Pave the cowpath - good and bad

'Pave the cowpath' is a term commonly used in workflows and BPM to indicate a rapid implementation that takes the current flow of work, automating it as-is. The cowpath (or goat-track) exists because its a commonly trodden path followed by workers to get the job done, some based on rules (there's a fence in the way and a gate to go through) and some based on convention (a well trodden track is easier to follow than blazing a new course through the long grass).

Yesterday, @dhinchcliffe 'tweeted' about a bulletin for the American Society for Information Science and Technology, which mentions the 'cowpath' in the context of social websites:

The Information Architecture of Social Experience Design: Five Principles, Five Anti-Patterns and 96 Patterns (in Three Buckets) by Christian Crumlish (curator of the Yahoo! Design Pattern Library).

This article presents paving the cowpath as the first of five patterns for the design of social websites and applications that are trying to put social elements on top of them. Crumlish's insights apply to workflow and BPM as well in my opinion:

[...]study some of your potential customers. How do they do what they do today? Yes, of course, the thing you want them to do will be better, but is it really entirely different? Can you offer people a way to continue doing most of the things they’re comfortable doing today as you introduce new possibilities into their lives, or are you really going to insist on them changing everything at once?

I know that this goes against the philosophy of process reengineering and methodologies such as six-sigma, since in paving the cowpath you are doing little to remove the wasteful activities in the process. Slapping a tool on to help move work around does not necessarily enhance the underlying process, but it does make it quicker and reduce wasted time while people wait for work to get to them (which is not necessarily a bad thing). In not removing waste or adding more rules, according to Crumlish:

Often the impulse is to stamp out these rogue behaviors and enforce draconian rules requiring only the behaviors you had planned for. This course of action really only makes sense if the behaviors you are trying to stamp out are truly destructive or evil. There are many anecdotes about thriving social sites that killed themselves off by legislating against fun and forcing their users into exile to find the activities they had been improvising “incorrectly” in the site they had to leave.

I also have many anecdotes about business workflows that were overly restrictive and prevented knowledge workers using their brains (1, 2, oh and the one I never wrote about 'cos its not a flattering story) that led to a similar effect: the application failed and was not adopted.

The great thing about paving the cowpath is that you can implement quickly, assuming you don't try and automate some of the completely ridiculous manual things that are done today, and just let people realize they are ridiculous and stop doing them in their own time. With the right tool, you can also play into the second half of Crumlish's cowpath definition:

A better plan is to support the behaviors your users are engaged in. Let your users tell you what the best and highest use of your interface may turn out to be. Don’t be so arrogant as to assume you know everything about how the social dynamics you’ve unleashed need to evolve.

In a business context this means that you, the business analyst or process analyst does not have to sit for 8 hours a day doing the work of the potential workflow users. Analysis can be done quickly and a flexible system implemented equally quickly. And do not assume you know best when it comes to the way workers must interact or handle exceptions to the rules.

Given some time and professional assistance from a process analyst and workflow tool, users can improve their own processes. The process analyst will need to be strong, since 'Management' will always have the desire to put in more rules, not take them out. Go ahead, pave the cowpath, and give the workers a chance to see how a new system might improve their operations. With a new tool in place (e.g. a lawnmower), users could be presented the freedom to forge a completely new cowpath that works better for them. It would be a shame to prevent user innovation through initially overdesigning a new workflow.

A post from the Improving It blog

Download the podcast of this blog post