The approach to both runways at San Francisco International Airport (KSFO).

The Failure of Crew Resource Management (CRM) – Part III

Part III: Asiana Airlines Flight 214

This is Part III in a series of blog posts on “The Failure of Crew Resource Management.” Click here to read Part II. 

NB: The following post is adapted from a script for a TED-style video talk that was submitted in Dr. Adam M. Grant’s Management 238: Organizational Behavior course (Fall 2014). Special thanks to Harikrishnan Joy, who was my teammate in research and execution of this project. We conducted an analysis of the Asiana Airlines Flight 214 crash in 2013 through the lens of the Hackman and Ginnett research presented in Part II. In order to paint a picture of what happened in the cockpit on that day, we listened to the tapes stored by the plane’s cockpit voice recorder (CVR). We later found the official NTSB report and CVR transcript and acknowledge several omissions in our own transcript that were picked up in the NTSB’s detailed audio analysis – so please forgive what may at times appear to be a dramatization from the original events.

July 6, 2013. Asiana Airlines Flight 214 is on final approach into San Francisco International Airport. The hot California sun beats down on the Boeing 777’s aluminum exterior. It seems like a pilot’s perfect day – light wind, no precipitation, maximum visibility, no wind shear. At the controls is Captain Lee Kang Guk, a 45-year-old with 9,793 flight hours under his belt and completing his required Initial Operating Experience training. It is Captain Kang Guk’s first time landing the Boeing 777 aircraft at San Francisco International, an airport notorious for its seaside runway requiring an approach pattern over the Bay. It is also his first time flying with the gentleman in the right seat, Captain Lee Jung Min, a 48-year-old Asiana veteran with 12,387 flight hours. As the highest ranking pilot on board, Jung Min has the title of “Pilot in Command,” assuming responsibility for the safety of the 291 passengers and 16 crew on board. Although this senior position would have generally placed him in the left seat of the cockpit, on this flight Captain Jung Min also serves as a “checkride instructor” tasked with evaluating Captain Kang Guk’s performance. Behind both Captains is their subordinate First Officer, Bong Dong-Won, 41 years old, and seated behind the cockpit in the cabin is yet another Captain, Lee Jong-Joo, 52 years old.

As the plane passes through 1,700 feet and three miles out from the airport, Captain Jung Min notes that the airspeed is too high. This leads Captain Kang Guk to disengage the plane’s autopilot and set the engine thrust to idle, which causes the plane’s onboard computers to switch the automatic throttle into a “Hold” setting. Although both pilots expect this, Captain Kang Guk appears to misunderstand this feature of the autothrottle, believing that it would hold the plane’s airspeed at a desired level, when in fact the plane was now responding purely to his controls. As a result, the plane begins to sink below its intended glide slope to the runway. About two miles out from the runway, First Officer Bong Dong-Won alerts both captains to the unusually fast descent, by calling out “Sink Rate, sir”.

No response.

“Sink Rate, sir” he repeats.

No response.

One final time: “Sink Rate.”

No response.

Finally, at around 1.4 miles from the runway, Captain Kang Guk responds to this warning and begins pulling the plane’s nose up – and with the engines still at idle, this action causes the plane to slow even more. At eleven seconds before impact, the airspeed warning in the cockpit begins to ring. Four seconds later, Captain Jung Min appears to take control of the aircraft, pushing the thrust levers up to maximum. Air Traffic Control, noticing the plane’s unusually low altitude, calls out: “Asiana 214, go around.”

It’s too late.

At 11:27am and 50 seconds, Asiana 214 slams into the seawall at the end of the runway. The plane tears apart, making a 330-degree spin before smashing into the ground and catching on fire. Three passengers, Chinese students on their way to summer camp, are killed.

