May 2002

Ridiculous Requirements. Meaningless Metrics. And that's just the fun part!

The 20 members of the North Jersey SPIN who attended the May 15 meeting were given an interesting challenge by presenter Bob Benjamin: forget everything you thought you knew about software engineering and start over. And don't be surprised if the market for technology skills starts to look fundamentally different from what we have been used to.

Ok, that might be overstating things a little bit. But just a little.

After beginning the evening with a brief discussion of what makes all too many IT projects go "BONK", Mr. Benjamin proceeded to turn some of the CMM commandments on their head. Specifically, Mr. Benjamin declared that requirements management (as it is currently implemented) doesn't work at many organizations, and collecting metrics often serves no purpose. This is part of a phenomenon that Mr. Benjamin calls the "Long Fuse Failure," where projects that are completed on time, on budget, and with zero defects still manage to provide no measurable improvements to the business.

As the initial shock of these pronouncements began to sink in, Mr. Benjamin observed that "the fundamental problem is that the IT department is brought in after the business case has been made, accepted, and is assumed to be correct," even though there are no business reasons for doing the things that IT departments are often asked to do. Weak business cases with little counter evidence and no QA pressure (that is, evidence of benefits from multiple sources) often lead to projects that have the wrong focus. Gathering requirements for such projects at this stage is a waste of time because it is already too late. Why? Becasue the requirements can not be based on any kind of portfolio management at this stage, and risk assessments have been entirely bypassed.

The perscription for curing Long Fuse Failure Syndrome includes closely tying requirements to business goals. It also includes tying metrics to "Portfolio Management." This means that instead of focussing exclusively on project metrics (time, size, cost, effort, and defects), Information Systems teams should also consider "balance sheet metrics" (liquidity, leverage, financial performance ratios), "operational metrics" (trends in cycle time, customer retention, market share, etc), and the "Five Driving Forces" metrics (supplier, buyer, substitution, competitive rivalry, and barriers to new entrants).

During this presentation, some NJSPIN members speculated that part of the problem might be that Information Systems personnel are not accustomerd to "standing up to business people when they propose questionable projects." A perceived second- class status for Information Systems (and manufacturing) contrasted sharply with higher-profile positions in sales, marketing, and finance.

The topic of outsourcing contracts (especially cost plus contracts) and their role in Long Fuse Failure situations also came up during the discussion.

Mr. Benjamin wrapped up the presentation by offering the web address for the Atlantic Systems Guild's online requirements management template.

To view the complete presentation, click on the hyperlink below.

May 2002 Program: Pre-Defining Success

Business Notes:
Important Announcement!!
Next month we are changing the venue of the meeting to the office of Telcordia Technologies. Details will be mailed to all NJSPIN members, along with a more detailed description of the program for the month. If you are not sure if you are on the NJSPIN mailing list, contact John Dworak.

-Next month's program: Jim Heil will do a presentation on software testing and software test plans.