Blog: Jim Gallo Hello, again! Hopefully, this blog will be where the rubber meets the road. This is our opportunity to create an interactive dialogue about the subject of Agile BI and to learn from each other's experiences. The focus is on Agile methods as they relate to delivering BI solutions. If you would like to suggest a discussion topic, an article or link, feel free to send me an email at Copyright 2012 Fri, 07 Jan 2011 07:24:16 -0700 Blame the Tool! In her book, Sudden Death, Rita Mae Brown wrote, "The definition of insanity is doing the same thing over and over again and expecting different results."[1]  Think about this quote for a second and ask yourself, does this quote apply to the way you deliver BI solutions?  Have you been doing the same thing over and over again expecting different results?  If so, then you might want to keep on reading.

I've been in the consulting business for most of my career and have seen too many organizations blame technology for project failures.  In the past couple of years, especially, I've seen scads of companies spending ungodly amounts of money to do nothing more than replace one BI tool with another.  I liken the decision to blame BI tools for lack of user adoption and downright project failure to Brown's quote.  I really have to laugh every time I hear an organization blame a project failure on the BI tool.  When the BI team (too often synonymous with the IT group) finds itself squarely in the middle of major business dissatisfaction and low adoption rates, there seems to be a tendency to place the blame squarely where it doesn't belong, namely the BI tools.

But is it really the tool?  How can these very same tools be in wide use across the globe and be an integral part of some very successful BI deployments?  If it were truly the tools' fault, the vendors selling them would have gone out of business a long time ago and one would expect that hundreds to thousands of successful deployments with a "bad tool" wouldn't be possible.

Psst!  Stop and take a moment to listen to what's really being said:

Company 1 - "We're going to replace Tool A with Tool B because Tool A can't deliver what we need."

Company 2 - "We're going to replace  Tool B with Tool A because our users aren't using Tool B."

Company 3 - "We're going to replace A and Tool B with Tool C because neither A nor B are cutting it."

Oy vey!

In my opinion, most BI tools available today provide much the same functionality - query, reporting, analysis, dashboard and scorecards.  Some are certainly better than others at any point in time, as like most products, they continue to leapfrog each other in features and functions. And for the record, I'm talking about the real BI tools, not basic reporting tools pretending to be BI capable.  So if the tools provide similar functionality and have been successfully deployed many times over, what's really going on?

I liken these "Tool Wars" to another manifestation of poorly executed BI programs - speadmarts.  To me, they're symptoms of the same disease - IT's failure to truly engage with the business to deliver impactful, insightful, meaningful and intuitive solutions.  Seriously, if the BI team were building solutions using business-driven techniques, the solutions couldn't help but be successful.

So, before blaming tools and looking for a replacement, I'd encourage BI groups (IT groups) to first conduct a self assessment.  Reach out to the business community for whom the solution was supposedly built and seek to understand what the real objections are.  I'd be willing to bet donuts to dollars that the solution is suffering from one or more of the following woes:

  • A warehouse was built by IT and "thrown over the fence" to the business community with the expectation that they'll find value in the raw data tables.
  • The "BI" implementation is nothing more than a glorified report building environment rather than a metrics- or objectives-based solution that can be leveraged to drive revenue, improve customer satisfaction, reduce costs and maintain strong financial controls. (see "A Business Context for Agile Business Intelligence")
  • Time to value is measured in quarters and years rather than in weeks and months because existing delivery processes are mired in traditional waterfall methods.
  • Adequate end-user training has not been provided.

In a nutshell, "good" and "bad" have very little to do with the BI tools and lots to do with providing true business value in a reasonable timeframe and at a reasonable cost.  When BI solutions are anchored in and aligned with business objectives, the front-end tool really shouldn't matter.  If you don't agree, try replacing your BI tool while holding everything else constant.  When you deliver the "new and improved" version, don't be surprised when the business looks at you like you're insane!

[1] Rita Mae Brown, Sudden Death (Bantam Books, New York, 1983), p. 68.  Author's note:  I found that this quote has also been attributed to Albert Einstein and Ben Franklin.  However, the earliest published reference I could find was Sudden Death.
]]> Fri, 07 Jan 2011 07:24:16 -0700
The 12 (or so) Myths about Agile BI - Myth 3 Agile BI Methods Do Not Support Sound BI/DW Architecture Principles.]]> Tue, 09 Nov 2010 12:34:31 -0700 The 12 (or so) Myths about Agile BI - Myth 2 There is No Need to Produce Documentation When Using Agile BI Methods.]]> Tue, 09 Nov 2010 12:30:28 -0700 The 12 (or so) Myths about Agile BI - Part 1 I hope you don't mind, but I've opted to create articles rather than blog posts about the 12 myths.  As I began writing about the first two myths I finally had to admit to myself that the word "terse" in no way reflects my writing style.

