Tableau may not have the largest market share in the BI space, and with 2012 revenues just over $127 million compared to billions for mega BI vendors, some might consider this BI start up more of an annoyance than a worthy contender. But in just a few short years, Tableau has gone from a niche visualization vendor to the preferred BI tool for companies just starting out on their BI journey. For mega vendors, Tableau has become the gold standard for ease of use and visual appeal, with most major BI vendors-- from SAP to IBM to MicroStrategy--all releasing products in the last year to stave off Tableau's encroachment into their accounts.
The company went public in May 2013 and its IPO price of $31 quickly jumped to over $50 with a peak of $74.70. Carving a distinct path, the company also decided to file not with the tech centric NASDAQ, but rather, with the home for many Fortune 500 companies, the NYSE. In the first half of 2013, the vendor's revenues have grown 67% to $89.9 million while many BI platform vendors are in single digits. I'm sure a number of software companies are wishing they had acquired Tableau in its earlier days, but even if the offers were on the table, the founders have repeatedly said they were not looking to be acquired. They've developed a product and fostered a culture that is hard to replicate.
its conference last week in Washington, D.C., felt more like a love fest than a user conference. Sure, there were the keynotes by the CEO and data rock star Nate Silver (of New York Times and now ESPN fame), but this company dares to put its developers on stage to demo the latest release. The crowd, with 4000 attendees, went wild, and the audience wanted these developers to succeed. The company even flies most of it employees to the conference, so they can meet with customers, see firsthand how the product is being used, and consult in one-on one "Tableau Doctor" sessions.
Tableau outlined six themes to its next release, with version 8.1 due in Q4 2013 and 8.2 in Q1 2014:
- Seamless access to data, in which bringing in data from even somewhat messy data sources will be easier
- Visual analytics for everyone, with R integration, better ranking function, and improved forecasting
- Fast, easy, and beautiful, with a lot of little user interface improvements such as a calendar control for date filters and enhanced presentation mode
- Storytelling with data in which dashboards are organized into story points that provided a more guided analysis experience, stepping a user through a decision-making or discovery process.
- Enabling the enterprise, in which the product will be fully 64-bit for greater user and data scalability. The cloud version will add support for SAML authentication.
- Visual analytics everywhere, with desktop authoring support on the Macintosh (8.2). Very few BI vendors support the Mac for authoring, but to be fair, most BI platforms use the Web, not the desktop, for authoring.
There was a long list of features and new capabilities showcased. The demonstration of story points was the most compelling in my opinion. Interestingly, "story telling" is a key theme of QlikView.Next , announced last year and currently in beta. At the ASUG BusinessObjects user conference in Anaheim last week, SAP also showed the concept of story boards in its upcoming Lumira SP12 due in a few weeks. So far, the Tableau demonstration seemed the more robust.
I thought it was interesting that the usually hyped topics of big data, cloud, and mobile were barely mentioned in the keynote. Perhaps it's because Tableau is already well into its journey in those areas, or perhaps it was because the audience was mainly business users, rather than an IT audience. Instead, the message seemed to be more about inspiring people to get to the next level of analytical thinking and creative break through.
Three areas that are not improved in the next release and that will continue to challenge Tableau are:
- Lack of support for multipass SQL: when users want to compare data from different fact tables or different levels of granularity, the DBA has to create a view or cube to enable this. Tableau supports data blending when data comes from different data sources, but not multiple queries from the same data source. As one customer pointed out, even if data can be blended in a work book, that's not the most scalable place to do so.
- Worksheet chaos: The data model in Tableau is embedded within a workbook that also contains the individual reports and dashboards, referred to as worksheets. An author can share the data model via the Tableau Data Server, but as users create more visualizations and reports, all those individual worksheets get added into one workbook, with no way to find or search through the worksheets and no way to logically organize them. At some point, Tableau has to better separate the data model and presentation components.
- Either/or in-memory model: Early on Tableau made the wise decision to leverage the power of existing data warehouses and analytic appliances. They only later added their in-memory engine as an optional performance improvement. Meanwhile, chief competitors QlikTech and Tibco Spotfire initially only had in-memory. A certain segment of customers didn't like that data had to be replicated into those vendors' in-memory solutions. Last year, both QlikTech and Tibco Spotfire introduced the concept of hybrid in-memory in which in-memory data can be combined with relational data that is selectively called from the data warehouse into in-memory. IBM Cognos with its dynamic cubes released last fall follows a similar concept. While the implementation, adoption, and marketing of this new hybrid in-memory approach vary by vendor, the hybrid in-memory seems to give customers the best of both worlds: fast and flexible. Scalable and performant. As in-memory has become more maintstream, I suspect hybrid in-memory will be the next generation approach. For now, Tableau is in the either/or camp: the data is either stored in its own in-memory data extract or stored in the underlying data warehouse.
There are other aspects to the Tableau product line that doesn't compete with large BI platform vendors (see our summary and detailed scorecards for side-by-side comparisons), and to this end, CEO Chabot was steadfast in the company's focus to help the everyday business user. The vendor will rely on partners to fill the gaps that Tableau does not intend to fill, as shown by recent announcements from Informatica and Teradata.
If you are a regular reader of my column, you might have noticed that I was a tad quiet the last few months. It's not because the BI market has been quiet, but rather, I was heads down finishing my next book "Successful Business Intelligence: Unlock the Value of BI and Big Data." I've just checked the final proofs and you should find it in stores or ebook in early October. Based on Tableau's keynote, I'm wondering though if it's not too late to drop BI from the title… perhaps I should have just called it Big Impact?