Power BI: What we love, what we hate, what works, and what to look out for.
During the 15+ years I’ve been working in data, I’ve used a lot of tools. My favourite, by a mile, is still the good old Excel and I will not apologise for that. That said, I have dipped my toe into other data lakes (get it?), with Tableau and QlikView being some of personal favourites.
Though I’m an Excel disciple, it is strange that, until recently, I haven’t used Power BI in the “real world” much. I’ve trained myself on it, but I didn’t need to use it for business purposes up until last week.
Running Brijj has been keeping me very busy. So busy, in fact, I’ve only just got around to building a “Business Overview” Dashboard. The irony isn’t lost on me, given that my product helps speed up data projects!
There are lots of metrics that matter to a start-up, and SaaS in general. But I needed to build a dashboard with some KPIs for myself and my co-founder. In a nutshell, we needed the following:
- Number of visitors to our website
- The proportion of those visitors engaging and becoming leads
- How many of those leads are likely to become paying customers
- How many prospects will we lose through churn?
- If we turn our marketing spend up or down, how will that affect visitors
- What all this means to our projected financials
Normally, I’d fire up Tableau as it’s my go to for building interactive dashboards, but this time I thought ‘hey, why not fall into the loving arms of Bill Gates & Satya Nadella once more? Let’s give Power BI a go!’
In about four hours I whipped up a dashboard profiling everything I needed. Which I can’t show you because I’m into data protection, of course. So instead, here’s a picture of Tech’s greatest dynamic duo:
What I love about Power BI
Modelling in Power BI is so easy! Getting all my data sources together was quick and easy. I needed to join data from Google Analytics and Mouseflow for our website. We’ve been running some add experiments too, so I needed data from platforms like Google, LinkedIn, Facebook, and Reddit. I needed data from MailChimp, Pipedrive, Help Scout and OptinMonster to get a clearer picture of engagement. Finally, I needed to import financial data from Xero and Revolut.
I did some of this through direct integrations in Power BI, for example with Google Analytics. For platforms with no direct integration, I used a combination of Smartsheet and Zapier, then linked to Power BI through Smartsheet. Easy!
Once that was done, Power BI came into its own. Power Query is top level for data preparation and transformation. It’s at least comparable with Tableau Prep, if not better. You can completely transform your data in quick and easy steps, building out a logical flow which you can amend on the fly. This gives you a structure of queries and tables which can relate to one another. Truly brilliant.
I also love the visualisation and dashboard aspect of Power BI. It might not be as intuitive as Tableau or Qlik Sense, but it does a great job at creating tidy, clear dashboards. It also has a couple of details which exceed Tableau from an aesthetic perspective – for example, the way that visualisations animate when drilling down into data. It’s a small detail, sure, but I like it. The choice of visualisations is also great. The debate around Pie Charts and Gauges will go on forever but whatever you stand, Power BI gives you 99% of the visualisations you’ll ever need. Tableau used to get very touchy about the ‘right’ thing to do with visualisations. It amazes me that it refuses to install certain types of viz for their customers. Give the people what they want, Selipsky!
What I hate about power BI
Just to be clear, Power BI is great. The issues I have with it are reflective of the issues I have with Microsoft in general. I know they are huge, with hundreds of products and varying development pipelines…so why is nothing ever easy?
I didn’t come across much I couldn’t do in Power BI. I just don’t understand why Microsoft makes things so unnecessarily complicated at times. Take Parameters for example. In Tableau, you can create a parameter which is a dynamic value you can assign to formulas, filters, and visualisations. It’s the go-to tool for what-if scenarios. When you try to find out how to use parameters, Tableau’s knowledge base outlines it in an easy way. They care about user experience and put in the time to make things simple. Microsoft? Not so much.
To show what I mean, imagine I was a complete novice, and I typed the following into Google: ‘I want to change a value dynamically…’.
If I add the words ‘to Tableau’, this is the first link that comes up:
Great, now I know that ‘parameters’ are what I’m looking for and I can easily find out more.
Now, if I add ‘to Power BI’ instead, the top result is this:
Don’t bother reading it. The bottom line is it wouldn’t help. The first article authored by Microsoft themselves is nine links down! Now, I know that Google search rankings aren’t Microsoft’s fault, but this small frustration is an example of something Microsoft fails at repeatedly.
Microsoft don’t appear passionate enough about user experience. Things are always a little bit…annoying. There are lots of examples of it in Power BI. Why can’t I easily build parameters? Why can’t I easily change the width of a bar? Why can’t I change a line graph to a curved line? And if I can’t do something, why is it always hard work to find out how!?
I’m aware that these are minor concerns, but the reality is that people want to use products that are as free from frustration as possible. I’m afraid that Power BI has inherited some of that Microsoft clunky culture.
Having gotten my general annoyance of Microsoft products off my chest, I have to say that Power BI is a good product. Its modelling and dashboarding is excellent. Although for adhoc analysis it trails Tableau. It’s a little clunky and has fundamental functionality missing which can be annoying. Having said that, in my view, Power BI is going to dominate the BI tool market for years to come. I’d bet it will become a fundamental business tool like Excel and the office suite.
Power BI is great for:
- Quick and powerful data transformation and modelling
- Functional dashboards
- Natural move forward for Excel nuts
- One of the more “affordable” solutions
Power BI is not as good as others for:
- Adhoc visual analysis. Tableau is better
- Non-Technical users. Again, Tableau is better
- Creating beautiful visualisations. You guessed it…Tableau is better!