skip to Main Content


Yeah I’m a Bit Slow…

Back in October, Microsoft released a roadmap blogpost detailing plans and direction for their BI platform.  I kinda pressed the Snooze bar on paying attention, because hey, I was super busy, plus…  I always tend to take my time with announcements like these.  I do my best “digestion” once the noise has died down, and my best thinking slowly, in quiet.

Spoiler: Mostly Positive

Words like “Bad” and “Ugly” are overkill, as my reactions are mostly positive.  I just love this movie and have been dying to create/pervert an image like the album cover shown here.  I was raised on Star Wars, yes, but also on Spaghetti Westerns, thanks to my father (my middle name is Clinton, because he originally wanted to name me Clinton Eastwood Collie – not a joke, my mom intervened).

What “negative” feedback I do have is constructive in nature, rather than disparaging.  So if you came here hoping for a bloodbath, move along, this isn’t the post you’re looking for.  It’s not all sunshine and double rainbows either, so if you’re a Microsoftie hoping to hear that I love everything, well, that’s not quite happening either, but please stick around and read what I have to say.

The Good

imageThey explicitly come out and say they want Power BI to be used by a Billion People.  I like this kind of ambitious goal, and it absolutely strikes me as Completely Feasible.  Heck, we have that many people (more, probably) creating and consuming Excel-based analytics already!  Why shouldn’t we set our sights lofty?  And besides…  goals like these are ONLY attainable when MS BI fully embraces the Excel crowd.  They’re not going to get there just by direct Power BI adoption.  So goals like these tell me they are seeing things properly (or at least, will be forced to over time, muhaha).

They are working on an “on premises” server for the publishing of Power BI Desktop visualizations.  Bravo, bravo, bravo.  Years ago when Microsoft announced its “Cloud First” mantra for BI, it was clear to me that they weren’t going to do on-premises versions of the new stuff unless they were forced to.  Well, they’ve been forced to, because, in my experience, the bigger the customer, the less likely they are to embrace cloud-based BI (and big customers are overwhelmingly the most important to the MS business).  For more than a year, I’ve been crossing my fingers that MS corporate would be pragmatic enough to start building out the on-prem version.  This is awesome news.

It sounds like they might be building an on-premises version of Web Excel that does not require SharePoint!  I’m reading between the lines here a bit, because the blog post didn’t explicitly say this, but the post does seem to strongly imply it.  I mean, they’ve already done this in the cloud, by allowing you to pin Excel reports into your dashboards.  This would be a HUGE development for on-premises customers, because a) SharePoint is one of the most fidgety servers in the world to install and maintain   and b) SharePoint tends to involve a heavy political cost as well, because it tends to cross over into other factions of IT outside of BI land – at one of our clients, the deployment of Power Pivot for SharePoint took more than THREE YEARS because of such politics.  Sidestepping all of that would be a big win.

Excel continues to have a prominent place at the table.  Keep in mind that the blog post in question was written by the SQL team, not the Office team, and as such it’s natural that they’d emphasize their own products more heavily, but Excel does appear in the roadmap as one of the four canonical types of reports.  In related news, the name “Power Pivot” will survive indefinitely within Excel – so much so that we were enthusiastic to maintain its presence in the title of our 2nd edition book – and even the SQL team continues to build around Excel (witness the ability to pin Excel reports to Power BI dashboards and the new Excel addin for doing precisely that.)

The Bad

imageHey, nothing in there is Bad, in the way Angel Eyes was in the movie.  (Some people are always expecting there to be Evil lurking behind the scenes at Microsoft, and no, there just isn’t any such thing.  Academic hubris and big-company-clumsiness are the only “evil” forces we will find.)

But if we’re talking “Bad,” in the way Lee van Cleef himself was bad…  you know, like Michael Jackson was bad, or George Thoroughgood was bad to the bone, well, I think our friends The Engines (as in, the DAX and M engines) qualify.

The Ugly


Or maybe we should call this section The Awkward.  Cuz that’s a decent umbrella under which to collect my thoughts.  (And besides, Eli Wallach, in hindsight…  ugly?  I don’t think so.  Did you see him in The Holiday?  Such a sweet man.)

