Qliksense vs Qlikview vs PowerBI

[R_Minty45]

I used QlikView for about 3-4 years and it was very powerful. It met almost all requirements we needed. The scripting and language was fast and intuitive. The only issue that it had was for end users, using it in the browser was slow and the visuals became a little dated. Using it on the desktop was significantly faster.

I started to look into Qlik sense, and they basically went all in on browser technologies. It looked faster and more modern, until you start to develop in it. That's where it became a pain. They moved the speed of development (scripting and loading data) and made it slow, so that it became faster for the end user in the browser. The Qlik sense desktop app is nothing more than a browser.

I gave up on Qlik sense to try Power BI ... which looked like it had all the benefits as QlikView and none of the drawbacks... and it was an utter failure too. Power BI was such a immature product (couldn't take care of 3/4ths of the QlikView features) and was exponentially slower to develop for (scripting is slow, any change you make takes minutes to update etc.). And sure enough, you look into it, and a lot more "web" technologies are underneath PowerBI.

So my question is, is this what we are left with? Slow, laborious, report creating? Is there a better product out there that doesn't fall into the same problems? Is Tableau any better? chatrandomazar

Thanks

[marcus_sommer · Accepted solution]

Of course all products will have their strengths and weaknesses and I don't know all and most of the few not very well - but I'm quite sure that there is no better BI tool than Qlik. Especially in regard to the speed and easiness within the development as well as the performance and usability in the front end.

But like always you need not only the right tool else you must be also capable to use it sensible. The tools itself aren't intelligent in any way although all the sales-people promise the opposite - just loading the data and all the magic happens by itself. For the few (prepared) demo-scenarios it worked of course but the real world is different and above all more complex. You don't need just solutions for each single challenge else all must work smoothly together - and here you will need some knowledge of your own to orchestrate all the tasks. Usually it's not a single application else n ones within an environment which itself is embedded within multiple other ones.

Quite often a task starts with grabbing some of the related data and loading + adjusting them. That's not generally wrong because you need to know them before you could make a sensible planning. But you need more - knowing the processes from where the data come, what they should contain, which information aren't not included and/or needs to be created from an ETL chain and which expectations and requirements comes from the business - and you need to comprehend this quite deeply before you could really start to plan how and when and with which resources should the development + deployment be done. Personally I prefer to use some paper and a pen to make this work (ok. that's quite old-school but the essential point is not to begin the work before you had planned them). The possible and available tools are widely inferior until this work is done.

Everything what's not properly done at the beginning - conceptual and within the datamodel - else which comes later step by step and/or is just tried to solve within the UI is much more expensive … and you could easily strike down the biggest server-cluster with a small dataset and a bad designed data-model + UI. In my experience projects do seldom fail because of the used tool else mostly because of the fact that they has not been placed correctly.

It's probably not the answer you wanted to read here … but maybe it's nevertheless helpful for you and others.

- Marcus