Crew Resource Management, or CRM, is a set of training procedures that focuses on decision making, leadership, and communication between pilots in the cockpit. The CRM concept was developed in response to the United Airlines Flight 173 crash in 1978, where two pilots failed to work effectively together in troubleshooting a mechanical problem. In a 1990 study on pilot interpersonal behavior, Robert Ginnett found “repeated evidence of poor crew work resulting in errors, accidents and incidents…” Following the crash, United Airlines revamped its crew training program, and CRM has since become the global standard for airlines. A 2014 study by Ford, Henderson and O’Hare in the Journal of Safety Research claims that CRM increases safety by reducing communication barriers in the cockpit and decreasing the traditionally hierarchical and authoritarian relationship between senior and junior pilots. This is evidenced by several pilots’ first-hand accounts of various accidents. As Captain Al Haynes (of the UA232 Sioux City incident) states: “…we had 103 years of flying experience there in the cockpit, trying to get that airplane on the ground, not one minute of which we had actually practiced, any one of us. So why would I know more about getting that airplane on the ground under those conditions than the other three. So if I hadn’t used [CRM], if we had not let everybody put their input in, it’s a cinch we wouldn’t have made it.”

What happened on Asiana 214 is a clear example of the failure of CRM. As pilot in command, Captain Jung Min failed to take any overriding action or make any commanding statement until it was too late. Caught up in an unbiased evaluation of Kang Guk, Jung Min may have been reluctant to intervene. Even more shockingly, both captains neglected the first officer’s repeated warnings. Kang Guk later said that it was “very hard” to make the decision to abort the landing, given the deference shown to superiors in Korean culture. By misinterpreting Jung Min’s silence as a sign of approval from a senior pilot, Kang Guk failed to act decisively even when he felt uncomfortable with the landing. And perhaps there was an element of face-saving – he didn’t want to look weak in the eyes of his superior by heeding a junior’s warnings. Furthermore, First Officer Dong Won may have been reluctant to speak up in the face of two senior pilots. His first two warnings contained the respectful suffix “Sir.” His third warning did not and was more authoritative in nature. Unfortunately, this came too late.

In an interview called “Why Teams Don’t Work”, J. Richard Hackman claims that many teams consistently underperform their potential. From his research, small teams who stay together for a long time perform the best. In the context of flight crews, the NTSB found that 73% of incidents occurred on a crew’s first day of flying together, before having the chance to learn how best to operate as a team. A NASA study found that fatigued crews who had a history of working together actually made about half as many errors as crews composed of rested pilots who had not flown together before. Hackman points out that while crews are cycled in and out due to financial and efficiency constraints, this often comes at the expense of team effectiveness. Decreased communication and less support break down positive team dynamics. Hackman found that the Strategic Air Command, a Cold War nuclear bomb squad, had teams that trained together and performed better than any other flight crew ever studied.

In Hackman’s eyes, there are 5 things for building an effective team. Teams need to have:

1) Realness

2) A Compelling Direction

3) Enabling Structures

4) A Supportive Organization

5) Expert Coaching

In addition, the use of a deviant, such as junior pilots speaking up, helps propel an effective team.

Moving forward, we suggest altering the process by which CRM is implemented. The breakdown of teamwork and communication between pilots can be disastrous. Even with CRM implemented, there have been many examples of improper decisions being made due to factors such as cultural norms. It’s critical for CRM training programs to explicitly acknowledge these teamwork impediments through the lens of previous failures. By using Hackman’s principles, we hope to see safety standards greatly improved and CRM failures reduced.

On Saturday, March 8, 2014, Malaysian Airlines Flight 370 disappeared while flying from Kuala Lumpur to Beijing. Although we do not know yet what happened to the plane, the dynamics in the cockpit that day will certainly be interesting to analyze in this organizational behavior context. Teamwork takes more than simply having a group of people; in the case of airlines and especially crises, a successful team needs to train together, stay together, and deal with the situation together.

Best Friends or Worst Enemies?

The Failure of Crew Resource Management (CRM) – Part II

Part II: An Evidence-Based Approach to Analyzing Failures of CRM

This is Part II in a series of blog posts on “The Failure of Crew Resource Management.” Click here to read Part I.

Continuing on my previous post, I looked into the surprisingly vast body of academic research on Crew Resource Management (CRM), and settled on two research works.