All of the on-prem stuff is being organized “under” SSRS.  Yep, SQL Server Reporting Services is being positioned as the Mother Ship for all of the on-prem investments.  As in, if you want on-prem Power BI Desktop publishing, or Web Excel, or Datazen, you need to install an SSRS server (as opposed to, say, Microsoft’s Hot New On-Prem BI Server, which is what *I* was expecting them to do).

As a technical detail, fine, OK.  But as a description…  look, Power BI Desktop publishing, on-premises, can only be sensibly described as “Power BI On-Prem.”  In any conversation with customers, that’s what we’re going to be calling it.  Calling it “SSRS Interactive,” as it’s described in the blog post, would be super, super opaque and I can’t imagine anyone in the real world responding well to it.

And isn’t SSRS a pretty non-sexy vehicle?  I mean, it’s been around forever, it correctly has a reputation as Yesterday’s Static Report Server, and it’s lost whatever “coolness” it once had in customers’ minds.  Why would we “stamp” all the Brand New Hotness as if it were yesterday’s old news?  If I want people to be properly excited about the new on-prem server, I’m going to (truthfully) tell them that it’s not SSRS, but instead Microsoft’s Next-Gen On-Prem BI Server.  Web Excel, Power BI, and Datazen have nothing to do with SSRS.  Why associate them like this.  (Yes, I get it, SSRS has a huge install base, so there is that as an advantage.  But install base is a very different thing from mind share and perception, and I suspect this branding will backfire significantly.)


New Official Names for Things We Already Understand?  Weird if so.

While we’re at it, they’re calling Datazen “Mobile,” and NOT using that word to describe Power BI (which they are calling Interactive), which seems to imply that Power BI is NOT mobile (wrong) and that Datazen is not interactive (also wrong).  This makes me wonder if these descriptors (Mobile, Interactive) are truly going to be official terms in the released product, or if they’re just part of this blog post roadmap thingy (boy I hope it’s the latter, cuz it’s all just super confusing – we didn’t need new adjectives for these things – adding a new description doesn’t magically converge and rationalize the many different offerings, it merely makes a decent story look murky).

Instead, I would have hoped that, by now, we wouldn’t be talking about Datazen and PBI as separate techs anymore.  Tell us a story where beautiful, interactive, mobile visuals are all going to be one thing.  I’m sure that’s the eventual plan (right, Microsoft?) but it’s non-ideal that we’re heading into a major release and still distinguishing them.  If I am about to get on-prem PBI, why do I need Datazen anymore?

Other Questions

I still don’t get the MS obsession with Hybrid mode.  As in, a BI platform that simultaneously has one foot in the cloud and one foot on-premises.  It’s one thing to have a mixture of on-prem and cloud data sources.  But why would I want my UI layer in the cloud (on and my data engines on-prem (on SSAS/Power Pivot servers that I maintain)?  Seems like a whole bunch of extra network roundtrips to me (slow), as well as additional configuration requirements (and failure points) in the form of gateways.  Not to mention, it STILL is likely to trigger enterprise customers’ “anti-cloud” immune systems.  (That’s been my experience so far anyway).  Seriously, if I’m wrong about this, please tell me, I actually DO want to understand better.  As it stands, I see Hybrid as not terribly practical, and the sooner MS internalizes this (or I discover why I am wrong), the better.

But hey, we’ve got an on-prem answer on the way!  If the above sounds cynical, well no worries, because MS is already solving that problem by giving us a full on-prem version.  As they should be.  And our larger clients will love this.  They will deploy lots of it.

Right?  Right?  It’s a FULL on-prem version.  100% parity with the cloud version, and vice versa.  That was hinted at in the blog post with words like “harmonize” and “align,” but it’s such a simple thing to say, and they danced around it rather than keeping it simple, which worries me a touch.

Hey, Rome wasn’t built in a day, I know they can’t deliver parity overnight.  It takes time.  And I also understand that the cloud versions will always be “fresher” than on-prem, because it’s just so much easier to release new code to the cloud.  But it’s such a simple thing.  It’s what EVERYONE expects:  the cloud and the on-prem servers are the same thing, with the same features, capabilities, and behaviors.  Why fight the way humans work and think.  Why swim upstream against that current.  Just tell us that perfect parity is an explicit goal, and that the on-prem thing isn’t intended as a stopgap.

