The End of the Swim Lane?

posted by WMC

Over the past few posts on Pulse, I’ve highlighted a number of business scenarios where you can use network analysis techniques to derive insights you just can’t obtain from other methods.  Here’s another interesting application – Business Process Analysis.

Below is our typical view today – a swim lane of a business process flow, but if you wanted to know what were the critical steps in this process to make it as efficient as possible, you’d have to develop and run a process simulator such as iGrafix; a non-trivial task as you have to add in process steps times and throughputs to every step in order to generate the overall simulation and see where the blockages occur.  And then it dawned on me that this is just another network graph – after all a network graph is anything where we can describe a relationship, or link, between nodes and we can certainly do that for business processes.  So I grabbed a swim lane diagram from a recent piece of work where I was very familiar with the process – here it is in anonymized form.

Pretty typical and if I asked you to guess which were the critical steps, this view suggests to me step #2 and step #16.  But what would happen if I graphed it and analyzed it as a network?  That was an intriguing thought. So I built the load file, plugged it into the analysis software and analyzed the network for both the “communities” present, and also for the parts of the network that control information flow. Then I generated the visualization and scaled the nodes so that the larger the node, the more control that process step has on overall information flow.  And then for good measure I analyzed it all again for community structure; here’s the graph.

The first insight was the community structure correctly grouped related tasks together (example #1,#2,#3 & #4) and by comparing this to the swim lane, I can see that activity is spread over 3 roles…hmm, probably way more efficient to combine those steps into a single  role – I didn’t expect that.  The second insight concerned the 2 process steps that exhibited most control over information flow, #9 and #10, again not what you might expect if you guessed #2 and #16, but I know from this project that #9 and #10 are in fact the key steps which slow down the whole business process and addressing that had been part of the recommendations of the original study.  The interesting thing here was I had derived this insight very, very much quicker than when I’d done the original project.  I’ve since tested this approach on a few other swim lanes I know very well and so far the results are consistent with what we see above; quite fascinating to my mind.

So is this the end of the swim lane?  I don’t think so, swim lanes are great to help people understand their role and their role boundaries, but I do think this technique can very quickly direct you to parts of the overall process flow that will derive most operational benefit if you simplify and optimize them.  That is certainly worth knowing, especially in these days of scarce resources where prioritizing efforts to derive greatest benefit is something we all have to do, but are often forced to rely soley on experience and, to be honest, guesswork.  No need to do that anymore.

Thanks for reading.


You can find Mike’s original article on LinkedIn here.


The opinions represented here do not necessarily represent WMC’s views as a whole.