Power BI not that great?

I’ve been talking with a guy at work who hasn’t used BO before and is an MS stack person. Having explained to him about the semantic layer and how it works with Webi, he was stunned that it’s on the downturn given that its functionality is far better (in his opinion) than what SSRS/Power BI offers.

There could be hope for us yet!


brucecy92 (BOB member since 2020-06-02)

I’m working with both at the moment - maintaining a legacy BO system while transitioning into PBI. What I’m noticing most is that it’s not so much about the front end display tools, but far more about configuring the back end warehouse/model. For the average user, both toolsets will do 95% of what is wanted, but PBI is slick and shiny and easily configurable, whereas BO seems clunky and requires major resource/money to upgrade anything …

Debbie


Debbie :uk: (BOB member since 2005-03-01)

The possible transition from BusinessObjects to another tool is always blocked by the absence of ‘universes’. How do you handle this hurdle ?


RikDeclercq :belgium: (BOB member since 2006-09-28)

As a user of both, the lack of a true semantic layer is a massive hole. Ive got approximations of similar functionality (i.e. publishing blank reports with the dataset built and then using the dataset as a datasource) but its not the same thing. Nothing can be done in SSAS to help either with a large backend database as SSAS OLAP doesn`t suit and SSAS Tabular is relatively simplistic.

Try working with PowerBI over a large (80+) table OLTP database and its in-memory model is also painful, its not like you can bring it all in and model it in the same way as a Universe as it would have to bring all the data into memory which is massively inefficient. Then add to that the fact that PBI is limited when it comes to joins. Then using directquery isn`t great in this scenario as it then limits some of the funky thinks you can do with DAX and Q&A.

3rd thing, PBI isn`t truly 3 tier yet, to use the tool properly you have to install and maintain a client due to some of the current limitations (e.g. only desktop can consume a PBI dataflow).

That said, I`m a massive fan, mainly due to the power of DAX and the M language.

What I`ve found though that that power sometimes gets put in the hands of the end user with their datasources all over the place, so you just end up with a data governance nightmare.

I could actually see a lot of usecases for supporting both in a business to be fair!


ABILtd :uk: (BOB member since 2006-02-08)

Hmm. We’re using SSIS dimensional models and not having any issues with joins. The database is complex with a lot of tables. Done properly, it’s actually quite slick and we’re finding we can model things that Bobj would never cope with. Pretty much everybody is using the online Power BI, with only a few superusers/report designers using full client.

The biggest issue I’m seeing is that with Bobj, if a user wanted a new object adding, it was relatively simple - check it in SQL, add to universe, test and push to live environment. Sometimes a ten-minute job. In the new world a simple change can sometimes require a lot of below-the-bonnet tinkering. It means the whole change-control process is becoming longer and more rigid - but I guess that can be a good thing as users consider what it is they actually want.

Debbie


Debbie :uk: (BOB member since 2005-03-01)

Debbie, so you’re replacing BO with SSAS combined with powerBI ?


RikDeclercq :belgium: (BOB member since 2006-09-28)

Yeah, bit of a change control nightmare in SSAS DMs, always was (I started using them with SQL Server 7 in 1998 - HOW OLD!). Theres that and they arent much use with OLTP databases either, great on a dimensional schema though (thats what OLAP cubes were built for to be fair).

Its worth looking at SSAS tabular models for more rapid implementations of ad-hoc things, as its more similar to building a dataset in PBI (i.e. you can use DAX). A little birdie tells me there might be a 2 way migration path between SSAS Tabular and PBI Datasets in the future as they are afterall essentially the same in memory technology.


ABILtd :uk: (BOB member since 2006-02-08)

Effectively, yes.

We’ve got a legacy Xcelsius suite of dashboards too - quick fix (to dump the servers and Xcelsius asap - it’s become very flaky now) was simply to take the SQL -generated XML which used to power the dashboards and use it to create a PBI dataset. Not particularly flexible, but a good interim solution pending proper redevelopment.

Debbie


Debbie :uk: (BOB member since 2005-03-01)

totally agree this!

I did BOXI since September 2000 and started using PowerBI since late 2017.
Did several PBI projects now and saved much more develop time, more happy people, more stunning visuals, much cheaper…

And remember you could treat a PowerBI Dataset as a BO universe.

Difference is that a dataset contains also the data but …performance is very much better. And you never, never, never have to deal with Webintelligence Processing servers and falling Tomcat servers :o


TurningPointHolland :netherlands: (BOB member since 2006-09-06)

Some of our universes contain more than 100 tables. I guess these can’t be transformed into Power BI datasets or SSAS models.


RikDeclercq :belgium: (BOB member since 2006-09-28)

We’re facing some of that data governance nightmare. With WebI reporting, we had “one source of truth”, now folks are going and finding and writing whatever they want … with many legacy systems and inconsistent data … the results don’t always match reality.


KelleyHux :us: (BOB member since 2002-07-05)

In PowerBI you can combine tables into 1 to make it more comfortable for you. You can also choose which tables not to load everytime.

It is also a good moment to re-think about your datamodel. And last but not least you can split your model into more than 1 PowerBI dataset


TurningPointHolland :netherlands: (BOB member since 2006-09-06)

We probably have models that big now. But most of our universes were built over relational models or operational data stores. With PBI, we are moving towards proper fully dimensional models. Seems a lot more work up-front, but t e results are impressive.

debbie


Debbie :uk: (BOB member since 2005-03-01)

Hi, the way you should play this is to force your users to use pre-built (by an expert!) and pre-published datasets or PBI dataflows as datasources. As an admin you can even certify datasets and provide links to documentation for that dataset. It gives you some governance and control as to what people are consuming in terms of self serve data. It often helps if people build documents in PBI service as opposed to PBI designer, as the latter gives them free reign (which Ive seen run rampant on many occasions!). Its the closest to 3 tier SAP BI that I`ve rolled out.


ABILtd :uk: (BOB member since 2006-02-08)