Rob Collie

Rob Collie

One of the original engineering leaders behind Power BI and Power Pivot during his 14-year career at Microsoft, Rob Collie founded a consulting company in 2013 that is 100% devoted to “the new way forward” made possible by Power BI and its related technologies. Since 2013, PowerPivotPro has rapidly grown to become the leading firm in the industry, pioneering an agile, results-first methodology never before seen in the Business Intelligence space. A sought-after public speaker and author of the #1-selling Power BI book, Rob and his team would like to help you revolutionize your business and your career.

This Post Has 18 Comments
  1. The Power BI Desktop is exactly 3 steps away from becoming a “Desktop Database” on steroids for the millions

    It has to :
    1. Allow Excel to Connect to the DataModel in PBI Desktop file
    2. Allow password protection of the the PBIX file
    3. Allow scheduled refresh

    Once this happens – the Access folks have to find a new job and Power BI Desktop will be on everyone’s Desktops

    All business users can then leverage

    1 Large data handling capacity of the 64 bit PBI Desktop (without having to upgrade to a 64 bit office) – and without having to worry about file corruptions

    2.The powerful ETL capabilities of Power Query

    3. The compression and Calculation capability of the DAX engine

    4. Extremely small file sizes (<100kB) of the Excel report connected to the PBI Data model

    The Cloud version on the other hand has no future at all

    1. Hi Sam! I agree with most of what you say here but I disagree about the cloud, which DOES have a future in my estimation. A bright future. But the key word is “future,” because it requires time, marketplace evolution, and an MS willingness to play it slow.

      Cloud is appealing in two places, in my experience: small and medium biz, and departments of larger biz. MS currently isn’t satisfied selling to either one of those, but the latter could make for some great infestations. for instance – was its success due to a series of IT departments saying “yep, we’re gonna go cloud CRM” or did it happen departmentally, under the radar, before IT’s sensors detected the “threat?”

      Maybe it’s too late to follow that model. Maybe BI *never* could have flown under the radar. I have to allow that both of those may be true.

      And hey, it makes TONS of business sense for MS to pursue a cloud strategy. Revenue streams like that, with no complementary costs (hardware and maintenance budgets for on-prem servers), is a golden goose for software firms and we are NOT going to ever see MS give up on it.

      But MS can’t just “make” customers come around. Right now, the prevailing enterprise mindset is No Way Are We Putting BI in the Cloud. Building more features doesn’t change that. Adding a Hybrid mode Franken-Compromise doesn’t change that. You can’t change the customers on this, you have to change to meet them where they are. And right now, that’s on-prem (unless you’re willing to “go small” and focus on SMB and departmental adoption).

      That’s where parity and naming come in. An awesome on-prem solution that looks, feels, and acts just like the cloud version provides an environment where it’s psychologically MUCH easier to “sell” an enterprise on a transition to the cloud 12-18 months after they’ve adopted on-prem. Calling the on-prem version SSRS, and having it offer a different set of functionality, is sacrificing that important psychology before it ever gets going.

      Lastly, one final way that the cloud can “win,” and it’s an unpleasant one, is via economic pressure. If the US experiences another sharp economic downturn, for instance, cost-savings suddenly becomes a higher priority for the enterprise, and suddenly, the cloud isn’t so bad after all. A weird thing to root for, but I do believe it would change the conversation.

      1. @Rob – I am not saying the cloud has no future – I am saying certain products are not meant for the cloud – “Modern Excel” is one of them – I am going to say the “Old” Excel with millions of Index/Match or VLOOKUP’s or SUMPRODUCT / SUMIFS is not suited either – It would take forever to open – Last mile connectivity is still an issues in large parts of the world.

