Lessons from optimization and automation

I rolled my eyes when I first saw this automation at a local hawker centre for brewing of the traditional nanyang coffee. It is with good intention and effort to improve efficiency, and possibly consistency with this automation . However, there could be better ways of improving or optimizing. You cannot simply "automate" a process by adding a machine to a human process without due considerations of the goals, operator and working environment. The process was originally for human operator. Adding a machine to a human process may be suboptimal without due considerations. There are also longer term issue like high total cost of ownership ( from maintenance) as well as service competency in this environment. As I reflect, I will share 3 potential lessons , and possibly how we can move on from current situations

Lesson 1: Focus on prioritized goals for optimization

If the goal is efficient ( e.g. average cycle time taken to produce a cup per customer), use of a robotic arm may be suboptimal for a process that is relatively consistent and standardized ( processes of brewing coffee or tea are almost similar). 

A continuous line production approach may be more practical . We can learn that from fast food chain using industry-scale drip coffee brew machine to make coffee for dispensing. It uses a large coffee filter, and has high efficiency output per liter. There is also minimal moving parts for maintenance. Some may argue that the quality of drip coffee may differ from nanyang coffee style of brewing. Nanyang coffee style of brewing will closer to a hybrid of drip coffee brewing with french press via using of filter sock. Again, if the goal is clearer and quality parameters could be set correctly, optimization can be further done. For example, modification could be made to an industrial drip coffee brewing machine so that it holds water in the filter longer. This translates to longer contact between coffee powder and water, and thereby giving a similar french press/nanyang coffee brewing effect to the coffee. Given this context, a time-based calibration of setup should be sufficient. Temperature sensor or other monitoring sensors, albeit at higher cost of maintenance and setup, may be required if there are goals to ensure quality that customer wants,and willing to pay for.

The lesson we have here is that we should identify and prioritize clearly the goals for optimization. Then tune the setup to the goal

Lesson 2: Have awareness of operator and environment

It is important to be aware of potential users and the environment for the machine. In this case, it is a hawker center where you are unlikely to have well-trained machine technicians to be able to do frequent and proper equipment maintenance. Having a complicated machine, like in this content, will incur higher cost of maintenance in the long run when it is not properly maintained. In addition, there may also be long downtime due inherent competency gap in the likely operator and nature of the environment.

The lesson we have here is that we should study the potential operator and environment before implementing any optimization. In this case, a simple to maintain and with redundancy options system for brewing coffee may be more practical.

Lesson 3: Understand strength of solution and it's relevancy to operation goals

Flexible arm solutions like this has been more accessible due to reduced cost over the years. However, it may not be the most efficient way of working for consistent and standardized operation with high output like coffee brewing in hawker centee. It is good for operations that has a lot of variability and changeover. In production, I have seen it used in context where there is a large variety of products with different setup used. The flexible arm allows flexible and consistent changes to production setup.. In this case, the flexible arm may be used to make individual customized cup of coffee with special requirements ( type of milk, amount of sugar etc). This is provided that customers are willing to pay for such customization.

The lesson we have here is that we should always look at the strength of every solution, and see if it is relevant to the goal of this particular optimization. The flexibility of this flexible arm solution may not be that relevant for coffee brewing in a hawker center environment

How can we move forward?

Interestingly, the situation of suboptimal optimization is not totally new.to just a context of automation in hawker centre. I am sure many may have seen situations where automation or machines are purchased with good intention but realized that they are not relevant. 

There are likely three choices. Firstly, it is to continue the suboptimal process and machine. Accept the issues with it. Secondly, it is to obsolete the suboptimal process immediately, and revert to something more practical. Thirdly, it is to think of a new use for it that brings new value to customers. Typically, most will adopt the first and second options in varying degrees due to sunk cost effects of equipment investment. 

A better idea might be to evaluate objectively the pros and cons of each choices in relation to the operation goals ( efficiency , short term and long term cost, customer experience). There may be hidden delighter needs that could be fulfilled by a machine that the customer is willing to pay for (Option 3). Out of the box, I can imagine using this flexible arm to perform the creation of customized specialty coffee much like how a Turkish ice cream seller or how western coffee barista performs. If all else fails, take a hard decision to choose option 2 (obsolete the machine) and use the space/machine for better value generation.

Automation and optimization are never a simple action of just doing it. Good intentions may not necessarily yield good results. You need to put your heart and mind into continuous improvement and refinement. Hope that you find the shared three lessons as well as lessons forward useful. Always open to hear your opinions and thoughts.

Written and copyright of Alex Goh