Pulling this out of the 2016 summary thread since I only want to talk about this one company. Saul wrote:

Then I have a new half-sized position: Hortonworks (HDP) at 2.3% and in 13th place. What in the world is a company called Hortonworks. Well it is an important seller of Hadoop (whatever that is). Actually, what Hadoop is is a database or a framework for distributed storage and distributed processing of very large data sets on computer clusters built from commodity hardware.

I hope it isn’t news that Hadoop was invented by Google, which naturally called it The Google File System (the Hadoop name came later), and that all the code for it is available free via Apache Open Source. There are various ways to get an Hadoop implementation, including through Amazon’s EC2/S3, Microsoft’s Azure, and the Google Cloud Platform, not to mention a slew of companies you haven’t heard of, like CenturyLink, Cloudera, and MapR. Oh yeah, Hortonworks, too. Note that what these companies offer are not all equivalent. There are all sorts of flavors of implementations and support available.

Hadoop is oriented around enormously large databases. For instance, both Yahoo and Facebook have ginormous Hadoop implementations. Whether other companies turn to Hadoop or not would be, in my opinion, dependent on whether they need such really large databases. Hadoop isn’t easy, especially the MapReduce parts, so you don’t use it if you have to, and that’s if you need a really really big DB.

So why did I buy it? Because I read the Prepared Remarks and the Presentation associated with the last earnings report…

Honestly, I don’t understand financials, but even if I did, they’d come second. First, I want to know how the company makes money and what their competition looks like. Hortonworks used to be a “Unicorn” - a private company expected to be worth over a billion dollars, but by the time it went public valuation had dropped over 40% to $592 million.

Hortonworks makes its money selling support (via a subscription model) and providing value-added services like consulting on projects or training. Hortonworks doesn’t currently have paid-for upgrades or connection modules, or hosting, etc. Those are common ways for a company to make money off of Open Source projects. Maybe Hortonworks is looking at doing that, but does anyone know if that’s actually the plan?

I have two initial impressions: First, the good: The trend is towards collection of lots of data (look at the Internet of Things, for instance), and so Hadoop will likely find new applications. But, second (the bad news) from what I can tell, Hortonworks doesn’t yet have the right business model to be profiting from this trend, and I don’t have any insight that leads me to believe they have a plan to get there. They might - I just haven’t done enough research to find it.


I did some research few months back on Hortonworks.

Cloudera, Hortnoworks, Mapr are all Hadoop distro + service providers. From what I recollect, the biggest of the 3 in terms of marketshare is Cloudera followed by Hortonworks. Cloudera is still private.

Cloudera has hybrid open source/ proprietary model. Hortonworks markets itself as pure open source. I actually like Cloudera model better as potential upside from proprietary software is much higher. But there will always be companies looking not to be locked into a vendor and Hortonworks will always have their appeal. Besides it is a very recognized brand in Hadoop world over the years and is not going away anywhere. Both have been in business for a number of years and have over a 1000-2000 employees with similar Glassdoor ratings (largely positive).

The closest analog business model to Hortonworks is RHT (Red Hat). So business model is straightforward - subscriptions + services. I am invested in Red Hat too, which at this point from Price/Cash Flow is fairly attractive and has been a great investment over the years. I see HDP as RHT in its early innings.

EV/Sales (consider deferred revenue for cloud company) for Horton works is ridiculously low for the fear of a secondary. The moment this company proves it is cash flow positive I do not see a secondary happening. Based on their burn rate they have a few quarters to go and based on management guidance secondary is unlikely at this point (of course we just have to trust the management here). Besides since many large companies are heavily invested with HortonWorks, the worst case scenario looks like an acquisition. So the way I look at it it there is likely a lower bound on losses. To turn cash flow positive, it just needs sufficient scale. Total addressable market is huge for Hadoop and it this point it is growing at 65%. It is hard to see how this will not be profitable eventually.

I purchased it around 9 and again around 7. There sure is risk with this investment, but it is potentially very high reward.


I’m not real keen on sub-$10 stocks… more risk than I’m comfortable with. I’d like to see this break $12 on increasing volume.

I’ve added to my watch list.