While most of my operations classes at Stanford focused on theory, such as around production systems design and supply chain management, this was a general course on operations management -- putting into practice through cases various higher level operations analysis techniques.
A lot of this wasn't new to me, but it was nice to see it in a different context and more from the perspective of a general manager. The objective of the class was to teach us how to use operations in an organization as a competitive "weapon" and not as a burden or friction to minimize.
Below are my biggest takeaways from the course.
- Companies fulfill their business strategy through operations. A company's goal is to make money, and it does this through its value proposition in one or more of the following areas: price, quality, time, and variety. The mix of these four determines the operations strategy that best aligns with the business strategy. For example, if variety and quality are key to producing some very customized product, a job shop manufacturing process with flexible, independent cells and scheduling would make sense, whereas if price and time are more important, such as in producing some commodity good, a continuous flow or assembly line process is a better fit.
- Manage by the bottleneck. We studied various quantitative techniques of process analysis and read The Goal to learn that the first area of attack when aiming to improve a system is the bottleneck. An hour lost on the bottleneck is an hour lost for the entire plant, whereas an hour lost on a non-bottleneck is a mirage.
- The bottleneck should be the most expensive resource. Otherwise, if the bottleneck is a cheap resource, it would be silly not to purchase additional copies of the resource until the point where it is no longer the bottleneck. We learned that at Amazon the bottleneck is an extremely large and complex sorting and box-filling robot; this is its most expensive piece of machinery, and the company does what it can to make sure it is operating at full capacity to meet demand.
- Don't optimize for efficiency, but for your goal (profit). Looking at localized efficiency (like the utilizations of labor or specific machines) is what many people do, and this leads to inferior global performance. As long as the bottleneck is at maximum utilization and running at a pace that matches sales, it is irrelevant how much of your labor or other machines you're utilizing (to a certain degree). You should make decisions that maximize profit, not ones that make your plant look "efficient" in as many areas as possible.
This is a similar lesson to what I learned in my Thinking on Your Feet class this quarter, which I'll post about later: the difference between productive and effective thinking. You can be very productive and efficient in taking care of lots of tasks or brainstorming lots of ideas, but if you're just doing that to be efficient with your time rather than actively working towards a concrete goal, this is often time and energy that's wasted. This now reminds me of the w4w lesson from The 4-Hour Workweek: "work for work's sake" as opposed to work that's globally required and productive.
- There are many non-manufacturing (such as service) applications for the Newsvendor model. This is a simple model we covered in class that allows you to decide how much inventory or capacity to build up in order to deal with random demand and differing underage and overage costs. I thought it was neat how we applied the model to service industries (like hotel and flight overbooking) and even to web advertising inventory planning.
- The psychology of waiting lines: We read some research that presented some common findings about lines that I've definitely thought about or personally experienced before.
-Unoccupied time feels longer than occupied time.
-Pre-process waits feel longer than in-process waits.
-Anxiety makes waits seem longer.
-Uncertain waits are longer than known, finite waits.
-Unexplained waits are longer than explained waits.
-Unfair waits are longer than equitable waits.
-The more valuable the service, the longer I will wait.
-Solo waiting feels longer than group waiting.
In class, we studied ways that queue waiting times can be minimized (analyzing various queue structures under Markov/Poisson/Exponential assumptions), and when they can't, there are ways to use the psychology results above to make waits "seem" shorter (providing entertainment, making the wait seem like it's an in-process wait, explaining waits, etc.).
- When redesigning queue layouts, minimizing actual service time by employing better servers/employees is more effective than employing multiple servers. In other words, one server who is twice as fast is better than two servers at half the speed. I thought this was an interesting result and shows how training and staff selection can be more important than simply hiring more people.