My goal is to address each myth in a bi-weekly article starting with today's edition - Myth 1:  Agile BI is radical and new.  Whether you're interested in operational business intelligence, business intelligence for banking, business intelligence for retail, or any other type of business intelligence reporting, I remain convinced that Agile BI is the way to deliver the solution.  Let me know what you think.

]]> Tue, 12 Oct 2010 17:21:40 -0700
The Twelve (or so) Myths about Agile BI Agile development has been applied to software development projects for quite some time. In a white paper that was first published by Robert Holler in the May 2006 issue of Better Software Magazine, entitled "Five Myths of Agile Development," Mr. Holler addressed the following myths with respect to software development projects.

  1. Agile Development is Undisciplined
  2. Agile Teams Do Not Plan
  3. Agile Development is Not Predictable
  4. Agile Development Does Not Scale
  5. Agile Development is Just Another Fad

In the past couple of years, we've begun to apply agile methods to business intelligence and data warehousing (BI/DW) projects.  Without a doubt, I've heard many of the same things addressed in Holler's white paper said about Agile BI from a number of sources.  Interestingly, when speaking with folks who are both familiar and unfamiliar with Agile BI, yet another set of misconceptions has arisen.  As a contributing architect in the development of Agile BI methods, I've compiled a list of Agile BI myths. As I sit here writing this blog post I'm not sure if I'll end up with ten, twelve or fifteen.  I encourage you to post your "Myths" here on my blog so we can discuss their validity. Who knows what I'll hear next!  At least for now, my list includes:

  1. Agile BI is radical and new
  2. There is no need to produce documentation when using Agile BI methods
  3. Agile BI methods do not support sound BI/DW architecture principles
  4. Agile BI displaces the software development lifecycle (SDLC) in its entirety
  5. Agile software development methods can be applied without modification to BI/DW projects
  6. When working as a part of an agile BI/DW team, anyone can work on anything, without consideration for roles and skills
  7. Agile BI only works for small teams
  8. Improvement in team velocity is all that matters
  9. Agile BI methods only work well with seasoned, highly talented and experienced team members
  10. All BI/DW tasks can start simultaneously because waterfall methods are no longer used
  11. The BI/DW team must be collocated in order for agile methods to be successful
  12. There is little to no correlation between story point estimating and project management estimating and budgeting

Look for future posts as I dispel each of these myths as we head down "The Path to Agility."

]]> Thu, 30 Sep 2010 09:38:48 -0700
Your Homework Assignment As we begin our dialogue on business and IT agility, I've always found it best to first set the proper context.  In my article, "A Business Context for Agile BI,"  my goal was to set the table for the macro view of Agile BI.  The next context that's important to understand is the foundation of agile methods.  While there are many flavors of agile, most of what I will be writing about is based on the Scrum method.  Having said that, I had a choice to make - should I first attempt to explain Agile/Scrum or should I assume that you are already familiar with it?  Hmm, what to do?  What I've decided is take the middle ground and point you to several great references in order for you learn more about the foundational concepts.  This way, credit is given where due and I can turn my attention to the nuances of the methodology and its application to BI/DW delivery.  In other words, you now have a homework assignment.

Without a doubt, the authors of the Agile Manifesto and the Scrum Alliance get credit for bringing us agile methods and principles in the first place.  Read the Agile Manifesto's 12 Principles of Agile Software as the starting point.  After that, take a tour of the Scrum Alliance web site as it's chock full of great information about Agile/Scrum and includes lots of great articles and where to go for training and how to become a Certified Scrum Master.  As you read through the information on these sites, keep in mind that they were developed specifically for software development projects.

Next, turn your attention to Ralph Hughes' book, Agile Data Warehousing to get a primer on Agile/Scrum and XP as it pertains to building data warehouses.  Hint: If you want to get the crib notes version of Agile/Scrum, read chapter 2.  From there, you can learn more about foundation concepts such as user stories, story points, burn down charts and other key components of Agile/Scrum.

Lastly, I'd encourage you to read Larissa Moss and Shaku Atre's book, Business Intelligence Roadmap.  In particular, Chapter 0, Guide to Development Steps, provides a great overview of the entire software development lifecycle (SDLC) in the context of BI/DW projects.   Why is this important to understand?  Because Agile/Scrum does not displace the entire SDLC, it augments and enhances parts of it.

After doing your homework assignment you should have a better understanding of the basics of Agile/Scrum and the BI/DW SDLC.  So what makes the contents of this channel different?  In a nutshell, the contents of this channel will morph the foundational concepts into a practitioner's view of the world, at times adopting the principles outright, and at other times, challenging them at their core.

As I'm planning writing several articles that explain Agile/Scrum within the broader SDLC (as defined by Moss and Atre) as well as methods for properly estimating projects beginning at Step 1 (Justification) and segueing those estimates into sprint cycles, I hope you'll take the time to get acquainted with the basics.  In doing so, we can start our journey on 'The Path to Agility.'

]]> Fri, 17 Sep 2010 08:55:09 -0700