Strategic: 2008 Microsoft BI Stack

by 15. November 2009 19:59


I’ve recently had the pleasure of working with Microsoft’s 2008 BI stack. I say pleasure because Microsoft has really pulled out the stops.  The architecture is sound, the performance is outstanding.  Gartner has ranked Microsoft’s BI offering in the top quartile of providers.  This puts them in the same league as Cognos (Oracle), Business Objects (SAP) and usual players in the space.  I have used Cognos and Business Objects in the past and have had the opportunity to use or see most of the other vendor offerings.   Each Vendor offers tradeoffs.  Below is my experience with Microsoft’s BI solution up to this point.

Microsoft’s BI offering is centered around SQL Server.  In this case my team and I used Microsoft server 2008 and SQL Server Enterprise 2008.  I wish I could tell you it was because we wanted to take advantage of the performance improvements, However our logic was simpler.  If you’re going to build a new stack just use the latest and greatest.  The most important thing to the team was to use SQL Server 2008 so that we could take advantage of what R2 was going to provide.

The first thing that really struck us was the modular nature of Microsoft’s strategy.  SQL Server is the core, but in terms of delivery Microsoft relies heavily on the fact that most business users utilize Microsoft Office as their main office productivity suite.  It’s a safe bet considering that Microsoft has an incredible penetration rate in this space.  However, external facing customers might prove to be challenging in terms of delivery.  We chose to put a proxy between our Reporting Services implementation and the outside world.  This allowed us to implement URL rewriting and offer another level of security.  Unlike Cognos, which offers an easy to implement portlet we had to custom make a solution for implementation into our ALUI portal engine.  Of course all of this can be solved if you just use Microsoft Share Point.  This was not an option for us.

Typically I’d view this module nature as being a hindrance, or even a lack of integration features.  What we ended up finding is this is actually a strength of the Microsoft BI stack.  Microsoft has a great API/services to access all the functionality.   Building out a custom solution for integration was not that bad compared to the limited options for customization that other vendors offer.  For example, Users can natively access cubes and analysis right from Excel 2007.  However, because the API was so rich, we chose to write our own ‘Excel tab’ to make accessing and running analysis functions as brain dead as possible.  A solution that users appreciate.

So far, as a strategic play, Microsoft’s BI offering is proving to be a good investment.

Calling All UX Designers: BI Needs You

by 16. April 2009 15:04


The numbers are in.  Despite the economy spending is projected to increase by 2.6% in IT.   The better news for those in the Business Intelligence field is that projected spending is projected to grow at a rate of 8.1% in 2009, and that’s just on platforms!  So why is Business Intelligence(BI) an afterthought?  More specifically, why do organizations spend millions developing web portals, marketing, and user experience but skimp on ugly BI presentations?  How many of IT organizations have scores of reports, but limited adoption rates?  Can UX designers help with the beginning of BI 2.0?

I write this as someone who manages a developer team and a BI team.  What I’ve noticed is that the best practices of UI design have not filtered down to BI.  Why?  I believe that there are four main reasons:

1)  BI is somehow perceived as ‘less sexy’.  After all, It’s just a list of information already entered.
2)  A lack of understanding about how people view and consume data.  Especially large amounts.
3)  Failure to see BI as an ‘experience’.
4)  Generally the design is left to the ‘Engineer’.

The first reason, that somehow BI is perceived as ‘less sexy’, is probably due to a few reasons.  First and foremost it’s the tool sets that are offered.  Most BI tools do not allow CSS, or their standard graphics and control packages are something less than desirable.  However, in the last few years there’s been a surge in third party controls that give BI platforms such as Reporting Services the ability to make and design eye popping presentations.  Additionally the IDE’s are generally difficult to work with and lack common tool sets to encourage design centric solutions.

Number two and number three on this list is related.  UX designers will focus on design sessions that utilize eye tracking software, mouse tracking and analyze web log information to death.  But when was the last time you ever heard of this happening in terms of BI?  Do users consume data differently in BI than they do on a webpage?  I would argue not really.  It’s just a different view.  You still need to highlight what the important information is in the view, and make it easy for the user to figure out what ‘action’ they need to take.  Action being the key word here.  Often times I see BI presentations where the end product just spits out a listing of information.  Not really helpful.  If you must ‘list’ then at least have a ‘dashboard’ like presentation in the header to help me instantly understand what information is present.  I call this the 15 second rule.  If your end user can’t figure out what information is on the report, what’s good about it, and what’s bad then the ‘report’ has failed the usability test.

