I am deeply grateful for the opportunity that I was given to attend Tableau’s annual conference, catch up with peers and see what was cooking over there. Each year I highly anticipate the keynote announcements much like a child anxiously waits to open wrapped holiday gifts. It was the 2012 Tableau Conference “Kraken” keynote where I knew that I had seen the next big thing in business intelligence. Right place, right time…and right solution with all the right features customers wanted being developed. The Tableau developers on stage were truly endearing, real software engineers. How could I not appreciate what Tableau had built and what they were doing for my beloved world of analytics?
Fast-forward to 2014 and my expectations now are much higher for the market leading data discovery vendor…are they possibly too high? I enjoyed the 2014 conference but I am not enamored by Tableau’s pursuit of basic charting for the non-analytic masses on mobile devices revealed with Project Elastic. I treasure Tableau because it has BOTH ease of use and analytic depth. If it were a basic analytics solution, I would not care much for it and could substitute it with any of a zillion other alternatives that seem to pop up weekly. Many people don’t realize that Tableau appeals to a wide audience of non-technical and technical data enthusiasts. I admit to being a Tableau user “outlier” and that my desired wish list of geeky analytic features won’t get many votes. Even though I left Tableau Conference a bit disappointed, I’d guess that most Tableau users are happy. Here is a summary and my perspective on what I saw and heard last week that I can share publicly.
In 2014, Tableau (aka DATA) is the most talked about visual analytics company in the world. Last week their annual conference was held in Seattle with a record-breaking, impressive crowd of over 5500+ in attendance despite the addition of numerous regional conferences around the globe. A litany of analytics ecosystem partners also participated including top big data players (Cloudera, Hortonworks, Amazon, IBM, MapR, Splunk, Google), a variety of ETL offerings (Alteryx, Informatica, Paxata, Trifacta, Clover ETL, Lavastorm), predictive solutions (Big ML, Rapid Insight), database vendors (HP Vertica, Teradata, Actian, Exasol), implementers, and other add-on offerings. The business intelligence and analytics industry analyst community was also there to cover what DATA had to say.
Without a doubt, this event has grown rapidly over the past few years much like the Tableau user base. Historically Tableau’s land-and-expand momentum has been exceptionally impressive with sales growth of over 80% to 100% each year – literally doubling sales year over year over year over year over year. My #1 question coming into the conference was can Tableau realistically continue growing at such extreme rates? Historically BI and analytics solutions are deemed a success if they achieve ~30% to 40% adoption. I have seen Tableau saturating the U.S. market. I am also noticing several compelling newer players (ZoomData, SiSense, Looker, Datameer, Platfora, DataWatch) this year and a few of the historical BI players are indeed improving (Microsoft, SAP, TIBCO Spotfire, Qlik, SAS). In my opinion, I feel the biggest threat to Tableau is other solutions becoming “good enough”.
My #2 question was around what improvements were being made in the areas of enterprise implementation. Tableau often spreads quickly within an organization without much planning or governance. Tableau users totally love it. The 2014 conference theme, “Art of Analytics”, was appropriate! Tableau is a great solution and blows away most of the competition in this space at the moment. However, Tableau users can easily go crazy with it, creating a mess after the first year or two. There is a bit of work entailed to sort through, reorganize and optimize Tableau messes to keep it scalable and manageable. Messy Tableau implementations actually look a lot like rogue database under the desk, Microsoft Access or VBA with Excel reporting projects with a faster Tableau Data Engine (TDE). No one seems to care about how or what they can’t see behind Tableau’s beautiful dashboards but the implementation sure makes a world of difference in enterprise scale out scenarios. The Tableau partner ecosystem has been trying to ease pains and add toolkits to fill enterprise voids. Although these issues can be mitigated with the right people, processes and tools, I was hoping to see version control, lightweight workflow and stewardship, metadata management, impact analysis, lineage, and improved performance news. Here is what I did see, what totally shocked me, what I felt was missing and a few thoughts on Tableau’s publicly revealed direction.
