2014-11-December-Wall-Street.jpg
You can bet that the folks at Hortonworks’ won’t sleep much tonight, instead they’ll likely be replaying and rethinking every move they’ve made as a private company and every aspect of their strategy. As the youngest of the three primary, independent, commercial Hadoop distribution providers, they’ve gone from saying that they wouldn’t be talking about an IPO anytime soon, to suggesting that it might happen in 2015, to secretly filing in August, to unveiling the filing last month, to setting the date for the actual offering -- it’s tomorrow.

What’s the sudden hurry?

As Gartner analyst Merv Adrian puts it, “There might not be a better time.”

And, at least in the near term, he may have a point. Unless it’s possible to go back in time, that is.

Is an 'Intel Inside' Strategy Viable?

Consider that, up until recently, Hortonworks was Microsoft’s Hadoop honey. Its distro (HDP) powers Microsoft’s on premises Hadoop offering, HDP for Windows, as well as HDInsight, Microsoft’s Hadoop service in the Cloud. Not only that, but, according to Curt Monash of Monash Research, in his analysis of Hortonworks IPO s-1 filing, $7.5 million of Hortonworks’ total $19.2 million in subscription revenues (over a nine month period) came from Microsoft.

That seems to be a fairly committed relationship.

Or at least it did.

At an October press event, Microsoft CEO Satya Nadella and his deputy Scott Guthrie, invited Mike Olson, an exec from Hortonworks competitor Cloudera to the stage to talk about his intention to certify Cloudera’s Hadoop distro on Azure. Why wasn’t someone from Hortonworks on the stage, we asked at the time. After all, they were already certified. This was Microsoft’s boiler plate answer:

While Microsoft has a longstanding partnership with Hortonworks, Microsoft is continuously evaluating its customer needs and requirements and adjusting our strategy as required. This partnership with Cloudera brings a leading distributor of Apache Hadoop onto the Azure platform that will co-exist with our existing offerings with Hortonworks.”

And, if you’re Microsoft, giving customers choices is a good thing. If you’re Hortonworks, whose strategy depends on being the untainted, 100 percent “Hadoop Inside” of everything, not so much. Cloudera’s moment on Microsoft’s big stage certainly rang an alarm for some industry watchers.

Setting Off Alarms

And there had been another alarm rung not long before when Teradata, which owns 7 percent of Hortonworks, strengthened its relationship with both Cloudera and MapR.

While this was an important move for Teradata because in its quest to become the “The One Stop Shop for Big Data,” it needs to be able to offer its customers more than one flavor of Hadoop, its impact on Hortonworks’ strategy could likely be negative. Why? Because if HDP was the default Hadoop distro Teradata resold, its revenues would be far more substantial than if it is one of three.

We should note here that HDP does power the Teradata Appliance for Hadoop and that we have no reason to suspect that will change.

Change in Strategy or Ride the Wave? 

While some market watchers, and even Hortonworks itself, will say that Microsoft and Teradata’s broader embrace of Hadoop vendors is a good thing, that a rising tide raises all boats, and that there’s plenty of room for everyone … if Hortonworks’ strategy is dependent on being the de facto standard, then two very large vendors have just leaned away from, rather than into, this line of thinking.

Does Hortonworks strategy need to be revisited?

Who knows if enterprise buyers are even aware of the fact that something has shifted? After all, potential investors are likely to be studying balance sheets and looking at the tech market, in general, rather than at the Hadoop market, specifically. Especially during a week when there are four tech IPO’s -- Lending Club, New Relic, Hortonworks and Workiva -- in two days. And, hey, today’s Lending Club IPO is doing especially well.

Title image by Alex Proimos (Flickr) via a CC BY-NC 2.0 license