I’d first like to examine the work of J. Richard Hackman, a leading organizational psychologist and mentor to my own professor (Dr. Adam M. Grant). Dr. Hackman “spent a decade of his career studying how to improve the effectiveness of airline crews.” Below is an Harvard Business Review interview with Dr. Hackman entitled “Why Teams Don’t Work,” along with a link to his namesake publication.

Why Teams Don’t Work – J. Richard Hackman, Harvard University

Harvard Business Review interview conducted by senior editor Diane Coutu:

Original Work:

Dr. Hackman’s research is really interesting – I remember learning about his Job Characteristics Theory in MGMT104: Human Resources Management (Fall 2013). I didn’t realize that he had spent so much time researching airline crew effectiveness. In particular, I found it really fascinating that two pilots will only work together as a team every 5.6 years. On one hand, this completely makes sense – in the race for pure efficiency and cost savings, airlines need to have their pilots be transferable to different routes and aircraft at a moment’s notice. A good example of an airline that really buys into this mentality is Emirates – I recall watching a documentary about their new hub airport in Dubai, where everything is timed to the second to ensure that unnecessary costs are not incurred. I wonder, though, if this drive for financial efficiency is implicitly exposing the airline to safety risks – and not the usual physical/engineering risks afflicting airlines, but rather risks related to the way their crews function together. Dr. Hackman would certainly say that the airline crews need to stay together to achieve the best performance as a team. At the same time, a good team requires a strong leader to ensure that group members don’t get complacent. Finding that balance between “individual autonomy and collective action” is especially crucial for pilots – and in the case of Air France 447, the individual decision of one co-pilot was at complete odds with the actions of the other co-pilot and the captain.

The Hackman paper is great context for the next reading, which is written by Robert C. Ginnett, a former Senior Fellow at the Colorado-based Center for Creative Leadership. Dr. Ginnett wrote a chapter entitled “Crews as Groups: Their Formation and Their Leadership” for the academic publication Cockpit Resource Management (Wiener, Kanki, Helmreich, 1993).

Crews as Groups: Their Formation and Their Leadership – Robert C. Ginnett (LinkedIn)

“Cockpit Resource Management (1993), Chapter 3:

One of the examples given in the reading involved United Airlines Flight 173, a DC-8 aircraft that crashed in 1978. I shared this with my grandfather, who worked on this type of aircraft in the 1970s and recalls the incident. He was able to shed some technical insight on the crash – and in particular, he noted that many of the problem with the DC-8 aircraft (including a conditionally-faulty fuel gauge) were widely known by ground engineers but largely ignored by pilots. All in all, it was the failure of the crew to stay cognizant of, prevent, and then respond to the technical fault that caused the ultimate breakdown in teamwork and resulting crash. What happened 31 years later with AF447 is almost identical.

This is Part II in a series of blog posts on “The Failure of Crew Resource Management.” Click here to read Part III. 

"Inside the automated cockpit of an Airbus A330—like the one belonging to Air France that crashed into the equatorial Atlantic in 2009." -Vanity Fair, 2014

The Failure of Crew Resource Management (CRM) – Part I

Part I: “The Human Factor” – Air France Flight 447

Over the past few weeks, I’ve been discussing the following article with Dr. Adam M. Grant, my professor for MGMT238: Organizational Behavior (OB). It’s an analysis of the 2009 crash of Air France Flight 447, and is truly a captivating piece of journalism. By reconstructing the plane’s last few hours (based on cockpit voice recorder data), the article discusses the failings of an organizational behavior system in helping to prevent this disaster.

The topic of OB as it pertains to aviation/aerospace is beyond fascinating to me – and it’s something we touched on briefly in class a few weeks ago when discussing the 1986 Challenger space shuttle disaster. Here are my preliminary thoughts on the topic.

