Ever heard of the Gimli Glider? It was an Air Canada flight back in 1983 that, get this, ran out of fuel. That’s right, a 767 with 61 passengers simply ran out of fuel. How did this happen? And what does this have to do with UX Research? Stick with me on this, because it’s pretty cool.
Air Canada Flight 143 was a flight route from Montreal to Edmonton with a stop in Ottawa. The aircraft, a Boeing 767, was relatively new, and one of the two fuel gauges tended to fail. Normally not a big problem, since aircraft like this have built in redundancies (in this case, a second fuel gauge). The real problem though, was that the second fuel gauge tended to read incorrectly when the first failed. Not great, but there was yet another redundancy. Ground crew, prior to take-off, would take a dripstick measurement. Next was a whole lot of math to figure out how many pounds of fuel are in the plane (aircraft fuel is measured in weight, not volume, since the volume can vary at different temperatures).
In Montreal, while fueling the aircraft, they discovered the fuel gauges weren’t working. The ground crew and flight engineer do the math by hand, and decide they have enough fuel onboard to get all the way to Edmonton. Only, they didn’t. They had made a mistake in their math. The plane would only have enough fuel to make the stop in Ottawa and about halfway to Edmonton.
Landing in Ottawa, the flight crew changed to the next shift and the ground crew checked the fuel again. The mistake could have been caught here with no harm done. Just top off the plane and they’re good to go. But, get this, the ground crew made the same mistake in the math! The plane took off destined not to reach Edmonton.
In the air, the plane runs out of fuel and one by one the engines die. The air crew did not even initially recognize the alarm that was going off because they had not been trained on this particular alarm. At the time, it was thought to be impossible for all of the engines to fail at once.
Years ago, pilots adopted checklists to use in emergency situations. They’re great because they force you to do things in order and help prevent you from missing something. The pilot grabbed the manual and frantically searched for the checklist for an all engine failure. Only, it wasn’t there. Just like no one had trained them on the all engine failure alarm, there was no checklist. The idea of all engines failing was so inconceivable, that an emergency procedure was never written. Air Canada Flight 143 is now a glider, losing speed and altitude and heading towards the ground.
So what does this have to do with UX Research? Well, to prevent accidents, the airline industry adopts the Swiss Cheese Model. The model suggests that no one safety check can catch everything. It relies on redundancies. Imagine each safety check is a slice of swiss cheese. If you hold one slice of swiss cheese up to the light, light comes through the holes. Add a second slice, and you still have light coming through, but much less. By adding several layers, several redundancies, you can greatly reduce the amount of light (accidents) that get through.
Now think about UX Research. No single form of research, no single project, will ever catch everything. Each piece of research is a slice of swiss cheese. It provides coverage, but it still has holes. That’s exactly why you need to adopt more research, more often. The more research you do, the more layers of cheese you have. The more layers of cheese you have, the lower your risk. Multiple rounds of research with varying methods are your redundancies.
A usability test is great, and it can help you determine if customers understand how to use your product. It won’t tell you if it meets a customer's needs though. You need exploratory research for that. Ethnographic research is great for learning in situations what people do and how they do it, but it won’t tell you if you should move forward with Product A or Product B, you need A/B testing for that. Just like redundant safety checks in the airline industry, you need redundant research.
Each research project is going to have its own nuance, its own specifics, even if there is overlap. There will always be different variables that will help you confirm, dispute, or expand upon your prior learnings. It’s why researching something you already “know” is a good thing! There is no silver bullet. There is no one, perfect piece of research or information that will ensure you don’t fail. You need redundancies. You need more pieces of cheese.
So what happened to the Gimli Glider? Well, it just so happens that Captain Bob Pearson, who took over the aircraft during the layover in Ottawa, was an experienced glider pilot. The flight crew managed to locate an abandoned airfield within their glide range, and prepared to land. As they got closer, they realized they were going too fast to land safely. Captain Pearson executed a maneuver known as a “forward slip” to reduce speed. This maneuver is commonly used in gliders, but almost never used when flying an aircraft the size of a jumbo jet.
The maneuver worked, and the flight crew got the plane safely on the ground. The plane suffered some damage because the front landing gear collapsed (without engine power, the front landing gear was unable to lock into place), but no one was killed.
The only injuries were from passengers going down the emergency slides, which it turns out were not long enough when the plane landed at an angle because of the landing gear collapse. Later, as part of the investigation, pilots were put into a simulator programmed to mirror what happened. Every single one of them crashed. It was Captain Pearson, an experienced glider pilot, the last piece of cheese, that prevented a catastrophe.
There is no right answer to how much research, how many slices of cheese, you need to prevent a product failure. At Key Lime Interactive, we can help you figure out the best research studies and methodologies to not only prevent a crash and burn, but also to help you design the best experience for your users. Contact us and let’s see what we can do for you.
Comments
Add Comment