This article is written by Matt Dent, BI Developer at Data3. It aims to help you choose which BI tool is right for you; whether you are using it for personal use or for a large enterprise. Which part of BI tooling is most important for you and your needs.


First off, there are lots of companies that take out super in-depth reviewing processes to rate and decide on which tools are the best over a variety of categories. Take bi-survey.com as an example, they’ve covered a huge array of BI tools and pitched and rated them at over a vast range of factors. Some of the outputs can be seen in the charts below:


Another credible source is Gartner. Every year they produce a magic quadrant for BI tools, this is the latest quadrant:

Microsoft (Power Bi), Tableau, Qlik and ThoughtSpot were the only four tools to make it into the leader’s quadrant. So, we decided to declare war between the top three ranked tools as determined by Gartner. To do this we recruited the help of Richard Speigal of Nationwide Building society and Matthew Williams of Lloyds Banking group, who know Qlik and Tableau inside out. Matt Dent of Data3, (who organised battle logistics) is a self-confessed Power BI guru.

So, this is what the BI armies looked like:

We decided we were going to battle over six key areas of BI tooling – six BI territories. Who would be crowned king/queen of the BI world, and claim their territories? We ran a scoring system at the end of each section. 3 points for a 1st place, 2 points for a 2nd, and 1 point for coming in last.


First up is Connectibility. Unfortunately, and in a rather dull fashion, we decided all tools are unbelievably good at connecting to data sources. This is exactly what they are designed to do. Richard made a very good point – if your data source isn’t easy to connect up to with one of these tools then it is most likely that you need to rethink your data source, rather than your BI tool. Three points apiece. 

In the second key area to fight for the BI thrown, PowerBi went first. In PowerBi, it is easy and intuitive to make simple dashboards and ones that look great. You can import background images, create theme files, and use standard visuals. These standard visuals will be all you need for 95% of analysis. For the other 5%, you can usually find a custom visual that fits the bill – it must be said though, some of these can be clunky and have limited formatting functionality. Overall, it is easy to make most of the analysis you’ll ever need to create look great and not take too long to make. 

Qlik entered the battle – Richard felt uneasy, he knows this is Qlik’s weakest segment. Qlik’s viz vocab has 50+ standard visuals, more than enough for enterprise dashboards; there are also extensions which allow custom D3 input, as well as third party extensions. With that said however, they are never going to look quite as great as the offerings of the other two contenders. 

Tableau gives you complete freedom to create anything you want – from Sankey charts and overlaid maps to where players are on a football pitch as it uses a Cartesian plane. It can be complicated and time consuming, but with a huge amount of online help it is doable for the more experience analyst. Parameters and sets put power in the hands of the end user meaning they can explore the data until their heart’s content. These can be built into the dashboard to make it intuitive to use. Tableau’s preattentive attributes – make differences as easy as possible to see – reducing cognitive load. Conditional formatting of colours, sizes and shapes can all be changed based on logic in a calculated field which makes it far easier to bring out insight

Power Bi – When Matt first came to Power Bi, he absolutely hated the languages that Microsoft uses in the tool. Creating calculated fields in Tableau seemed simple, then creating measures and columns in PBI seemed overly complicated. You have DAX which is a bit like the formulas you’d write in your Excel formula bar, you can also write M query, which is based off F#, and then you can also integrate with SQL, R and Python. For a beginner there are too many options. For someone advanced, it’s great to have the flexibility to interrogate it multiple ways. For a beginner coming into PowerBI, it is straight forward, most functionality is drag and drop, and if you stick to DAX it’s quite straight forward and relatively similar to Excel, however more different to Excel formulas than you’d expect. 

Qlik has a super intuitive and powerful cross filtering experience, with especially great micro interactions. The selection bar builds up automatically which is powerful and easy to store selections based on slicing and filtering the data. Of the three tools, Richard believes this was the quickest to pick up and use in anger. 

To create graphs, Tableau has the show me button which allows you to create simple but good looking viz. For the creation of dashboards or reports, unlike PBI, you create your vis first and then go through the process of the dropping the viz you built into a dashboard, there is a variety of ways to do this including tiled, floating and containers (a combo of both) which makes it slower than Power Bi

In Power Bi, you have a dedicated area for transforming your data, which is powerful. You have two main options – writing with the m query language in the formula bar or using the advanced editor. All is tracked in the applied steps area; you can click back on previous steps to see a snippet of your data set for each stage which is excellent. You also have quite a lot of in-built functionality, such as: replacing errors and values, detecting data types, and transforming data types. There are some basic in-built statistical functions as well as being able to write R or Python to help manipulate the data. However, the more you do in transformation area, the more steps that you have created, the slower your report responds. The view tab is particularly powerful to have a good understanding of your data.

