Richard Hackathorn: July 2009 Archives

Founded in 1993 from Sweden, QlikTech International developed QlikView, a point-and-click data analysis tool. Presenting are Anthony Deighton, SVP Products and Strategy (remotely via Webex), John Trigg, Global Product Manager (also remotely), Brian Wahl, Solution Engineer, and Tony Setter, Account Executive. They have been noted as a fastest-growing BI vendor with $120 M in revenue, 50% growth in 2008. They have 530 employees in 22 offices in 12 countries. Now, QlikTech has over a half million users in 11,500 customers in 95 countries.

Deighton started with the argument that traditional BI is dead due to commodity consumer software, technology savvy of users, and anger of business managers. He refers to this trend as the consumerism of enterprise software from easy-to-try, easy-to-buy, scales and UNsupported. In another slide, Deighton notes from surveys (2004-2006) that traditional BI has been a failure because of average implementation time of 17 months, over million $ in software, 31% success rate (however measured!), and only 36% users are confident in reports.

The market opportunity lies in simplifying analysis for both trained analysts and untrained masses, as shown here. It seems that many vendors make the same claim. I would like to see a deeper statement about learning process of untrained masses and the like.

QlikTech claims that their approach is a fundamental difference from other vendors with patented in-memory associative technology. As Curt Monash noted a year ago, the use of "associative" is unclear in this context. So do not think of associative in a technology sense, but in a psychological sense.

Wahl started demo of sales by customer, products, region, sales rep, etc. A few interesting points: can move/copy analysis objects from QlikView to Excel/PPT. smart cache of dynamic aggregation as use clicks with sharing among users,

Deighton remarked, "Advantage comes when you are solving a business problem and you can do it with speed. Scalability to many TB does have limited; however, those limits are increasing over time. Almost all users are server-based, rather than client-based."

Final comments about QlikView 9 launch... has instance running on Amazon EC2 so that a server can be deployed "within 15 minutes", has iPhone and Java Mobile clients, and a free developer tool for personal use.

My Take... A 16-year old company who has not only survived the tech wars but has flourish in recent years. Although many other vendors make the same claims in the same market segment (pervasive enterprise analysis tool), QlikTech has carved a niche for itself within the international arena with a proven mature product. I like the 'consumer' flavor to this enterprise tool as a mechanism for maturing the user base. An issue is whether the depth is sufficient for the power users. I would recommend BI professionals invest the time to download the free version and apply it to a nontrivial problem. 

Birst is a Software-as-a-Service (SaaS) company focusing on BI tools and solutions. Presenting are Brad Peters, CEO and Stefan Schmitz, Sr. Director of Product Management. Product launched in September last year and has some large customers, such as RBC Wealth Management, Spectrum Consulting, and Securian. Birst currently has more than 50 customers most of whom do not want their name released. A typical customer has 50 users with approximately 500 gigabytes of data. However, the largest customer has around five thousand users, terabytes of data, dozens of data sources, and displaced Oracle Financials with same functionality at a much lower cost. Birst is staffed with experienced BI veterans and funded by Sequoia Capital and Nummer-Winblad Venture Partners. "We are well funded," said Peters with hopes of being profitable next year. 

Peters described their target market at the gap between end-user ad-hoc reporting and departmental analytics, as shown with the dotted oral in the figure.

Birst emphasized the economics of deployment of the application, not necessarily the usability. "Birst has rationalized the infrastructure and re-engineering the IT supply chain, like Home Depot and Walmart has done in their markets," said Peters. @NeilRaden asked how does Birst do the integration of all that messy user data? "You can not turn straw into gold," said Peters. But, Birst does do a reasonable metadata analysis and generates the likely multi-dimensional model, which can then be fine tuned.

Birst differentiates themselves from LucidEra and other BI SaaS vendors by focusing on data integration capabilities. @NeilRaden quipped, "Data integration is like making sausage. You really do not want to know much about the process!"

Infrastructure for Birst has been built internally. They did not use open-source software since "it is not that good," remarked Peters. Birst uses both row and column oriented relational engines.

Demo time! Schmitz took us on a end-to-end solution scenario. Built a template with 'automatic' data upload, creating a dimensional metadata model. This model will then create keys, etc. The automatic mode is for simple users with simple data, but an advanced mode can be used by a data modeler to handle complex data. Point to your data files, or connect to a relational database, or connect to popular applications like Google Analytics and Salesforce. In 3 minutes, it produced a clean model and set of dashboards of the Northwind database from a zipped copy of an Access database. There is a nice way of displaying the dimensional model, allowing for fine tuning to that auto model  ...and nice set of pre-built analytics on all combination of attributes, measures and time-series ...and fairly comprehensive set of options for report/chart formatting. Schmitz took us on a tour of various reports/charts flipping through the attributes/measures (facts/dimensions). Sharing of reports/charts is done via folders with customized permissions.

Discussion of future features under NDA...

My Take...  Birst does have a value proposition for business users suffering under a weak corporate IT infrastructure...especially based on quick time for deployment, low skill requirements, and low pay-as-you-go cost structure. This implies to companies of any size. Peters remarked, "Smaller companies have difficult, even complex analytical problems." As wasvtweeted by @NeilRaden, who continued, "The less you know about IT, the more difficult questions you will ask!"

The long-term strategy would be enhanced if there was a better value proposition for the enterprise IT folks. At the moment, there can be a tense relationship. There are several ways to make enterprise IT happy, such as: data integration-as-a-service, dimensional-modeling-as-a-service, report-generation-as-a-service, chart-publishing-as-a-service ... all integrated back into the enterprise data warehouse. I would suggest that we should think of Birst as the front-line troops that capture new territory (unsupported BI applications) that will eventually be blended back into your IT empire. Now that is exciting!