2 Comments

Summary:

ParAccel’s competition all got bought, leaving the company standing all but alone as an independent company dedicated to the cause of big data. But with a solid product and a steady business channel to boost a large vendor’s bottom line, it shouldn’t be alone for long.

bouquet toss

Greenplum, Netezza, Vertica, Aster Data Systems — one by one, they all got bought, leaving ParAccel standing all but alone as an independent company dedicated to the cause of analyzing big data. But CEO Chuck Berger doesn’t feel slighted in the least, doesn’t feel like the last kid picked for a playground kickball game. To the contrary, he thinks the recent craze around analytics M&A has left ParAccel in a great position, and he’s probably correct. For starters, the spate of acquisitions isn’t necessarily over yet — there are still a handful of large vendors that could stand to step up their big data efforts — and even if it is, ParAccel is well suited to playing the role of the “scrappy” little guy pushing the innovation envelope and keeping large vendors honest.

On paper, ParAccel looks like a prime candidate to make headlines as the next big data acquisition, possibly with a premium price tag to match its now-exclusive status. Thus far, EMC, IBM, HP and Teradata have spent an estimated $2.5 billion on their big data acquisitions, largely involving ParAccel’s primary competitors. Berger says that ParAccel — which sells a column-based, shared-nothing analytic database with a focus on high-end applications — is almost always competing against Netezza, and occasionally against Vertica in lower-end deals. ParAccel wins its fair share of customers, too. Berger cites between 30 and 40 customers, including two very large global banks and many household names, including Nielsen, Moody’s and Merkle. ParAccel’s focus has been on financial services, intelligence and retail customers, although Berger noted that social media is fast becoming a focus among potential customers.

Presently, Berger said, ParAccel is operating as if it will go forward as an independent player, but he noted that he’s ran five other companies with that mindset, and they’ve all ended up getting bought. If a large vendor does come courting, ParAccel could benefit from the law of supply and demand. Responding to a question about whether ParAccel might be able to command a higher price to a potential buyer late getting into the game, Berger said he’d never go into a deal thinking the company could name its own price, but he’s aware that the supply of available analytic databases “just got awfully scarce” in the past few months, and it makes sense that a buyer determined to get into this space at this point will most likely have to pay a premium.

Berger’s confidence is probably well founded. Aside from the business it already has been able to secure in its relatively short time selling its software, ParAccel has one pending deal that illustrates the value in being small. Berger said ParAccel is presently trying to win a deal that its competitors might not be able to close now that they’ve been absorbed into large companies, because large companies generally aren’t able, or willing, to adjust their sales models on a deal-by-deal basis. There’s also the concern — true or just perceived — that companies lose their innovative edges, and certainly slow the pace of innovation, after getting acquired.

ParAccel, on the other hand, is adding new features all the long, including a forthcoming connector to Hadoop environments. CTO Barry Zane explained that ParAccel already has MapReduce functionality built into its database, so its goal isn’t necessarily a synergistic relationship between ParAccel and Hadoop environments but, rather, a migration of Hadoop data into ParAccel. Once there, the ParAccel Analytic Database can process unstructured data a la Hadoop via custom algorithms as well as built-in functions that are non-SQL-oriented but still use the SQL language. It’s actually quite a bold approach to Hadoop that ParAccel is taking; most its its already-acquired competitors have partnerships with Cloudera around hybrid SQL-Hadoop environments, but ParAccel wants to do it all. Zane will be discussing risk analysis in financial services at our fast-approaching Structure Big Data event in New York City.

Still, it seems as if acquisition is ParAccel’s ultimate fate, as it’s just too valuable an option for a vendor with an existing storage or server business to pass up. Not only is the software itself valuable for deriving insights from big data, but, as Berger explained, it’s designed to run on high-end systems with as much disk and memory, and with as many processors, as a vendor wants to pack into them. That means anyone buying ParAccel can drive up revenue across its server, storage and software product lines. Oracle plans to make billions per year from this model, and EMC has already rolled out a high-end appliance based on its newly acquired Greenplum Database. Teradata and Netezza (now an IBM company) both have strong analytic appliance businesses, too.

Who might come courting, though, is anybody’s guess. I got the impression from a recent conversation with a Dell VP that his company isn’t planning a move into the big data space, which, if true, could end up being a big mistake. But that’s Dell’s prerogative. Current ParAccel technology partner NetApp would make sense, though, especially after it just bought Engenio and gave itself a high-performance hardware business that might make a nice foundation for a ParAccel analytic appliance. But my crystal ball is a bit murky these days: I had Dell buying Aster Data Systems.

Image courtesy of Flickr user jmayer1129.

You’re subscribed! If you like, you can update your settings

  1. Thanks for the nice write-up. I’d like to clarify ParAccel’s position with Cloudera and Hadoop.

    For the highest performance, SQL-integrated analytics, ParAccel is the clear choice. However, Cloudera and Hadoop are great products in their own right and offer the benefits of open-source.

    Indeed, we certainly don’t see Cloudera and Hadoop going away! This is why we are currently developing a connector “bridge” to Hadoop for those projects where the customer has a Hadoop preference for some or all of their MapReduce code.

    Barry Zane, CTO, ParAccel

  2. It seems like ParAccel are trying desperately to position their smaller customer base, with smaller user counts as making them the last independent player, which to me makes it look like they are desperate for acquistion. If you discounted SAND and our 100+ customers, Infobright, Calpont, Exasol and many others then their position as the last independent could be true. Perhaps the last independent column database with 30-40 customers?

Comments have been disabled for this post