        But the Office 365 is a good concept – where the installation (,msi) files are on the cloud -but the application is installed on your local computer and is auto updated

        However “Excel Online” -which is the “actulal” cloud version of Excel – is only good for shopping lists and making “Excel Online” – identical in all respects to the “Desktop” version is still a distant dream

      2. Just wanted to disagree that Microsoft isn’t interested in selling cloud to departments of large biz – that’s actually whey’re they’re placing most of their Azure sales focus. Most EAs already include an Azure commitment, so they just want customers to get in there and consume what they’ve already paid for, and PowerBI / machine learning at the department level is a fairly easy sell.

      1. Giang – Yes I know – but the local host No keeps changing (every time you close the PBIX file) – and it requires a 3 rd party add-in and is not officially supported – and hence can’t be used in a real life project

  2. “my middle name is Clinton, because he originally wanted to name me Clinton Eastwood Collie – not a joke, my mom intervened”

    Rob, should one hit a rough patch work- and otherwise it’s always uplifting and helpful to read your pen. Thank you.

  3. My big concern at present is the release date of this on-prem stuff which to my understanding is promised as a vague “some time after release”.
    (I’m part of one of those enterprise customers who can’t legally put our data into the cloud so it’s on-prem or nothing for us).

  4. My largest concern is that all of this “goodness” (specifically relating to the on-prem capabilities) will not be in SQL Server Standard Edition. Working with customers in clinical environments both cost and privacy are key concerns and Microsoft has been making it harder and harder to push for their platform since many of the key features are Enterprise Edition-only. Don’t get me going on the BI edition that does not include data compression or table partitioning in a data warehouse environment.

  5. Great post Rob. The differentiation between PowerBI and DataZen is particularly confusing. I was under the impression they would be taking key elements from DataZen, such as the striking UX design and integrating it with PowerBI. Why have two tools that can deploy to Mobile? I assume DataZen may read from PowerBI on-Prem engine in the future as it doesn’t have much in the way of ETL or modelling.

    1. Thanks Jon! My travel schedule hasn’t gotten started yet for the year, so I’ve enjoyed getting back to writing here 🙂

  6. Good Post! I just finish a major project and came along this post. Result – 48 hours of getting knee deep with Power BI and very little sleep. In doing so I continue to discover some flaws in my PowerPivot Data Model. I understand that currently there is a one-way communication from PowerPivot to Power BI, however are you able to update that communication with changes on PowerPivot?

  7. Hi Rob – This is Carlos from the Excel Team. I can address a couple of specific points in your post regarding Excel.

    1. “It sounds like they might be building an on-premises version of Web Excel that does not require SharePoint!”

    We are indeed decoupling Excel Services from SharePoint in the new 2016 release. The functionality previously available through Excel Services will be available through Office Online Server. However, SharePoint is still required at least in those BI scenarios that require SQL BI Add-ons for SharePoint (e.g. Scheduled Data Refresh). Also, SharePoint 2013 does *not* have to be upgraded to work with Office Online Server 2016. And in such a topology, a lot of BI scenarios *will* be supported. Specifically: working with a data model and external data access (e.g. accessing organizational OLAP cubes).

    The key benefits with this model includes removing the double deployment needed previously with Office Online Server and WAC in 2013, and allowing continuous delivery to customers (which we can do with WAC and could not with MOSS), all this without losing functionality compared to MOSS.

    We are also currently working on a blog post on this topic with more details (when ready, I can post a link), and we will have a dedicated session on this at the “Microsoft Data Insights Summit” in March. (

    2. “Excel continues to have a prominent place at the table.”

    Absolutely! Excel continues to be a key part of Microsoft’s BI Strategy. We work very closely with partner teams internally like PowerBI to drive for a “better together” story for customers. (e.g. Here is roughly how I see things:

    * For deep and powerful analytical capabilities and flexible ad-hoc work – Excel

    * For dedicated visual exploration and building interactive reports – PowerBI Desktop

    * For live, connected reports, and flexible dashboards regardless of the source (e.g. PowerBI Desktop, Excel, etc.) – PowerBI Service

    There are of course some overlaps between the offerings, but the above is where I see the current center of gravity and strategy for each. I think this overlap is OK, and that giving customers flexibility on the tools they use is important.

Leave a Reply

Your email address will not be published. Required fields are marked *