Shocking “Project Elastic”
The opening keynote was jam-packed with vNext announcements. If you missed the opening keynote, you can still watch it on the conference web site. This year the biggest surprise by far was a new mobile application targeted towards general user audiences, everyday consumers, being developed under the code name Project Elastic. At first I was shocked but after thinking about it for a few days now, this strategic move makes a lot of sense. Data analysts are a smaller, limited market in comparison to an Excel spreadsheet, Google Docs or Apple Numbers user market. Today those spreadsheet users do consume data on mobile devices more so than ever. Spreadsheet apps on tablets and smart phones leave a lot to be desired due to super limited interaction and poor ad-hoc visual exploration capability. Mobile BI players like Microstrategy, RoamBI and DataZen are nice but those offerings lack strong mobile visual exploration and authoring – Tableau’s core strength areas. It is rarely known that Spotfire’s base cloud mobile authoring is actually pretty good. I have not tried Qlik Sense on a mobile device or tablet yet but they have shown touch-friendly demos in their natural analytics campaigns. By adding a new mobile BI product into the mix, Tableau is both diversifying their portfolio and expanding market potential.
- New Analytics tab/pane for easier drag/drop forecasting and statistics (The existing features shown were basic but it has fantastic potential.)
- Totally innovative data cleansing from a view (I have not seen anyone approach data prep or cleansing from a visual. This will be fascinating to see evolve.)
- Limited data prep features, pivoting and splitting rows in the data load window (Power Query in Excel has FAR better features in this area. Other spreadsheet-ish ETL solutions like Informatica Springbok and Paxata have similar capability.)
- Added 9/19/2014: New “smart” import capability of unformatted spreadsheet data that will intelligently remove headers and find the data to load into a table format (Super great feature and really innovative! No one else has done this that I am aware of right now. It eliminates the “yuck” work in spreadsheet data prep.)
- New, cool map lasso selection (similar to Qlik and Spotfire lasso features), radial selection and geographic search on maps
- Better Tableau Server administrative views that visually show permissions and data sources used in a published Tableau workbook (Sigh, baby-steps…we need more for enterprise roll-outs.)
- Parallel query performance improvements versus serial query, query engine scale across CPU processors and also improved persistent caching capabilities (Tableau continues to grow up with these needed enhancements, OAuth, central credentials and also the added Kerberos support.)
- Improved calculation interface allows editing in-place versus pop-up windows and an interesting drag/drop experience to drag a calculation to a model (Nice, user-friendly)
- New table calculation scope highlighting (Yay!)
- Finite control of the blah-grey storytelling boxes (You can now make them prettier, smaller and have the story text fit appropriately. Thanks Tableau for listening to my cries about that feature.)
- “DRIVE” a recently provided process for Tableau enterprise deployment (Tableau’s response to enterprise needs.)
- Completely rewritten mobile BI app including a new offline image feature with refresh, added capability to edit calculations in mobile BI app versus the previous read-only web authoring on mobile and favorites improvements
- “Project Elastic”, described earlier in this article (a WOW announcement)
- Cloud news ??? I missed it since I had a hard time paying attention. I was stuck on what does the new Project Elastic mean, how will it play with existing Tableau mobile BI, and so on. Added 9/19/2014: Tableau Cloud has live RedShift connection capability. There is also a new Tableau Cloud feature called Data Sync in the roadmap that allows for scheduling of data transfers to Tableau’s cloud via installed agents. (Sounds similar to Microsoft’s Data Management Gateway in Office 365 Power BI).