In Qlik, you have Drag-n-drop data loading. Qlik’s strongest advantage is its built-in ETL: Powerful load scripting and modelling capability. Some would argue this is another language to learn, however Richard is adamant it is very intuitive as it is a bit like SQL with the bits you wish SQL had: Loops and procedures, stepped debugging; all easily achievable. There is also an ecosystem of additional products, ideal for enterprise users. 

Tableau Prep – Tableau transform solution. This sits outside of Tableau itself in its own dedicated programme. It is similar in nature and somewhat in appearance to Power Bi’s offering. This area is powerful but can be time consuming and limited in functionality.

For PowerBI, The main option is to publish your report into the PowerBI web app. You can create workspaces that you can publish your reports or dashboards into. There is quite a lot of management options in the app, where you can setup things like scheduled refreshes for reports that have used the import mode to digest data. There are two main problems at the moment with this for me. Licences – to view anything in the app you need to have a PBI pro licence – these are quite cheap at £7.50 per month per user – but this cost could ramp up quickly if you had a lot of people needing to access these reports. This can also be a blocker for some clients with the extra IT resource and management this requires. The other problem is gateways – to have automated refreshes on data, your data either needs to sit in the cloud or you need to setup a gateway between the PBI app and your datasource, I think this is an overly complicated solution from Microsoft. If you need something more powerful there is premium licencing – embedding workbooks into your own workspace with dedicated capacity – which is great but expensive and only really an option for larger organisations. 

For Qlik, individual viz can be exported to image files or Excel. Qlik has an area called “Stories”. This area is like building a slideshow – easily shared to PowerPoint or PDF. Nprinting is an area in Qlik for automated reports and alerts via triggers. At Nationwide, Richard and his colleagues discourage the use of Nprinting – they want a full self-service offering and user experience. 

For Tableau, there are four options:

1. Tableau Public – share public datasets and viz using latest version of Tableau; this solution is free but not private nor is it secure.

2. Tableau Reader – use and interact with Tableau viz for free if you get sent the Tableau workbook (no live data connection). This package requires quite a lot of manual labour – so if you have regular reporting this won’t be the solution for you.

3. Tableau Online – hosted by Tableau but can be customised to your company/agency. Is outside of company’s firewall (direct live data connection) but you need a licence to view this area which can be expensive.

4. Tableau Server – Upload desktop workbooks to internal server space (direct live data connection) – again it is necessary to have the correct licence package.

The final key factor, the final area to take – Alertability. The last two rounds have kept the scores even across the three tools, did the last round split the tools apart?

Power Bi’s alert functionality is basic. A user can set up alerts for themselves which can push notifications; however, they will only be built up for that user. Alters can only be setup on a small number of standard visuals. Alerts are setup in the Power Bi App, however only in Dashboards, not reports – Dashboards are a collection of visuals from multiple different reports. 

Qlik doesn’t do Alterability particularly well either. Nprinting doesn’t do this to the standards you may expect. Qlik does have insight bot, which is smart, using natural language, which are user driven and can be integrated with tools such as slack. However, it is an additional and expensive product. 

Tableau’s functionality is similar in nature and function to Power Bi. You can send alerts from Tableau server when data reaches critical point or threshold. Able to select specific users to send the alert to in email with a custom subject line. You can’t do this in Tableau Desktop, so would need the relevant licence for this. 

If altertabilty is a definite requirement for you or your business, then a different tool, such as Looker or SiSense may be a good route to explore.

THE FINAL SCORES:


The final round shaped the podium, with Tableau claiming the win from Power BI by one point, and Qlik a further one point adrift. It must be said, all three tools are excellent, no tool out of these is a bad choice. Your decision comes down to what you are using it for, who you are, and your BI skillset.

Honourable mention – Google Data Studio.

Google’s offering is free, very simple to use and is great to create simple and informative dashboards. It isn’t as powerful or as functional as the big three, however, for a free tool – it’s exceptional. And it comes up trumps for shareability, especially if you are using the google stack. 


The Big BI battle was originally in the form of a webinar. If you would like to watch the on-demand version of this battle, here’s a recording for you: