You may have noticed that one theme in this blog is change management, which I think must be central to a BA’s approach to managing requirements. Actually, I take that back…. To some degree, it seems that there is an overemphasis on the requirements themselves, at the expense of managing the business case for satisfying those requirements.

I will try to state my concern philosophically first, by asking the question “Where does the BA stand? What ought to be the BA’s primary point of view?” Here are the stands I see taken by BAs, to varying degrees: 

• The requirements themselves – extremely challenging to set one’s bearings on the requirements themselves, because they will change! 
• The requirements process – obviously a more stable view, giving one the prism through which to view and manage the constantly changing requirements 
• The business case itself – this is the most advantageous point of view from which the BA should start his or her day

Saying it another way, if my mission is to manage the requirements, I could lose sight of the business case itself, which will change as requirements and solution elements (scope, schedule, cost) change. But if my mission is to keep the business case alive and fresh, by continually reflecting within the business case, in business case terms (benefits, costs, risks), the changes to requirements and solution elements, I have not only (a) taken a stand that puts me closest to my stakeholders from a business benefit point of view but also (b) made change management my core activity, because the business case cannot be kept fresh without it.

Do you agree? How central is continuous business case management to the way you work? If it is not central, what are the obstacles/challenges you face in making it so?

Congratulations to John Dean for his analysis of what to do with a nearly infinite list of specific identity transactions (see March 2nd blog for the list, and John’s April blog for his ideas). 

Improve on it if you can, but I like his idea of grouping identity transactions by the “certainty” required, and using known technologies (Human Chip Insertion, Fingerprints, DNA, Retina prints, id cards) to “produce” the needed certainty.  Each “identifier” stakeholder can simply select the level of certainty they require from such a system.  This is a huge simplification, and seems to be a doable analysis.  There is evidence for the accuracy of many tests, AND we will have to consider the usual 4-D time/space complications (DNA is reliable, and is easy to plant at a crime scene).

Condemnation upon John for his focus on the “identifier” stakeholder, instead of “we the identified”, who live on the other side of such transactions (mostly). 

Leaving out the identified in these analyses is (increasingly) the cause of problems like identity theft, loss of privacy, deprivation of due process (sure, you can fix your erroneous credit record – have fun with your new full-time hobby), erosion of civil rights, and much, much more (the teen suicide who was berated on-line by the adult mother of another teen?).

Leaving us out of the requirements gives John access to a seemingly EASY solution – “chip insertion”.  The risk of crooks “stealing” someone else’s chip for illegal purposes can be dealt with by rigging the chips to self-destruct if removed or fiddled with.  I leave the problem of “counterfeit” chips for a later discussion (solution, as usual, is premature,so nyah, nyah, nyah, Mr. Dean).

This kind of solution is especially tempting for management, in spite of being clearly odious to the “workers. 

The day is coming where, if you are a convicted drug user, or a crummy boyfriend, or a whistleblower, or just a cranky person, people will be able to single you out electronically – even treat you the way we treat convicted pedophiles, just by accessing your history based on your chip id. 

Convicted pedophiles and other sexual criminals already know what this is like – no chance to really start over, always labeled.  The process created by the “identifiers” is great for labeling one for life; not so great for allowing for change and forgiveness. 

The current system protects the interests of the “identifier” stakeholders (John, language is important – can we just call them fascists?).  It assumes recidivism, and does not measure rehabilitation.  Remember – Civil Liberties are not just for people that you like.

One can claim that measuring rehabilitation is a different problem from identification.  Unlike John Dean, my perceptive readers immediately realize that if we can’t measure rehabilitation, then crime, or any “unapproved behavior” becomes a life sentence in the system John advocates.  Do you really want the letter L tattooed on your forehead for life, just because you once littered?

We’ll deal with such “identifier” requirements next time.  In the meantime, as a proof of concept for John’s analysis approach, would you agree with the following estimates of how much certainty is currently required for the following identity transactions?.

Amount of certitude required to identify:

1. A terrorist = 0.1% (if we harass 1,000 people to actually catch one terrorist, we are satisfied – witness Guantanamo).
2. Oneself for a driver’s license = 10% (or commerce will shut down – i.e., it is easy to fake the id required)
3. Oneself to buy coffee and donuts = 95.0% (5% of transactions are “fraudulent”?)
4. A death penalty convict = 70% (30% of death row inmates may not belong there, the states’ “management” are not investigating their own mistakes, of course). 
5. A safe sex partner (varies by individual from 99.5% for those who want to witness your HIV test in person, no cheating, to 0% for Larry Craig, Senator from Idaho, who would prefer to know nothing about anyone next to him in a bathroom stall).

I finish with an important success point – small teams.  I want to make the point that our two person “team” (John Dean and myself) is (so far) enormously productive.

In perhaps 20 total person hours of actual focused, asynchronous but responsive work, we have enough grip on the scope to already despair of the sheer scope of the problem – this is good progress!

Next month (unless John tees me off again) we will contrast John’s “certainty” requirements with the requirements that “we the identified” might want to add.

Stay tuned – if we get the requirements right, it may only take one generation to finish this, and none to soon!