
- Process Control/Improvement
- Design of Experiment
- Analytics
Process Control/Improvement (PI)
I recently went to a mixer at the University of Washington for current ChE grad students to meet with alumni in industry. People were surprised I made the transition to business and especially when I mentioned that my ChE thesis was in process control. "Oh god," the reaction would typically begin, "All I remember is all the Fourier & Laplace Transforms and I lost interest." This was disappointing, as Process Control/Improvement is one of the most useful concepts I learned in ChE and one that I've been able to apply time and time again in business. A central Buddhism tenant is that suffering comes from struggling with what is (article). Much of our lives are spent trying to enact change, and process control (in practice) lays out the perfect framework:What you hope to effect = what you have control over + what you don't + what you don't even know
or
Controlled Variable (CV) = what you hope to effect
Manipulated Variable (MV) = what you have control over
Disturbance Variable (DV) = what you don't have control over
CV = f(MV,DV) + unknown
Process Control is really about improving what you have control over and understanding what you don't have control over, given your resources. Remember our AT&T case from last time? Customers were leaving due to poor customer service. Better plans didn't affect customer retention, so it would have a negligible effect on churn.Design of Experiment (DOE)
Before tackling any project, it's worth using Design of Experiment to validate: "Do the resources and levers I have at my disposal have any effect on changing my Key Performance Indicators (KPIs)?" Let's say I want to increase revenue for my wholesale distribution system. Using the marketing material balance:?Customersin - ?Customersout = ?CustomersAccumulated
My customer retention could be better, so I've asked a multidisciplinary team to brainstorm on solutions:- Twitter / Facebook campaign
- Account management practice
- Quality control (on-time & error-free delivery rates)
- Product offerings
- Etc.
- Effectiveness
- Cost
- Time to implement
Healthcare Example
Typical CVs & MVs in healthcare are patient outcomes and revenue (CVs) to staff and patient beds (MVs). If you hit a bottleneck, hire or build. In an increasing competitive and cash strapped environment, these are fewer available MVs (article). Process control teaches us that MVs can number 1 to n with varying degrees of bang for your buck. Creativity, innovation, "boot strapping" (a favorite venture capitalist & entrepreneur term) can come from asking the question"What can I do with what I have?" or "What MVs do I have at my disposal?" Your CVs become your strategic objectives and your MVs are the resources that you can bring to bear that can actually make a difference in your CVs.Analytics
Most engineers in business school find that if they could understand systems of differential equations in undergrad, business analytics should be quite easy by comparison. The analysis techniques I most frequently use for my clients in industry come from the following core topics:- Statistics [is this data meaningful?]
- Multivariate analysis [how do I use data to predict behavior?]
- Time-series analysis [how does behavior change with time?]
- Optimization [how do I maximize KPIs given constraints?]
When have you used your ChE background as a corporate mechanic?
Photo: Robert J.Pennington, www.rhizomeimages.com (C)2011 Arkan Kayihan, used with permission
Comments
- Log in to post comments