Ever wished you could change the dimension in a chart on a dashboard? Ever wanted to give viewers the ability to slice the data by the dimension of their choice? Well, Dynamic Dimensions may the solution for you.
Power BI Trumps Tableau
… Or Does It?
By Dwight Taylor
VIZYUL LLC Founder and CEO
As a kid, we used to play a game on the playground called King of the Hill. We played it in the winter when the playground was plowed leaving large snowbanks at its perimeter. Someone would race to the top of the snowbank and claim to be king, willing to take on all who dared to take him down. In our bulky winter coats and gloves, we would attempt to push, pull, drag and tackle the king off the summit so we could lay claim to the throne.
The game may no longer be a mainstay of the playground, but it’s played frequently in corporate America where one company dominates a market until another comes along and attempts to take it down.
Tableau is finding itself in that position. The data visualization platform has been the industry’s gold standard for nearly two decades. For the sixth straight year, Tableau is classified as a leader in the Gartner Magic Quadrant for Analytics and Business Intelligence Platforms.
Similarly, Gartner recognizes Tableau’s closest competitor — Microsoft’s Power BI — in the leadership quadrant (there are only three with Qlik being the third). According to the most recent Magic Quadrant study, Tableau and Power BI are equally on par in their ability to execute. However, in 2016, Power BI leaped past Tableau in the completeness of its vision.
I am a firm believer that Tableau is a much better platform than Power BI, but I also believe Tableau has to keep a close eye on its competitors or risk having Power BI become King of the Hill.
The vast majority of data analysts I engage firmly believe that Tableau is the better platform. It’s not even a contest in their eyes. I find it almost surreal that nobody believes Power BI is a threat to Tableau. I liken it to the 2016 presidential election where all the pundits and pollsters were handing the White House keys to Hillary Clinton. As we now know, the experts were wrong.
How did we get here?
The Rise of Tableau
In 2003, a small unknown company from the San Francisco Bay Area came into the data visualization market with a tool that could help the most non-technical person see and understand data with ease. At the time Tableau debuted, the data market was owned by goliaths like IBM, Microsoft, Oracle, SAP and others.
Tableau’s objective was to disrupt Business Intelligence incumbents with technology that made the democratization of data and self-service analytics possible without dependence on IT.
Tableau had a “land-and-expand” sales strategy. It targeted selling its data analysis visualization tool to the person at a company who had the biggest pain point. It often was a data analyst who had to synthesize files and data from a myriad of teams and divisions – an exhausting task.
From that single initial sale, Tableau expanded its footprint inside each customer’s company, and it didn’t take long for it to become a disruptor to the Goliaths.
Power BI’s Ascent to Prominence
Microsoft took a different approach. First, Microsoft targeted data analysts who were most comfortable using its Excel spreadsheets. Second, Microsoft incorporated smart analytics into its strategy. In the early 2000s, Excel was the No. 1 data visualization, data discovery and data analysis software program in the world so it was a no brainer for Microsoft to target those users. Between 2006-2012, Microsoft sold add-on software that allowed analysts to give a different look and feel when presenting data visualization.
Simultaneously, Microsoft was pouring millions of dollars into its own standalone product and unleashed Power BI in 2013 to go head-to-head with Tableau. The Power BI black box offering included Microsoft Excel add-ins, SQL Server Report Builder components, Machine Learning and Artificial Intelligence. The first release of the tools fell far short of competing with Tableau, and thus was ignored by the King of the Hill.
From inception Power BI was presented the perception that it was the more cost-effective data analysis and visualization tool. That perception has caught on, and the majority of customers I’ve spoken with believe Power BI is cheaper than Tableau. Microsoft has done a masterful job masking the overall cost of its Power BI platform, especially for on-premise deployments. Perception can trump reality.
