The Hortonworks IPO, not adding up for me. #BigData #Hadoop


There’s Money In Those Elephants….

I have to admit I got a little excited when I saw a small piece on the newsletter about Hortonworks filing for an IPO. Are we to see a Hadoop company push themselves into the big wide world, finally a time for elephant wings to be unclipped and set free in the enterprise loving stock trading folk?

For me though something didn’t sit right, looking to raise $100m via an IPO doesn’t seem enough. Why? Well look at the competition around you. The likes of Cloudera already have raised $1.2bn over seven rounds compared to Hortonworks  $258m in five rounds.

At present Hortonworks operates on a loss, this surprised me a little (well a lot) with the first nine months of 2014 with revenues of £33m and an operating loss of nearly $88m. With over 250 customers the company are looking at some long term goals to become a $1bn company into 2017/18.

At first I was surprised at such an operating loss. When you think about it though Hortonworks are still in product development mode with a lot of engineering working on the bespoke components of their ecosystem. That’s probably where the money is being spent.

According to Mattermark with an annual burn of $120m in the previous year and $42m cash at present, there’s an urgency to raising some money.

BigData is Long Play

If you look at the major Hadoop distribution vendors out there at the moment: Cloudera, Hortonworks, MapR and EMC, they all have similar features.

  • Their offering is based on an open source stack (Apache Hadoop) with modifications or feature enhancements.
  • The business model is akin to the old RedHat model, selling on support. Download for free but if you want training and help then you’ll have to pay for it.
  • The addressable market for large scale customers is i) average and ii) easily transferable between vendors.

We’re still in the early days of Hadoop and BigData in the everyday lexicon of enterprise software. Right now we’re in the opening acts of the show, still ironing out the toolset, “yes we will need to figure out how to SQL transactions over the cluster as that’s what our customers are used to”.

And when we say “Hadoop” what we really mean is:

  • Hadoop and YARN
  • Hive
  • Impala/Drill/
  • HBase
  • Kafka/Flume/Storm

Plus a continuing list of technologies that are handy to know to get a solution in place. A steep earning curve (I’ve been using Hadoop for three years now and I’m quite cool with it all) means that there’s an opportunity for support and consultancy.

The big customers will always go to the big boys. There’s an opportunity for the mid tier but I’m not going to talk about that.

Hortonworks! $100m! Really!

Vendor support is going to be a long play at least 5 years but more than likely into 10 years. So $100m isn’t going to be much of a runway. I firmly believe they need to capitalise on their “we run on Windows” (Microsoft is a big customer) play a lot harder and if possible have another serious Intel-like amount investor so Hortonworks have the means to stand shoulder to shoulder to the likes of Cloudera (Let’s be honest $1.2bn is a lot of bacon and data is the new bacon).

Perhaps Yahoo! could return to the fold, Marissa is always on the lookout for great companies to put money into. And with all that data from Yahoo! News crunching from Summly and cat pictures from Tumblr, well Hadoop processing is going to fit in real nice. Let’s not forget that Hortonworks was a Yahoo! spin out from 2011.

Marissa, company shopping, yesterday.
Marissa, startup shopping, yesterday.

Do You Need A Vendor?

When you are dealing with an open source technology the question will always remain, “why do I need to pay you when I can get it for free?”. Now some companies like the comfort blanket of paying for support. It’s there and we access it any anytime.

With Hadoop there’s engineering effort involved. And for the larger companies they have the money and the ability to invest in their own engineering effort and not rely on the consultancy and support from the big vendors. If a large company decides to partner with a vendor it’s either down to technical gap filling or PR opportunity.

So…. Finally

Back to the main topic of this post, Hortonworks. Will they be a $1bn company by 2018? I’m not so sure, this is a long play and I wouldn’t expect to see the fruits of investor labour making Hortonworks into a $1bn company until 2020/21. But I’m prepared to be wrong. If I had all the data I’d spin up a Hadoop cluster and know for sure 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: