I must admit that I am writing this blog entry for the second time (the original one disappeared somewhere on my hard disk). The d?j? vu of a rewrite seems to remove any passion - I hope I do not bore you to tears. Here goes anyway.
I said in the original phantom blog that I am trained in operations research and that this unfortunately tempts me to consider every thing to be a variant of an optimization problem. And, yep, I do cast business analysis the same way. A typical optimization problem has an "objective function" and a "constraint set". The goal of optimization is to maximize the value of the "objective function" given the "constraints". Business analysis is the art of achieving enterprises goals (objectives) given its limitations (constraints). The business analyst tries to find the best solution - but I suspect that often times they get as far as finding a reasonable feasible solution - finding the best is almost impossible.
You should naturally complain that this simplification of what a business analyst does is unfair. So let me elaborate on three weaknesses of this viewpoint.
First, the "business" constraint set is outstandingly difficult. It is defined by existing customer commitments, regulation, marketing "must haves", legacy systems that will break if you look at them the wrong way, board level strategic mandates and, most of all, the way people are used to working - the ingrained business practices that every organization feels is essential to who they are. But - understanding these is why it is called business "analysis" - it is the raison d'?tre of having business analysts. It is why I do not see analysts and Management Consultants to be all that different.
Second, this optimization point of view ignores the dynamic nature of business. You can say optimization assumes a statically defined objective function and a constant constraint set. But in my defense - optimization does present another view - it asks the questions: Which constraints really bind? Which constraints - if loosened - would have the greatest impact in allowing the business objectives to be even better met?
I see this as the crux of business analysis. When a business analyst can say "just change how interactive voice response works and we will be able to reach out to 50% more customers in debt" (in other words defining IVR based payments as the critical constraint), then real value is being provided. But - and this is a huge "but" - the fact that a business analyst can point to a constraint as needing to be loosened is not enough. Every constraint has a vested interest - someone who owns the process / practice. A business analyst, without sponsorship from whoever has defined the "objective function" (the CFO, CMO, CEO), cannot fight these interests. Vested interests have years of experience and domain experience. Being right is not enough to win the argument.
There is no doubt that Enterprise software is a vested interest. I will leave the question open whether vendors see their software as such, and whether they are aligned with the need to make sure they are not being a primary "binding constraint". All I will say is that embracing SOA is not enough.
Third, optimization is ultimately about how problems are solved. It is essentially applied maths. What can it tell business analysts? Linear optimization problems (as used in the production planning) have beautiful algorithms that get answers quickly. Integer programs (as used in allocating the best people to the best jobs) take pure brute force. Some problems (especially in logistics) succumb to heuristics. But it is the non-linear problems that are the real bear. They are all about the art of finding solutions and iterating around them to see if a better solution exists.
I think business problems are very non-linear. And I think that art of solving business problems is getting the "not quite the right" solution and being able to change course quickly and cost effectively. I think this is the trickiest part of business analysis - putting their enterprises in a position to make mistakes "on purpose" - making sure technology and business process decisions are not written in stone.
I think Enterprise software should be more about how problems can be solved, then resolved? and less about "here is the solution". I think business analysts hardest job is making this dynamic measurable. How can the TMF help? And how can we (vendors and users alike) help the TMF.