Microsoft Power BI: Agility is Strength
Power BI is a de facto option for companies with Microsoft platforms like Office 365, Azure, Microsoft Excel and that have Enterprise SQL Server agreements for reporting. According to Gartner’s Magic Quadrant respondents, the platform is easy to use and has a free option, as well as an attractive entry point for the Power BI Pro version. However, this version requires that data be stored on Microsoft Power BI servers in the cloud. Many large companies and enterprises are reluctant to store data in the cloud, so the only viable option for prospective Power BI customers is the more expensive one.
“Microsoft\’s customer reference scores place it in the top quartile for ease of use, with 14% of customers citing this as the main buying criterion,” according to the February Magic Quadrant comparison. The report states that “BI\’s reference scores also place it in the top quartile for visual appeal. Winning customers within the ﬁrst few minutes has been part of Microsoft\’s ‘ﬁve by ﬁve’ strategy — ﬁve seconds to sign up and ﬁve minutes to ‘wow’ the customer.”
And it does all of that and more.
Power BI’s automated analytics, charting, custom visual platform and pre-built dashboard templates for products such as Google Analytics and Salesforce give it an interesting market position when placed next to Tableau, the current King of the Hill. However, Power BI’s highly technical nature of the data models is a barrier to entry for data analysts unfamiliar with complex data modeling concepts.
Tableau: Ease of Use and Self-Service Analytics is Great, But . . .
Tableau also comes in three versions – free, low-cost cloud-based and enterprise premium, allowing users to see and understand their data. Tableau specializes in the art of muting the complexity of the software and intelligently guiding the user’s attention to what’s most important, the data visualization, Tableau is ideal for the non-technical data analysts that need to create interactive data visualizations. The free version is a fully functional version of Tableau Desktop with data volume and data connectivity limitations.
With nearly a quarter century of business intelligence and programming experience, what I have witnessed and hear is that data analysts love Tableau.
Tableau earns impressive customer reference scores, with 94 percent of them scoring the platform as “excellent,” according to the Magic Quadrant report, and 55 percent “using it to empower centralized teams to provision content for consumers in an agile and iterative manner.” Additionally, 64 percent use it to enable decentralized analysis by business users, according to Gartner.
Tableau really invested heavily in integrations and connections to a wide variety of data sources. Its strong ability to obfuscate the complexity of data modeling and data integration remains one of its top selling points.
With its solid customer scores and a continued focus on reducing the user’s time to insight, no question that Tableau is currently King of the Hill, but Power BI’s “five by five” strategy has this offering making its way to the pinnacle of the hill.
The Power BI Cost Perception
When Tableau brought on Adam Selipsky as its CEO in 2016, the former Amazon web services executive changed the pricing model to a subscription-based model. The pricing conversion reduced the entry price point, which made Tableau a plausible option for companies where high initial costs were a barrier.
Being the motivated consultant that I am, I use and help clients achieve their objectives with both platforms. They are equally impressive and deserve to be the top two leaders in the Gartner Magic Quadrant for Analytics and Business Intelligence. No other platform is even close.
Now let’s look at the perception surrounding cost.
In the following graphic, I’ve presented a side-by-side on-premise deployment cost comparison of the two platforms for 50 users (10 content creators and 40 content consumers). Please make note of the assumptions following the graphic.
1 Labor costs not included (consulting, solution architects, etc…)
2 On-premise server deployment is a customer requirement
3 For high availability and redundancy, the customer has requested a 3-server cluster
4 No vendor discounts have been applied
5 Subscription pricing (not enterprise agreements) is preferred by the customer
6 Customer requirements do not include embedding needs
7 Server hardware and operating system licensing costs not included
When you crunch the numbers, Power BI is clearly the more expensive option for an on-premise deployment. Not to mention the fact that the cost grid does not include the cost of the SQL Server cluster necessary for an on-premise highly available Power BI Server deployment. I encourage prospective data visualization customers to also take a look at the TrustMap on TrustRadius; written by platform end users.
Tableau Making Moves to Stay Ahead
Smart Analytics is the incorporation of Machine Learning and Artificial Intelligence into platforms to automate data prep, data analysis, data visualization and data driven decisions.