- Added 9/19/2014: Improved Salesforce report embedding via the Canvas framework instead of the current inline frame approach allows for an improved combined solution experience with much better reporting than Salesforce offers
When reviewing and monitoring vendors, I also pay close attention to what is NOT said and the actions not taken. Last year I gave Tableau my wish list of features. I provide wish lists and feedback to a few analytics ecosystem vendors now that I am out on my own. My #1 feature request that I felt would deliver the biggest positive impact was proactive, contextual alerting. I wanted the type of alerting that Targit BI and Microsoft Reporting Services provide today for my accounts. I feel that modern dashboards should be “smart” and forward-looking in addition to letting you know what happened in the past. I want smart tools that reach out to me when there is a possible issue or item of interest – manage by exception but better. Sure I can build predictive queries and code alerts myself but I wanted an easy, breezy and powerful right-menu option. I guess that leaves room for Tableau predictive engagement work for the time being. Tableau projects are generally tiny in scope so I need something to keep me working or I’d starve to death out here.
Other missing items that I hoped to see include:
- Dynamic parameters? (This one is a need – not just a wish.)
- Better implementation of tree maps
- Address level geocoding for maps like Excel Power View mapping
- Less complicated conditional formatting (usually requires calculation logic code)
- Easier mass subscriptions (currently requires tabcmd code) versus onesie subscriptions
- No infographics (actually good since infographics really is not best practice data visualization)
- No operational reporting (still need to supplement with Reporting Services, Jasper or Crystal)
- No data profiling or quick descriptive statistical skims over loaded data sets to see qualities like correlations between columns or distributions
- No new predictive algorithms, prescriptive or advanced analytics that I noticed mentioned
- More visualizations? What vizzes were new? Did I miss them?
- Etc., etc., etc. (I will spare you my ridiculous wish list of nice-to-haves.)
If a few of the above missing items were covered, please let me know and I will add the correction.
Take-Aways from Tableau Conference 2014
While I am not thrilled about Tableau showcasing basic charting with Project Elastic, I do understand why they are moving in that direction. The deeper analytics market is dominated by powerhouse SAS, SPSS, R and TIBCO Spotfire today. The ROI potential of my geeky feature wish list would be far less than the ROI of serving an untapped, basic reporting, mobile device user audience. I do hope that Tableau’s current Desktop and Server solutions continue to evolve and remain deep analytically. It would be a shame to see a beautiful analytic offering becoming a basic one, eventually fading away into the sea of other rudimentary tools. Added 9/19/2014: Tableau representatives confirmed that they are committed to building solutions that help people get smarter with data. It is not their goal to become a “basic” solution. Project Elastic is for different audiences and use cases than Tableau’s Desktop and Server. Tableau’s Desktop and Server will continue to evolve in the areas of analytics that I was distressed about not seeing in the 2014 keynote. Great and important clarification – thank you for sharing that insight!)
So with regards to my #1 question coming into the conference, I left feeling that Tableau could indeed continue to grow exponentially overseas in 2015. The move of Tableau’s CEO to London is noteworthy. Right now there is a wonderful opportunity to grow Tableau overseas. In Italy for example, no one in my business intelligence presentations was using Tableau! It could have been a language localization issue, lack of awareness, strong competition from SAS or Microstrategy overseas, or for other unknown reasons. I don’t think extreme 2x growth will continue in the U.S. market. There are more and more new players pounding down the door of the poor U.S. Excel user with the same exact message and similar looking offerings.
As for my #2 question, sigh… Apparently it looks like Tableau is leaving it to the partner and implementer community to work through recent DRIVE guidance and add value where there are voids for the large enterprise deployments. This is another opportunity area for Tableau consultants. Added 9/19/2014: Tableau representatives confirmed that DRIVE guidance will be made available to customers, partners, and also as a services offering through their Professional Services Group. They will be continually improving and adding enhancements in this area of their offering.
All in all, there are a lot of pleasant little Tableau improvements on the way. Tableau has a great summary of those here that you can review. They always have fabulous speakers, content and sessions. The Tableau community is extraordinary and a lot of fun to be around. So am I too harsh? Did I expect waaaaaay too much from this market leading data discovery vendor? I do know that I can be over the top at times with things that I am passionate about! This might be one of those times.