Bill Gates once made the comment that applying a technical solution to a broken process just makes it worse; while applying technology to working processes makes them more efficient. After hundreds of technical solutions implementations I can say with certainty that these are true words of wisdom. Our most challenging projects (the ones that have often gone poorly) can generally be identified by an attempt to implement a technical system where the human work process is not yet clearly defined. On the flip side our best projects have come from improving an existing business process by automating already decently defined human activity.
As our business has evolved we’ve added processes to better flesh out process before we design a technical solution, such as business process modeling (BPM). While we’ve seen improvements as a result, many challenges still exist. I believe the reason for this is that business system end users have a very difficult time imagining how they work so any attempt to invent process through visual or written description will end up playing out differently than expected when actually put to practice.
The Pizza Game
While still in the corporate world I participated in a training exercise sponsored by our division VP that I still reflect on often today. The basic idea is that two teams are formed to operate a pizza production system with what seem like impossible requirements for speed and accuracy at first. After a few iterations though, not only are the requirements met but even more stringent and complex needs can be met. I often find myself asking why we can’t do the same thing with our own business or the clients we serve on a regular basis … usually the rationale is that our process is much more complex or that we can abstract it to something like making paper pizzas.
Real World Games
Several months back we identified that the top challenge preventing operational growth for our company boiled down to operations scheduling. So… as a tech company we decided that we needed the right tool. Since we develop technical tools it was clear the answer was to develop our own tool. After several design iterations of possible technical solutions we realized we needed more process definition in order to build the tool, so we initiated a BPM process which went through a few more iterations until it looked great on paper. The immediate and obvious outcome was that we had simply uncovered more questions than answers and had designed something we all knew would not work well.
Rather than build the solution we decided to try another approach… we allocated a resource to “try” the process manually using limited technical tools such as emails, phone, chat and spreadsheets but to try to follow the process. Immediately, it became clear that the BPM served well as a guide but needed refinement to work in reality. We also found that much of the process could be supported with existing systems and tools without requiring a lot of technical work. We also found that as some of the basic requirements were achieved through a low cost implementation, the need for a bigger and more sophisticated process emerged that would not have been visible without operating the more granular activities.
Automating business process is a great way to scale a business and support growth but automation needs to be focused on already well-defined processes to succeed. Invention of new processes or automation of broken processes needs to begin with simple exercises to define and test execution in low cost, low overhead ways before investment in technical solutions should commence. Some ideas for how this can be achieved are:
- Create simplified BPM flows using a tool like Intalio|Create (www.intalio.com) and test them with human handoffs before automating them.
- Walk through possible workflows as a story to see if they hold up (i.e. first I do x, then you do y).
- Consider strategies to simulate or play act process to see the results and try different approaches.
- Identify the lowest cost method to get to an initial working process, even if it’s not a good long term solution and then refine based on the experiential outcomes.
- Use a task based BPM tool and simulate all process steps as human activities and then gradually introduce automation.
Want to see what a BPMN workflow looks like? Check out Meghan’s Bairat’s blog on BPMN here.