The last point, design is left to the Engineer, is not really a surprise.  As an Engineer, and I think I speak for most Engineers,  I personally take no offense to it and I’m quite aware of my shortcomings as a designer.  That’s why we have UX!

Recently I had the pleasure of working with a BI UX specialist, David Crandall.  Talk about an eye opening experience.  You don’t realize how bad something is until you see it done correctly.

I asked Dave how he sums up the current state of BI UX design.  He replied:

The customer can have any color he wants so long as it’s black.” – Henry Ford in regards to the Model T.

Even though the data might be life changing, those tons of reports are often tons of ugly reports.  We laugh to think that someone once thought that people would be happy with an ugly car, but we’ll throw millions of dollars at our company’s business intelligence and leave it ugly.  Obviously “pretty” is not the whole solution, but it has to be considered a valid component when trying to convey information effectively.

 

I would recommend checking out Dave’s blog.  Or if you want to see what I’m talking about in terms of rethinking BI presentation check out his resume marketing presentation.  It’s data about Dave, just not like you would think (plus it uses Prezi which is an awesome service).

What are your experiences with BI?  Are we devoting enough design time to it?  Do you think we’ll ever see a Web 2.0 phenomenon in BI?  One thing is for certain.  We will need the UX folks onboard!

Why Does Enterprise Business Intelligence Fall Short?

by 18. February 2009 18:41

 

Would you like to make a lot of money in IT?  Just sprinkle some Business Intelligence buzz words in your resume.  No really.  The Business Intelligence space is filled with impostors, resume inflators, and snake oil sales people.  If I haven’t offended everyone yet then let me add that the Business Intelligence space is also littered with vapor ware vendors.  Is it their fault?  I would argue no.  It all starts with strategy behind the Business Intelligence initiative a company initiates. 

Let’s start by talking about how your Engineering team views Business Intelligence.  Rome wasn’t built in a day, neither were your systems, and success in  Business Intelligence will follow the same logic.  So from day one your Business Intelligence strategy needs to be  at the forefront of your engineering design.  You can design the most elaborate database structure and applications that contain the most complicated business logic.  But if you can’t get the data out and make sense of it then it was all a waste of time.  More times than not this is the symptom I see in IT.

Another prolific problem occurs when the Business Intelligence initiatives come from the top.  How many times have you heard ‘we need dashboards’ or ‘we need KPI’s’.? Your Business Intelligence team then has to go figure out what these folks want to see.  Sometimes the business knows and provides you with some spreadsheet that the kid down in Finance put together from data that isn’t anywhere in your source databases.  Other times it’s in an even more colorful spreadsheet that has the executives all excite, but upon further analysis you realize that the logic or data is flawed.  Then you’re stuck in the precarious position of having to explain why the report that has been touted as the ‘ultimate report’ can’t be implemented.  All of this is symptomatic of a Business Intelligence strategy that doesn’t provide the users with what they need, a self service method of accessing the available data and reporting on it.  You need to develop a strategy to implement this, because the chances are that your Business Intelligence team is not the ‘business experts’.  The kid down in Finance is.  Provide him the data and the tools and let him loose. 

So why do I say that the the fault lies with you and not the snake oil sales people or the vapor ware vendors?  Because all they do is come in and show you the latest wiz bang gadget filled with colorful dashboards and capabilities to help you decide if and when your shoes need to be tied.  The responsibility to setup a structure that takes advantage of these capabilities is yours. 

Here are three simple thinks to keep in mind while trying to optimize you Business Intelligence development:

1) Get them involved in new development initiatives as early as possible.

2)  Concentrate on the data structure, design and processing. 

3) If you do nothing else.  Concentrate on delivery mechanisms and self service tools.  Your team is probably not business subject matter experts, but your end users are.

I’m by no means the master of BI so I welcome any and all additions, suggestions or comments!