TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Analysis: Behind QlikTech's Acquisition of Expressor Software

QlikTech says its Expressor acquisition has everything to do with metadata and almost nothing to do with data integration. Industry watchers disagree.

• By Stephen Swoyer
• June 13, 2012

QlikTech Inc.'s acquisition of Expressor Software Corp. has almost everything to do with metadata, very little to do with data integration (DI) connectivity, company officials say

“[Expressor's] key differentiator is that it's a metadata-driven solution. It builds on top of prior success. You can read [from] your existing deployment, get the metadata out of it, and use that as a foundation for expanding,” says QlikTech CTO Anthony Deighton.

QlikTech's pitch with its flagship QlikView analytic software has deemphasized the importance of metadata. No more, says Deighton. “Part of the disdain that [we've had] for metadata is its traditionally top-down approach,” he explains. “The traditional metadata model is 'Come in, knock down the entire company, and try to build the uber-metadata model for everything that you're going to do.' Then consider what the business value is. That's kind of the counterstrategy of QlikView.”

Industry watchers aren't convinced, however. Certainly, they say, metadata management has been a problem area in QlikView, but it's unlikely that parent company QlikTech would allow Expressor's DI connectivity to wither on the proverbial vine. “A lot of people use QlikView over a [data] mart or other such thing where ETL plays a role. Metadata was a considerable weak spot [for QlikTech], just like [it is with] Microsoft and Tableau,” concedes Mark Madsen, a principal with information management consultancy Third Nature Inc. “I suspect [Expressor is] going to become the preferable alternative to the [existing] scripting interface to get data into QlikView when [it requires] more than mapping and less than big bludgeon ETL.”

Why is QlikTech downplaying the importance of Expressor's connectivity expertise?

What’s the Connection to Connectivity?

Expressor isn't a new DI player. It's been a going concern for more than four years now, and it isn't an obscure player. When it launched, it boasted a rockstar-like management team: chief scientist John Russell (now with Pitney-Bowes) was a Yahoo veteran, CEO Bob Potter was a veteran of agile data warehousing specialist Kalido, and senior marketing executive Michael Walclawiczek was a veteran of both Kalido and the former Streambase. (Both Potter and Walclawiczek are still with Expressor. According to QlikTech's Deighton, they're expected to remain in their roles.)

Nor is Expressor a stranger to the QlikTech world. Back in January of this year, the two companies signed a formal partnership accord. The companies have a history together, albeit a brief one.

In all fairness, it isn't surprising that Deighton and QlikTech would seek to play up Expressor's metadata management facility: a nonconformist approach to metadata has been a big part of Expressor's pitch from the very beginning. "What we provide is a semantic metadata repository. We built a set of tooling around the repository. It's a role-based, collaborative system, [the] aim [of which] is to build a central [collection of] metadata by rationalizing physical metadata from different source systems and maintaining that relationship in the semantic dictionary," Walclawiczek told BI This Week back in May of 2008. "Everything that we do in the system is based on those common, rationalized business terms."

That's what QlikTech says was the impetus for its acquisition.

Expressor emphasized its nonconformist take on metadata management, but it also trumpeted its leaner approach to enterprise DI. In fact, lean data integration was a major component of the version 3.0 reboot that Expressor announced 18 months ago. That release, said Potter, amounted to a substantive redesign of the Expressor platform itself.

“We decided we really had a good vision, but we didn't have quite the right product architecture to execute that vision," Potter told BI This Week in late 2010.

He described Expressor's “smart semantic” approach to data mapping as a boon to DI, chiefly because it eliminates the need for 1:1 mappings between data targets and sources. With some DI tools, Potter claimed, “every new project is a brand new project; there's no reusability. What we allow users to do is to build a canonical model or an abstracted model based on [data] types, which means that you map only once, then you reuse over and over and over again.”

Although there's much to recommend QlikTech's metadata-first spin on its acquisition of Expressor, there is reason for skepticism. First and foremost, QlikTech has a slightly longer-standing partnership with Informatica Corp. In conjunction with the release late last year of QlikView 11 -- a major refresh of the QlikView platform -- Informatica announced a PowerCenter connector for QlikView. QlikTech officials even made the QlikView Connector for PowerCenter part of their version 11.0 pitch. “Instead of having Informatica dump [data destined for QlikView] into a flat file or whatever, [IT] can have it directly write to what we call a QVX file, which is our optimized load format,” said Jeff Boehm, vice president of global product marketing with QlikView, at the time.

Informatica's QlikView Connector is a one-way-only proposition: it isn't designed to support bi-directional connectivity, and it only supports extraction into QlikView. It's less a general-purpose connector for DI and more a fit-to-purpose connector for a specific vendor.

In light of its history with Informatica, some industry watchers suggest, QlikTech is loath to say very much about Expressor DI as a potential competitive offering.

This is in spite of the fact that, according to Deighton himself, QlikTech plans to continue to sell Expressor DI -- to existing and to new customers -- in the long-term. “In terms of support, we'll absolutely continue to support the existing [Expressor] customers,” he says. “[T]he approach that we have in the short term -- and quite frankly in the long term -- is that ... the Expressor product will remain as an independent product ... and people can buy it, and existing customers will [find that they] have use cases that mash up [with QlikView].”

Deighton doesn't see the acquisition as conflicting with -- or as functioning to somehow deprecate -- QlikTech's existing partnership with Informatica, either.

“It doesn't change the partnership significantly. The focus for us as [the acquisition] relates to Expressor is metadata, not [data] integration,” he explains. “The value of Informatica is in its integration. In a funny way, Expressor helps us ... because it provides us with a way to consume the metadata ... that Informatica provides us.”

Is QlikTech really going to refrain from pitching Expressor DI to small and midsize enterprise shops -- particularly in cases (such as those suggested by Madsen) which require more than scripting but less than full-blown ETL? Deighton demurs, noting that many of QlikTech's customers also have large Informatica implementations. “We're interested in that [Expressor DI], but many [of our] customers -- in particular, our large customers -- have very large implementations with Informatica, and that's their preferred tool for integrating the data.”

An industry-watcher who spoke on condition of anonymity describes QlikView's partnership with Informatica as a “stopgap” measure. “There was a big cultural mismatch [between QlikTech and Informatica], with Informatica being the big corporate data integration tool and QlikView being the quick-to-deploy BI tool,” this person says. “Informatica is a lot of things, but agile isn't one of them. They provide infrastructure, which, once in place, is great. [Providing infrastructure] isn't what Qlikview is about. They needed something to better support the data ingest process than the basic data mapping they had.” Expressor gives them that, this industry watcher claims.

Third Nature's Madsen, for his part, says he's “surprised” by the acquisition.

“I was expecting either an internally developed product or a partnership,” he notes. “An acquisition makes sense, and Expressor was small, culturally compatible, and filled [a need] with metadata and integration. Plus at about 20-ish people, [it's] easy to digest.”