As King of the Hill, Tableau has established itself as the self-service analytics disruptor to every incumbent Business Intelligence vendor. However, Smart Analytics is the disruptive force on the horizon. The vendor that successfully delivers Smart Analytics will ultimately emerge as King of the Hill.
On the playground, retaining King of the Hill status demanded vigilance and a healthy dose of speculation. Vigilance because one slip could end a reign. Speculation because identifying formidable contenders from mere wannabees doesn’t always yield expected results.
Tableau finds itself at this exact point of decision.
As a consultant on the ground with a wide variety of clients, I believe Power BI is that formidable contender. With a clear edge in ML & AI analytic capabilities, Power BI clearly has its sights on the throne; and Tableau is playing catch-up.
Tableau needs to do three things well.
- Flawlessly and quickly integrate ML & AI analytic capabilities into its platform with the August 2017 acquisition of ClearGraph and the August 2018 acquisition of MIT startup Empirical Systems. Also keep a close eye on ThoughtSpot.
- Arm every Sales resource with an ironclad total cost of ownership narrative making it crystal clear that Tableau ultimately is cheaper than Power BI.
- Don’t slip, trip or falter while executing on the existing Tableau product roadmap.
Power BI is a good example of how marketing impacts perception, and how that perception can create a window of time long enough to reshape reality. I believe Tableau clearly trumps Power BI when it comes to those telling stories with data
For now, Tableau continues to reign as King of the Hill; but for how long?
Want to share your thoughts, great. Share them below. Please keep comments clean.
FREE 45 min. Consultation
The thrill of the new Tableau server honeymoon is over…and now the hard work of Tableau Content Management begins. The first article in this series will answer the question \”why should I care about Tableau Server Content Management\”; and will demystify the various types of content stored in Tableau server.
Why should I care about Tableau Content Management…isn\’t Tableau a self-service platform?
It\’s a valid question. The answer may prove to be a bit shocking at first…SELF-SERVICE isn\’t really self-service. Yes, Tableau server is designed from the ground up to encourage and nurture a thriving self-service culture. However, over time, this self-service freedom comes at a cost. Here are just a few thoughts to get your Tableau Content Management juices flowing.
- Multiple copies of the same or very similar data sources published to the Tableau server
- Performance degradation
- Multiple versions of the same analytic content
- Back-end source system performance degradation (overwhelmed SQL Servers, Teradata, Oracle, etc…)
- Outdated/stale content
Finally, if your Tableau environment experiences any of the scenarios below, some focused attention on Tableau Content Management is a great place to start.
- Tableau server users have trouble identifying which data source is the most current.
- Tableau server users have trouble finding the dashboard they want (multiple copies of the same content or outdated content).
- Tableau servers overall performance keeps getting worse.
- Database administrators claim Tableau server is negatively impacting the performance of the databases they manage.
- Users cannot find the Workbook or Data source that contains the data field or formula they\’re looking for.
Tableau Server Content Types
Let\’s begin by identifying the various types of content found in your Tableau server.
- Sites – A Site in Tableau server is an actual content boundary. Cross-site content communication is not possible.
- Projects – A Tableau Project is an organizational object in Tableau server, not an actual object.
- Workbooks – A Workbook is a package for everything needed to display dashboards.
- Views – A View is a dashboard or worksheet published to the Tableau server as part of publishing a workbook.
- Worksheet – A Worksheet is a single visualization within a Workbook; whether it\’s published to Tableau server or not.
- (Unmanaged) Data Sources – An Unmanaged Data Source is a data source that can only be seen and used within the context of a single workbook.
- Published (Managed) Data Sources – A published data source can be used across multiple workbooks.
- Metadata – Metadata is data about things like data sources, worksheets, dashboards, Tableau Storys, actions, filters, etc…
- Schedules – Schedules are definitions of when tasks are run
- Tasks – Tasks are created when a refresh or subscription of a workbook, dashboard or data source has been attached to a specific schedule.
- Extract Refreshes – An extract refresh contacts back-end systems to pull new data into a published data source hosted on the Tableau server on a designated schedule.
- Subscriptions – A subscription sends the recipient a current version of a View on a designated schedule.
- Alerts – Alerts, a content type new to Tableau server 3.X and above creates an automated notification based on user-defined threshold criteria.
- Groups – A group is a collection of Tableau server users
- Users – A user represents a user account on the Tableau server.
Yeah, there are quite a few content types living on your Tableau server.
Now that we know what type of content that lives on the Tableau server, the next article in the series will introduce some of our favorite Tableau Content Management techniques.
Receiving regular notifications that your automated data extract refresh has failed is a great new feature of Tableau server. There are times when you may want to unsubscribe from receiving the updates. Here’s how…
- In the upper-right hand corner of the screen, click the arrow to the right of your name and select My Account Settings from the menu options.
- At the very bottom of the page you’ll find the Email Notification section. This is where you can disable extract failure notifications. PLEASE NOTE: This will turn off notifications for all of your scheduled extract failures.
While executing a routine I’ve successfully done a number of times in this exact environment, I learned something about the Tableau Server TABADMIN command line tool. IT CAN’T SPELL.
Take a look at lines 67 and 68 (sorry for the small image). They’re ALMOST the same; well sort of….er.
The subtle difference in these two lines kept me up all night. Why? because I’m a good IT community member, I run big jobs during off-peak hours and it took 2 hours for this job to fail each time. AAARRRRRGGG!!!
I digress, back too my discovery. After a little digging I found the TABADMIN configuration I needed to address the query limit timeout error (backgrounder.querylimit). An error that created a mild interruption of the GS/TO game. The second two hour window passes and BANG; error again. At this point I’m thinking, those darn DBAs, why won’t they let my query run…LOL. It’s a really good thing no one was available to field my tech support complaint (almost kick yourself).
So this morning I connect with the DBAs, explain my dilemma, and they respond by saying, well, on our side your query ran successfully three times last night. What?
So I do what any rational admin does, I open a ticket with Tableau. While waiting for support to call me, the light came on. Something’s wrong with the configuration property I set. The portion to the left of the period is missing theer.
Moral of this story? Tableau Server admin/architect, check your spelling because TABADMIN won’t correct your spelling and also won’t warn you about a bogus configuration setting. As good admins, we must avoid kicking ourselves at all costs!
Have fun out there!
You install Tableau Server , and being the good Tableau Administrator you are, as soon as the installation finishes you head over to the server Status page to make sure everything’s running as expected. You anxiously await the page loading only to find the File Store process synchronizing.
At the onset, this doesn’t seem to be something to worry about. Since the File Store is a new process introduced in Tableau Server version 9.X, it’s reasonable to assume that it would need to synchronize the first time out of the box. Reasonable thinking on your part. However, after 30 minutes, then an hour and a full day when the process hasn’t finished synchronizing, you have a reason to be concerned. Aaaarrrrggggg!!!
Reproducing the Behavior
According to the Tableau Server documentation, this new service handles the storage of extracts and in highly available Tableau Server architectures it ensures that extracts are synchronized to other file store nodes so they are available if one file store node stops running. But wait! I don’t have a Tableau Cluster, I have a single node installation of Tableau Server. I know, it threw me for a loop for a while as well.
Choosing NOT to install the sample content while configuring Tableau Server during the installation routine is the condition that creates this scenario? When Tableau Server starts up for the first time, and there’s no content to actually synchronize, you’ll see the File Store process behave in this manner.
This may seem laughable, but I’ve done it repeatedly, so I know it works. Simply publish a test Workbook to the newly installed Tableau Server, wait a few minutes and watch the File Store process stop endlessly synchronizing. Then simply delete the Workbook from the Tableau server.
PLEASE NOTE: Adding Worker nodes to the Tableau Server installation will not resolve the issue.