Automation? Isn’t this lesson supposed to be about how “lean” starts at home? Think again. Unless you are working at a major car manufacturer, the home is actually one of the most heavily automated places you will ever be. Washing machines, dryers, dishwashers, coffee machines, toasters, and blenders have unlocked tremendous productivity in society and are doing their jobs so well that we hardly think about them at all.
Our cupboards, basements and garages are also full of automation equipment that we thought would have similar impact as the items above, but then turned out to be not worth the trouble. Either these machines take too long to setup, take too long to clean, are stored too far away, break down too often, or simply don’t work as advertised. This is not any different in a manufacturing plant with the difference that those machines cost hundred or thousand times as much, making following a gut reflex to get one very costly.
Back to the dishwasher. It is one of the successful devices as the labor savings are tremendous. Once the machine has been loaded, it can be set and forgotten. Driven by a highly competitive industry for multiple decades, maintenance and repairs are now minimal. It’s only drawback is that it can disrupt “flow” and acts as a bottleneck, for example during a busy Thanksgiving dinner. Even that does not matter, however, as its predecessor, the sink, is still around to step in if need be. This is often not the case in manufacturing where a new machine usually replaces the old process, which is then unavailable in case of malfunction or increased workload.
Unfortunately, not all machines are similarly well received like the dishwasher. Sometimes, simply changing the battery of a device is perceived too much hassle, putting an end to a device’s life. More often than not, it turns out that a “more efficient” machine has a cheaper per-item cost as a human worker, but disrupts flow to a point where a manufacturer discards it. This can be due to limits in what the machine can handle, wait times that are unacceptable when there is rush, or malfunctions that preferably happen when the machine is needed most.
Even when provided with an optimal trade-off between cost, functionality, and flow that has been proven valuable for decades, it is tempting to further increase automation. Why not constructing a robot that unloads the dishwasher? How would such a robot look like? It would require an arm long enough to reach everywhere into the dishwasher, a mobile platform to reach the shelves, and a lifting system to reach the different shelves. As all of this sounds very expensive, the first step in designing such a robot would be to analyze the task as a whole to find ways to get by with a simpler solution. Not surprisingly, many of these improvements are very, very similar to those that we found using the lean method.
For example, an automation specialist will quickly advice to place all cupboards close enough to the dishwasher so that a single robot arm can do the trick. Although you haven’t been ready for this improvement thusfar, you finally give in to this structural change to be able to get the benefits of a robotic solution. The overall value that this robot provides is now calculated as follows
savings obtained by robot = savings from lean improvements + savings from robot doing the labor
The first insight here is that the first half of the equation can often be achieved without buying anything. The second observation is that doing two steps at once is usually not a good idea (see also “kata” in the previous lesson). First, we will need to make sure the lean “improvements” will actually not create problems down the line. Second, it is unclear whether the robot fully replaces the human, possibly disrupting flow. The smarter approach is therefore to first optimize the system as much as possible, then specify what the robot should be able to do, how fast it should be, and look for appropriate solutions.
This is not to say automation or robots are bad. Robots can also increase flow. For example, a robot that is waiting for a picker in a warehouse (and not the other way round) might allow for designing a system in which everyone does what they can do best: the human retrieves items, the robot transports. A robot might also load and unload a machine and thereby increase flow. Such a machine possibly saves money two ways: a worker that arrives too early to tend to a non-automatic machine needs to wait, whereas a worker that arrives too late disrupts flow. These savings are set off by the time required to set the system up and system malfunctions. They are non-existing when parts need to be made that the automated machine cannot handle at all. In sum, the resulting savings might not be sufficient to justify the purchase. Depending on what the goal is, the manufacturer might be better off by simply improving the way the machine is monitored (see also “Industry 4.0“) and adding a pull system that notifies workers to tend to it.
Increasing the level of automation therefore follows the exact same rules as the lean process: define value, identify the value chain, increase flow, introduce a pull system, pursue perfection. A machine that passes this process is very likely a great investment.