Tableau has done for the discovery of data what Netscape did for the discovery of information, with the first web browser – empowered the masses. For data discovery, Tableau makes it simple to connect to some data, slice and dice it, and create some cool visualizations. It more than satisfies a simple equation for a software product:
Love = Results – Effort
That is, if the results for your users are way larger than the effort they put in, you have a winning solution: and Tableau kills it. Tableau’s timing was perfect, end user empowerment, the proliferation of data, just at the same time traditional command and control analytics was reaching a user frustration tipping point. Tableau provides an incredibly level of interactivity to “play” with the data, without requiring IT.
And there is one other timing aspect that Tableau has continued to capitalize on: a sustained vacuum of analytics vision from Microsoft, because they'd been asleep at the wheel around analytics. For a long time, Pivot Tables and Microsoft Analysis Services were the last great analytics innovations from Microsoft, and those introductions disrupted vendors (I worked at a vendor on the receiving end, and it sucked). But after those introductions, it has been a nuclear winter. That absence enabled Tableau to spawn a new industry – empowering users to explore data, and to thrive.
The Browser Wars of the Mid 90s
Similarly, when Netscape first appeared, with the growth of the Internet, Microsoft was essentially asleep at the wheel too. At the peak, Netscape had an 80%+ share of the browser market. Fearful that Microsoft was late to the Internet, Bill Gates led the led the call to arms with a letter to focus on the tidal wave. One of the areas: Netscape. The strategy was to put their full weight on changing Netscape’s dominance, with (love it or hate it) - Internet Explorer. Netscape quickly lost share as IE simply became the default - dropping to less than 1% share by 2006.
Gate’s Internet is Nadella’s Cloud and Data. One of the cornerstones of Microsoft’s strategy is not just cloud, with Azure (which now is second only to AWS) – empowering developers to create cloud services, but also tools and services to empower users to work with data.
The announcements around analytics have come quick and fast, PowerBI; PowerBI Desktop; PowerBI Mobile; PowerQuery; Azure Stream Analytics; Azure HDInsight; Azure Machine Learning; and Cortana Analytics. For the PowerBI suite, the price is right - PowerBI is free, and PowerBI Pro is $9.99 per user per month – where you get more data, more refreshes, on premise connectivity, and more collaboration features.
The Coming Data Discovery War
So I tried out the web flavor of PowerBI a few months ago, bringing in some data from Salesforce into a prepackaged web dashboard, and it was cool, but to be honest the results were too limited – you couldn’t really play with the data enough. Definitely a threat to some cloud dashboard providers, but no threat to Tableau for real empowered data discovery. It’s more for consumption of analytics, but not playing with data. It fits into a data discovery framework, but isn’t the whole solution.
Fast forward to last week, where I tried out PowerBI Desktop. PowerBI Desktop is basically the equivalent of Tableau Desktop. And the interplay is similar, where users create rich analytics with the client, and then publish to the web to share the results.
But what blew me away was how PowerBI Desktop stacks up....
Let’s start with the data sources. They’ve done a great job of adding a huge number of sources – the usual suspects like Excel, text files and database sources, but also supporting a wide range of big data sources, social sources, ERP and CRM sources etc. It looks like they’re working with ISVs to add sources at a frightening rate. Getting access to data is often one of the big stumbling blocks for data discovery (and I think one of Tableau’s weaker areas) – and it looks like Microsoft is really focused on cracking the code here.
So then I thought I’d get my hands dirty and give it a little test drive with my favorite old time schema – Northwind (which I was pleased to see Microsoft still use for on-stage demos!). It’s a relational schema, and PowerBI Desktop did the automapping for me, then enabled me to easily make some changes to the joins. Nice and straightforward and very usable, and easy to visualize the relationships.
Finally, for the really fun bit, some data discovery. And this is where it was shockingly good. From soup to nuts, from data to dashboards, I built the quick example below in about 20 minutes. And it checks all the boxes. On the right is an easily field selector, there’s a rich array of visualizations – traditional charts, heatmaps, gauges, geospatial charts (more visualizations can be added by third parties) etc. All of the visualizations have strong data flexibility, so I could easily change the data that I’m seeing in then chart, filter it, use TopN/BottomN etc. I found myself easily slicing around the data, trying out different views, just like Tableau.
Some of the cooler stuff is how the dashboard components automatically snap together, with no effort at all, so for example, when I click on a region on the map, my other charts automatically orient, and it’s easy to create a book of dashboards, calculated measures etc.
Oh, and publishing is simple too.
So, is Tableau the New Netscape?
Which brings me back to the comparison at the start of all this. PowerBI Desktop does what 90% of people need to do with discovery tools, and it’s free, and nicely integrated with Office. So why use Tableau then? Sure, Tableau is still better in some areas for sure – more visualizations, it chooses the right chart automatically, Mac support, and I’d say it still has a slight edge in intuitiveness for data discovery. But here’s the kicker, Tableau is 10+ years old, PowerBI is 1.0 – and it’s tying into Microsoft’s broader strategy around Azure, Office365, and Cortana. Brutal.
I’m sure there’s chatter going on in the halls of Tableau on PowerBI. But to be sure, the threat from PowerBI perhaps means considering additional options around predictive analytics, or moving towards an applications strategy beyond tools.
Of course, if I were to take the Netscape analogy to its ultimate ending, out of the ashes of Netscape rose Firefox – which came to haunt Microsoft. I’m not sure this story will end in the same way.