Friday, June 5, 2015

Yes, You need a Sandbox

I recently ran across a very insightful article By Bob Becker of the Kimball Group. If you have not heard of "Kimball", as in Ralph Kimball, he is arguably one of the best known founding fathers of data warehousing and the data model known as "star schema". Mr. Becker's article caught my attention because it asks and answers the question "Does your organization need an analytical sandbox?".

Take a moment to go read the article here.  Then come back and let's think more about it.


After my first read through the article, my reaction was: "of course!", and then I thought about all of my IBM i clients around the world and wondered if this applies to them, and if so how would it apply. And maybe more importantly, can this idea be simplified to help move an IBM i shop forward towards providing more value sooner than later.

The first observation that caught my eye is this:

" today’s competitive world, organizations need to be more nimble. They want to quickly test new ideas, new hypotheses, new data sources, and new technologies."

This echoes what my team has been saying to executives and technical leaders for many years. You must establish an organization that uses modern tools and modern methods to develop applications that are flexible, extensible, scalable and timely. Employing best practices in the science and art of DB2 data centric design and development is the foundation.

The second observation that caught my attention is this:

"A key objective of the analytic sandbox is to test a variety of hypotheses about data and analytics. Thus, it shouldn’t be a huge surprise that most analytic sandbox projects result in “failure.” That is, the hypothesis doesn’t pan out as expected. This is one of the big advantages of the analytic sandbox. The data utilized in these 'failures' didn’t and won’t need to be run through the rigor expected of data contained in the data warehouse. In this case, failure is its own success; each failure is a step towards finding the right answer."

Folks, this is profound. It is what I refer to as "lose to win".  Think about it. Better yet, do it.

The third observation that resonated with me is this:

"Having the right skills in house is critical to the success of the analytic sandbox. The users of the analytic sandbox need to be able to engage with the data with far fewer rules of engagement than most business users. They are users capable of self-provisioning their required data whether it comes from the data warehouse or not. They are capable of building the analytics and models directly against this data without assistance."

One topic I continue to shout out from my soap box is this: you absolutely need a DB2 for i database engineer.  Better yet, you need a database engineering team. These are the men and women who will have the skills, knowledge and mission to use the sand box effectively and efficiently - with a positive ROI.


During the iAdvocate session I hosted with Jon, Susan and Paul at the Spring 2015 RPG & DB2 Summit conference, we tossed around the idea of using a skunkworks project to learn, innovate and demonstrate value sooner than later. We also discussed the benefits of keeping such a project initially off the radar screen and NOT asking for permission. Once there is a eureka! moment, the team can go into marketing mode and share the results and/or show off the prototype - thus winning support and moving into the realm of official projects.

Granted, what I just said is not necessarily natural to most IT folks, and there is likely some new skills required - but hey, this is what we are here for.

One alarming trend I also shared with the Summit conference audience is that traditional IT organizations are increasingly being bypassed. The business units are embarking on data centric projects without assistance from IT. While this phenomenon is a topic for another time, let's just say that bypassing IT, and the wisdom embodied within, is not likely a good long term strategy. Never-the-less, it is happening. Why? In my opinion, it's because IT is not meeting the real or perceived requirements of the the business units. How about we change that...


So, what's my simple idea?

Build out a small, modern and independent DB2 for i sandbox.

Figure out how to buy, beg, borrow or steal some newer IBM i infrastructure that is outside of the normal control and governance scheme. By the way, this is NOT a development box. I consider designing and developing code a "production" activity critical to the day to day business operations, and as such, the development system is part of production. We don't want to mess with that.

Obtain and install the latest version of DB2 for i (that would be what's inside IBM i 7.2 + TR2), stand up some modern development tooling (IBM Data Studio for example), and add in DB2 WebQuery plus the DataMigrator ETL extension.

Get educated on relational database fundamentals (i.e. data modeling, SQL DDL and DML, set-at-a-time, etc.) and become familiar with DB2 Web Query and basic ETL.

Now, here's the fun part... go buy one of the business users a cup of coffee. Find out what they are doing with the data they are extracting from the production system. You know, the big data set that is pulled down to their PC periodically via QUERY/400 and manipulated with Excel. Learn their process for analyzing and formatting the data. Then go back to your DB2 for i sandbox and experiment. Try to replicate what they are doing.

Fail. Try again. Fail. Try again. Eureka! it works.

Now, enable the information to appear on a tablet. Congratulations, your information is mobile.

Ok, time to get some informal but important feedback...

Go buy the business user another cup of coffee and show off the results of your experiment. If they are excited about your prototype, encourage them to spread the news. If they provide some meaningful suggestions for change, take it and iterate. This will help to strengthen the new relationship and build trust, not to mention giving you more opportunity to learn and practice.

The bottom line is this: be proactive, be a hero!


If you want to discuss the "sandbox" idea further, or if you need help it pulling it off, please let me know.  We are here to help make you successful, and more valuable!

No comments:

Post a Comment