This is the End of Hadoop as We Know It (And I Feel Fine!)
The stocks of both companies soar instantly after market and as soon the announcement of the merger was made. Overall market participants seemed pleased by the outlook of a companied company. Across the media widely positive reactions sprawled and Forrester for example is sure that this is a “A Win-Win For All”. So all fine!?
Well, there is at least one that strongly disagrees with this assessment, although this might mainly be because his job title suggests so. In this regards did the CEO of MapR, John Schroeder, say:
“I can’t find any innovation benefits to customers in this merger”
Apart from this almost sole opinion, is this deal really the success story everyone believes it to be? No, it is not in my opinion and as the dust settles it be comes more and more obvious that this deal is surrounded by dark clouds.
Last Resort: Hybrid
A lot about the deal has also to do with timing and must be seen in the context of a series of recent announcements surrounding both companies. Hint: Cloudera has it in its name and it was the original strategy of that company from the beginning, but never really successfully executed on it.
But before we look into the broader context of the recent announcement, let us reconfirm why companies do mergers at all. Strategies by which companies do mergers are as such:
- Increase Supply-Chain Pricing Power
- Eliminate Competition
It it quite obvious that this merger is either an effort to combine synergies (1) or to eliminate competition (5).
A strong indicator for the latter would be a high premium for the shareholders of the target company. It is probably fair to assume that in this case the target company would be Hortonworks. But does this deal hold a premium for Hortonworks shareholders? Here are the details of the proposed deal so far:
- Cloudera stockholders to own 60% (evaluation before deal ~ $2.1 Billion)
- Hortonworks stockholders to own 40% (evaluation before deal ~ $1.8 Billion)
- Hortonworks CEO retires to board
Doing the math:
- Cloudera: 2.1 / (2.1+1.8) ~ 54%
- Hortonworks: 1.8 / (2.1+1.8) ~46%
From pure numbers and structure this doesn’t look like a (high) premium to me. Does it to you? So this merger is a clear synergy strategy! Good!!
Because Minus Times Minus is Plus
Combining synergies are good. Synergies increase the performance and lower the costs, which is very likely the main reason why this deal was so well received by the market and the main moitviation behind it.
But here comes the fly in the ointment that spoils the soup for its investors:
Hortonworks and Cloudera have failed and will likely never be able to show, that they have a sustainable and profitable business model!
And here comes the timing of the deal that I mentioned earlier, both companies join forces at a time where they are trading below the last private offering. This is a deal underwater for both companies.
Both also never before recovered truly from their initial public offering and have never in their history generated any profit.
At a time where both companies were fully committed to achieve profit and demonstrate a healthy business case, suddenly they raised a white flag and acknowledged their defeat.
So, this deals seems like the last resort to actually achieve sustainability. It appears as if both companies realised, that they would never be able to achieve profitability unless they join forces. And the new company success can only be measured by this promise.
Together, Cloudera and Hortonworks Stand a Better Chance of Profitability
Together, Cloudera and Hortonworks stand a better chance of profitability, but this will have an effect for utmost to the employees of both companies, but will likely also hold consequences for different Apache projects for example.
Surely we will also see that R&D will get trimmed. Will for example the ORC file format (Hortonwokrs) be sacrificed in favour of Parquet (Cloudera)? Will Hive (especially LLAP – also Hortonworks) still be strategic for the combined company?
The only thing certain is the uncertainty that comes with this deal and that this clearly marks the End of Hadoop as we know it. But I feel fine about it. How do you feel about it?
Kubernetes won – so now what?[*]
Another aspect about the timing of the deal is part of a series of announcements lately made by Hortonworks:
Looking at the list of announcments and cadence it becomes clear that Hortonworks is in a hurry to catch up.
Cloudera was founded around the concept of an cloud offering for Hadoop, hence the name. But it seems like it failed to deliver. Hortonworks for long just tried to tie itself to Microsoft and just in recent years extended offerings for other cloud vendors. And now looking by the latest announcements of both companies this is just the next attempt to survive the all-consuming cloud wave.
Other companies like Qubole, Databricks and Snowflake seem already better positioned to take on that business. Even SAP with it’s distributed processing solution SAP Data Hub for Kubernetes seems better potitioned in the new wave of data processing. MapR also claims:
“MapR has been innovating and delivering a better data platform for years, …, and deliver containerization with Kubernetes.”
So Hortonworks and Cloudera are defeated and Kubernetes has won. So what now?
The End of Hadoop as We Know It
I personally don’t like the in my opinion senseless talks about Hadoop being dead. Hadoop to me means much more than just comparing S3 with HDFS or the like. And I would not like to see this article to be associated with such discussions.
But to me it’s clear the the merger of Hortonworks and Cloudera marks an end to an era, hence my statement. But at the same time I am optimistic about the future of Hadoop (whatever it means to me) and also about the future of a combined Cloudera-Hortonworks company, if they are able to adapt and show profits rather sooner than later, therefor: