4 of 4 people found this helpful
You ask some great questions! You might try re-posting this question in the main forums where the question would have more exposure and likely get more responses (with luck you’ll get some on this from the folks who follow me).
Some resources for comparison include BARC, Howard Dresner’s Wisdom of the Crowds surveys, and of course Gartner (in particular check out their Critical Capabilities report in addition to the Magic Quadrant). Also IT Central Station and TrustRadius have various of user reviews, and Ken Black has been doing a series of Tableau and PowerBI comparisons at https://3danim8.wordpress.com. Plus I recently responded to this Quora thread on Tableau, Qlik, and SAP where I added some thoughts on PowerBI: https://www.quora.com/Which-tool-has-a-better-future-Qlikview-Tableau-or-SAP-Lumira/answer/Jonathan-Drummey-1.
It’s hard to get super specific about pros & cons of the different products because of the variety of use cases, organizational resources & requirements, data availability, etc. is such that what may be an awesome fit for one organization could be an awful fit in another. Another piece is that where MS and SAP offer “full stack” solutions and Tableau doesn’t we can end up with apples to oranges comparisons unless we are careful to isolate the variables. My suggestion is to take the time to do a side-by-side evaluation with one or more specific use cases for your organization.
Part of the reason I suggest the evaluation is that what I’ve seen and heard the most from others is that a number of features in in Power BI and SAP Lumira are “skin deep” and just not as advanced as where Tableau is at, three examples are quick table calculations, control over dashboard layout & interactivity, and geospatial visualization. So it’s worth taking the time to get beyond the cool demos that each tool can do and find out what’s right for you.
When I hear through the grapevine about orgs going with PowerBI or SAP over Tableau it seems like the primary driver is perceived cost. On the licensing side PowerBI is cheaper to license than Tableau and SAP has been known to toss in Lumira licenses with the rest of their stack (so it’s hard to beat “free”), on the resource side organizations may have an existing investment in the MS or SAP stacks and want to leverage that. Plus if the organization has ETL needs that Tableau can’t meet then that effectively increases the cost to fill that gap. The reason why I write “perceived cost” is that the bottom line cost of licensing & maintenance is faster/easier to see and evaluate than the costs associated with time to insight, quality of insights, and time to delivery of insight that is what Tableau is very purposely designed to improve. Also Tableau has an advantage with Tableau Public and the Tableau community. Tableau users can share and download workbooks and how-tos in ways that I don’t see in the other tools, though I imagine PowerBI will get there out of the likely large volume of users.
In terms of performance questions I break that down into four parts:
a) Firstly Tableau has been in the wild several years longer than Power BI and SAP Lumira and applied to a wider variety of data sources so we’re more likely to hear about performance issues than the other tools.
b) Could Tableau issue more efficient queries to improve performance and reduce load for live connections? Yes. The sheer number of ways Tableau is used is such that there’s no way the Tableau devs can always make sure that the query plan is always the most efficient possible plan. Also a known issue with Tableau’s performance is large fact-to-fact joins with different granularities, Tableau is working on that.
c) A certain number of Tableau performance issues can be attributed to dashboard developers ignoring existing performance recommendations, for example putting too many worksheets on a single dashboard, using a ton of only relevant values quick filters, and or making ginormous text tables (potentially with conditional formatting). Ideally Tableau would “just work” in all these situations (and I know the Tableau devs would like that to be the case) but it’s not perfect. I always recommend to clients to follow the guidelines in the Designing Efficient Workbooks white paper.
d) What we see again and again in Tableau deployments is that Tableau enables organizations to ask more and different questions of their data and invariably that starts highlighting situations where the existing data sources (including data warehouses) have not been designed or optimized to support answering those questions. Tableau can then get blamed for bad performance when it’s attributable to any number of other reasons. Here’s an example that I’ve seen multiple times: a given org’s existing data infrastructure supports either detailed reporting (a larger number of dimensions) in a very small window of time (days up to a month or two) or very coarse reporting (a few dimensions) over a large window (like years of data), and once they get Tableau in place they want to build views that display multiple years of data across a large number of dimensions and that’s just not supported by the existing infrastructure. In those cases we start looking at optimizing the data source (a knowledgeable & responsive DBA can often do wonders), using data extracts, aggregated extracts, and/or investing in better software & hardware.
All the tools you mention can connect to data sources that support a huge number of records. Tableau also has the data extract engine to deal with sources that are too slow (and sometime in the next year or two will be replacing that with Hyper to improve performance), I don’t know what equivalent Power BI & Lumira might have for that. The display side is a little different. Power BI and SAP Lumira are both limited to displaying thousands of marks and will do sampling in views that would be generating more marks. The sampling works for a number of people, however for my work I’ll regularly want to visualize 1M+ marks (while using position, color, etc. to help me see patterns in the data) and to the best of my knowledge that's not possible in those other tools. BTW the most I’ve heard anyone render in a single Tableau viz is 800+M marks.
3 of 3 people found this helpful
Jonathan's answer is right in so many ways and definitely check out the links he posted.
I know some people get asked to go do research to compare Tableau against Power BI or other competitors with the end goal being a checkbox/checklist of features. Let me caution you that this can lead to a lot of heart ache since not all features work the same way. A great example is something like grouping in Tableau vs. Power BI: In Tableau you can visually "grab" all the points you might want to group together using a box, a lasso, a circle, whatever. In Power BI you are required to control-click on each single dot/mark you want in your group. Might seem like a small difference on paper, but try it on a map with thousands of dots with some overlapping. A thing that takes 1 second in Tableau suddenly takes 5 minutes in Power BI. It's these types of differences 100 times over that really start to add up and why doing a side-by-side comparison of solutions is the only way to do a real evaluation inside an organization.
But again, sometimes you need to get a feel for features gaps, etc. I know lots of people will go take a look at Power BI's Ideas board (and just let me say, kudos to them for having something like this for their users to add their voice to what they want) to get a feel for what they currently don't have features wise and what they are working on: Power BI Ideas: Top (6962 ideas) – Microsoft Power BI
As of today it looks a good chunk of the most voted ideas for Power BI are all things Tableau already has:
- Previous level drill down
- Ability to deploy premises (i.e. not in the cloud)
- Global Filters
- Desktop for Mac
- Quick filters that support dropdown, multi-select, calendars, relative dates, wildcard, etc.
I know a couple other areas where Tableau customers who looked at Power BI sometimes got surprised were things like the 3500 data point visualization limit (you can't build a scatter plot, map, treemap, anything really that has more than 3500 points displayed at one time). Also the 1 GB extract limit is a showstopper for some folks, as is the 10 GB total storage per account.
Again, all of this really has just convinced me that whether it is Power BI or SAP or any other solution, the only way to get an accurate comparison it to figure out what is critical to your organization and then put each solution through the paces in real world settings with real users.
Hope that helps.
1 of 1 people found this helpful
Check out the video below where I compare Power BI vs Tableau and many important aspects. Let me know if it helps.
Just What I'm Looking For!
Nice and good Video. It is very useful for me to learn and understand easily on Comparision of Tableau vs Power BI. Thanks for sharing your valuable information. I suggested to my colleagues and they also liked.