Graham just wrote a post on his new blog about the difference between Business Process Automation and Business Process Improvement. Both terms are literally self-explanatory, at least at a shallow level; Graham explains them in more detail. Specifically, he claims that you should automate (BPA) before improving (BPM) your business processes.
BPA vs. BPM is very much the same debate as efficiency vs. effectiveness. Efficiency means doing something with low costs; effectiveness means doing something with high results. Of course, both are important, but I think that effectiveness trumps efficiency every time: you’ll never reach your goal if you set out in the wrong direction, regardless of how fast you can travel.
You can also see this principle in terms of software. BPA is sort of like performance optimization: you’re trying to take the existing processes and make them work faster, but largely in the same way. But all the tweaks in the world won’t save an algorithm that’s fundamentally inadequate. Replace that algorithm and you may get several orders of magnitude more benefit than the tweaks brought you. Quicksort vs. Bubble sort is the textbook example of this.
Graham claims that automations are a much faster/simpler change to make (relative to more strategic shifts) and thus provide quicker, more realizable performance gains. That certainly can be true in some cases, and in some organizations that may be all that’s possible. However, I don’t think that’s always true. Automating some processes can be extremely difficult and time consuming; in fact, it may even be impossible given current technology or impractical due to the costs. Indeed, improving the process may mean altering it so that it can be more easily automated.
In the end, automating business processes before managing them may be wasted effort if those same processes are going to be replaced.