks prices have finally recovered with recent quarters showing revenue growth and loss declines.
But the rising tide has not come without its share of heavy waves.
For one, Hadoop no longer has big data analytics to itself. Cloud providers have disintermediated the underlying platform with a plethora of Spark, machine learning, streaming, and data pipeline services. MapR and other Hadoop providers now have a growing roster of frenemies.
Then there is the persistent question of whether Hadoop is a viable market and when Hadoop companies (even if they don't call themselves that) will ever get profitable. For most providers, the first step toward getting there -- going cash flow positive -- keeps getting pushed back. When we spoke to Mills when he took over as MapR CEO in late 2016, he told us that he expected MapR would hit the cash flow breakeven point sometime in late 2017. That never happened. But it hasn't happened yet for its rivals either. Cloudera doesn't expect to hit this threshold until sometime in calendar year 2019. Only Hortonworks, which predicts it could happen in the current quarter, is bullish.
MapR has always been a fairly frugal company; for instance, it has not spent heavily on marketing. And in returning to the CEO (and chairman) post, one of Schroeder's goals is getting costs under control.
So why is the Hadoop or big data platform business so expensive? Surprisingly for a technology platform that has not yet matured, R&D is not the big culprit. In large part, the open source provenance of much of the technology has helped spread development costs.
Instead, the biggest costs have been with customer acquisition. While Hadoop has passed its first decade, it has only been an active commercial market for about five years. For on-premise adoption, implementation remains complicated, largely limiting the addressable market to those with sophisticated IT teams and data scientists. For MapR's public rivals, the costs of cashing out deferred stock options has also taxed the bottom line.
But remember, we spoke of a rising tide? The good news is that, even in an era where the cloud offers less capital-intensive alternatives, the land and expand strategy is still working. MapR reports for the previous 12 months a 150% net dollar expansion rate (customers grew their contracts by 50% at renewal time). That's similar to results from Cloudera and
MapR Hortonworks, who in their most recent quarters reported 130% and 142%, respectively.
And so the expand, rather than the land, part of the business is where Schroeder intends to focus. While MapR is not ruling out prospecting for new customers, Schroeder's goal is doubling down on the existing base. He hints that, even in 7 - 8 figure deals, there's still ample headroom for growth. "We don't need 10,000 - 20,000 customers to bring in a billion dollars in revenue," he said.