It would be quite reasonable to assume that the recent acquisition of Pentaho by Hitachi Data Systems (HDS) was yet another story about the managers of a successful smaller business finding the perfect exit strategy.
For Pentaho Chairman and CEO, Quentin Gallivan, the truth is a bit different, not least because `exit’ does not form part of the game plan. He sees the acquisition giving Pentaho the scalability and resources needed to expand rapidly into the new market for data integration in the big data analytics sector – and in particular its application into the world of the Internet of Things (IoT).
But it would not be that difficult to see Gallivan with at least half an eye on a longer term goal - the potential for Pentaho to drive HDS further and faster into the IoT and Industrial Internet marketplace:
We are a commercial open source analytical platform with the key capability of data integration that is needed for big data because there is a great deal of unstructured data such as clickstream weblogs, social media data and the rest. Most big data applications require the blending of that unstructured data with structured data. We’re an infrastructure business around big data, so we’re an arms merchant. We work with customers that try to get an RoI out of their big data analytics platforms.
According to Gallivan the company has 1,500 production customers, in a market where most of them are still trying to figure out where the return on their investment lies. In his view, many have not figured out a strategy beyond telling the IT department to `buy Hadoop’:
So the IT department does that, to be promptly faced by their CEO asking, `OK, and what are we getting out of it?’
The relationship between Pentaho and HDS goes back some 18 months and started as an OEM agreement, with HDS building Pentaho’s analytical platform into its IoT applications. Gallivan sees HDS as one of the industrial giants which has recognised that IoT is not just an important adjunct to its manufactured products. In his view it has also spotted an opportunity to become a significant software player in an important market sector as it starts to grow:
They see this as a company-wide strategic initiative to really help transform a lot of industrial businesses. They see the Internet of Things as the next wave, as a business venture. HDS has been charged with taking IoT to the market outside of Hitachi as well as internally.
The positive interpretation of this is that HDS has both an understanding of the need for IoT, and the ability to provide, a soup-to-nuts Industrial Internet environment across all manufacturing and industrial sectors. The company already has a partnership with Hadoop distribution provider, Hortonworks, and in that soup-to-nuts context the Pentaho acquisition can be seen as an important potential addition.
This does raise some questions, however, about the future of Pentaho’s existing relationships with other IoT players – not least of which is what one might assume is a potential arch-rival, the Software Division of major US industrial conglomerate, GE. Gallivan sees no problems here, however, and in fact is much taken with the potential that the HDS acquisition offers not just Pentaho, but its existing and future partners:
This is a broader subject than just GE. In the broad sense it is all about acceleration and expansion. We will be operating as wholly-owned subsidiary and as an independent company. Acceleration comes from the fact that I have 35 sales staff, HDS has 800. We have five data scientists, they have 200; so that allows us to scale, but they are buying into our stand-alone strategy. The expansion part of it is the IoT market, and Instead of just being an infrastructure technology component provider, with HDS we’re in the game.
This does give Pentaho scope to move in a number of directions, such as an expansion of the company’s product line. Gallivan acknowledged that he is well aware of the potential and need for increasing complex infrastructures with IoT, where physical or logical clusters of sensors are aggregated into localised data collection and control nodes - `medium-sized’ data lakes and analytical systems so to speak – where data would be collected, collated, cleaned up and even analysed where purely local issues need local management or remedial action.
But that data, or summaries of datasets, would then need to flow up through a chain of command of nodes to the central management system which, by that stage, would in effect simply be an industrial policy management environment. And yes, the word `simply’ is used for dramatic effect. The reality will be significantly un-simple.
Part of Gallivan’s positive view of the HDS acquisition is the move towards the industrial application of IoT providing such soup-to-nuts Industrial Outcomes Management, as highlighted by GE’s Bill Ruh. It is possible that Pentaho could be in a position to drive this forward within HDS:
They are looking at us for a lot of capabilities, but they are going to have their own capabilities and it will be a big customer. GE have taken some components, and we’re one of them, and they added real value. They have some real smarts.
And that is where he sees HDS heading as well, but not as competition. He sees HDS and GE as very complementary with the latter having a greater focus on IoT software applications. And because of that complementary status, Pentaho’s position as an HDS subsidiary and a GE partner could open up a role for the company as the glue between that complementary pair, although this has still to be determined.
Feeling the buzz
The very newness of the Industrial Internet sector has its risks as well as rewards, of course, and does pose the question as to whether Gallivan might find himself, in five years’ time, muttering: `damn, I wish I hadn’t done that’. He says:
Having sold the company? Not yet. We’ve obviously thought hard about it, but with the synergies we’ve got and the discussions we’ve had it is `so far so good’.
And he readily acknowledges the buzz he is getting from where the takeover can lead. Not for him the archetypal Bay area acquisition response of kick back and go sailing. Gallivan certainly seems up for the coming challenges:
Yes, at a personal level I now really think we can do this because we’ve got the scale and support, and I do want to see how this thing plays out. For example, HDS has one machine in its labs that has 100-times more compute power than we have in our entire lab. Toys matter to data scientists, and that is a big toy.
Oh, and if anyone is expecting details of the acquisition price, Gallivan is not saying, because HDS is not saying. He did, however, smile at the question:
Several analysts have made estimates, and quite a few are in the same ball park.
Make of that what you will.
IoT, Industrial Internet, Industrial Outcomes Management – call it what you will, this is going to be a huge market, and should two `ifs’ come out right, Gallivan’s enthusiasm for the future should bear fruit.
`If No. 1’ is that Pentaho can claim, long term, a goodly slice of the big data/analytics data integration market. The vast gobs of unstructured data that sensors and the like produce will take some managing for anyone to get any sense out of them.
`If No 2’ will be how well HDS manages Pentaho as a free agent which could become the focal point and driver of its own IoT aspirations. It just might find it has unwittingly engineered a `reverse takeover’ it wasn’t expecting.