Getting the right answer is important. But it is impossible to get the right answer unless we ask the right question. Time and effort invested in better questions pays off in better answers.
There used to be a zebra crossing across the main road. Pedestrians could get across easily and motorists were discouraged from using the side roads, because they had to wait for a gap in the traffic before being able to turn into the main road. Then the zebra was replaced with traffic lights. Pedestrians have to wait much longer for their turn to cross and it’s much easier to turn out of the side road. More people use the side road as a short cut, accelerating towards the lights to squeeze through on green. So then the side road needed traffic calming measures, with humps and layout changes to slow the traffic down. That hasn’t worked terribly well, so now new and humpier humps are being proposed. But that’s focusing on the wrong part of the problem. Getting rid of the traffic lights would be a far more effective solution.
There is a security door which is opened by a pass card. The card reader is on the same side as the hinge, the door is quite large and opens outwards automatically. If you are standing in the right place to reach the card reader, you are standing in the right place to block the door as it opens. The door opening device has a sensor to stop the door crashing into people. So if it detects somebody in its way as it is opening, it reverses and closes again. It is common to see people dancing around to be able both to open the door and get through it before it decides to close again (their confusion demonstrating a singular lack of affordance). The sensor certainly stops people being hit. But that’s focusing on the wrong part of the problem. Moving the card reader to the opposite wall would be a far more effective solution.
There is an online transaction which because of its overall structure sometimes asks for the same information at different points in the process. That introduces a risk of inconsistency, so there are rules in the system which test whether the same answers are given to the same questions and flags an error if there are any discrepancies. The inconsistency risk is managed effectively. But that’s focusing on the wrong part of the problem. Streamlining the process to ask for each piece of information once would be a far more effective solution.
Once you start looking, it’s surprisingly easy to find problem which exist only because they are the solution to an earlier problem. It’s terrifyingly easy to get trapped in a framing of the question which focuses attention on the immediate symptoms rather than on what caused the symptoms to be generated in the first place.
This shouldn’t be hard. An analysis technique as simple as the five whys would help with each of those examples, but all too often not even that level of structure is applied. Too often we focus on the quality of the answers to questions. Too often we would do better to focus on the quality of the questions we are trying to answer.
This is such a fundamental point that we miss every day. In addition to the five why’s there’s a need for the discipline to look at a problem within the context of the whole system. Local optimization is a big culprit for even creating systemmic issues while making one little silo look better (on paper at least, and only when you look at it in isolation).
Great post, I look forward to more!
Local optimisation is a good way of summarising a lot of this – where local may be by time or scope as well as by place or by organisation.
I fully concur! It is human nature to jump to solutions before understanding the problem. In fact, your first example is an example. The signal was installed because there was a problem with the intersection, it probably had nothing to do with the pedestrian crossing, it did not “replace” the zebra crossing. So, what was the problem? Why was the signal installed? If we understood the problem, we could identify better solutions (being rhetorical to make the point).