The concept of Crew Resource Management (CRM) became widespread across airlines in the 1990s. The idea was to encourage junior pilots to speak up when they felt that their senior captain was making a wrong decision. At the same time, massive technological advances meant that these junior pilots were not receiving the same quantity or quality of flight experience as their seniors (many of whom had flown for their countries’ air forces or had otherwise received intensive “fly-by-stick” experience when this was the norm). This automation in modern aircraft was designed with two fundamental expectations: a) that CRM is practiced by all pilots present in the cockpit, and b) that in the event of a technical failure, “pilot knows best.” As a result, you have planes that require pilots to communicate very effectively and fly the plane completely manually in an emergency. In the case of Air France 447, neither of the junior pilots were truly comfortable with the aircraft – indeed, the article notes that each pilot only had 4 hours per year of actual hands-on-stick flying time – but they assumed that the built-in autonomy would take care of them. This was combined with an unusual deference of the junior pilots to the senior captain, and a complete breakdown of effective communication. The CRM model failed, and the results were catastrophic.

I notice an interesting pattern between the events described in this article and a more recent crash: Asiana Airlines Flight 214 in 2013. The planes involved in flights 447 and 214 – an Airbus A330 and a Boeing 777 respectively – were both introduced in the early 1990s. Both aircraft were products of the same technological expectations described above, and both crashes involved some failure of the pilots to communicate effectively and respond adequately to the plane’s autopilot. Furthermore, the article mentions how the breakdown in CRM may be caused by cultural norms, and for this reason I wonder if the deference to one’s elders rooted in Asian culture may have in some way played a role in the Asiana tragedy.

This is Part I in a series of blog posts on “The Failure of Crew Resource Management.” Click here to read Part II. 

Qantas Ground Crew


Earlier this week, I read about a Cypriot low-cost airline that coincidentally went by the name of Helios Airways. The airline ceased operations in 2006 due to a tragic accident caused by an inadequately prepared flight crew – an event that ultimately led to the manslaughter convictions of five senior officials for their oversight in risk management and compliance. The story of this airline was a sobering reminder of just how risky this industry is. After reading the article “Startups Rarely Do Anything Well” by Eric Paley, I feel that my “boundless ambition” as an entrepreneur in the last few weeks has led me to overlook the one aspect that is ultimately paramount to success in this industry: customer safety. Unlike most other service-based firms, airlines have responsibility over their customers’ lives. This is a major part of the competitive landscape: airlines with the best customer safety procedures will thrive, and a single event caused by even the smallest oversight can serve to tarnish an airline’s reputation and send it into bankruptcy overnight.

Qantas Airlines has built one of the industry’s strongest brands around its accident-free record, by implementing one of the most complex and rigorous safety compliance systems in the world. From checking every bolt on every aircraft, to training flight crews in crisis management, Qantas has invested millions of dollars in ensuring that its passengers are safe. Here is the airline’s risk management model:

Qantas Airlines: Risk Management Model

Qantas Airlines: Risk Management Model

What this says to me is: at the end of the day, you can invest in the highest capacity, most fuel-efficient aircraft out there, but neglecting to invest in passenger safety and crew training is a recipe for failure. Managing the increasingly complex external risk environment is key to market dominance. Aviation entrepreneurs tend to avoid the subject of accidents – after all, no one ever wants to even imagine it happening to their airline. But this is something that needs to be discussed. I did a small analysis of the accident rates between helicopters and fixed-wing aircraft, using US data from the National Transportation Safety Board.

Type of Aircraft Accidents per 100,000 flight hours
Helicopter (rotorcraft) 9.47
Fixed-Wing (single or multi-engine) 8.38

Although the accident rates are very similar between aircraft type, the worldwide perception is that helicopters are far more dangerous than fixed-wing aircraft. A simple Google search will yield one of the largest passenger concerns: whether a “helicopter will drop like a rock if the engine dies,” although there is a significant body of evidence against this. Clearly, it would take a lot more than several compliance procedures to convince individuals that Helios Air will get them to their destination safely.

To be honest, I am fairly certain at this point that my idea for Helios Air will not come to fruition. Given the serious pricing issue I raised last week, along with other factors such as the immense amount of capital and operational risk involved, it is hard to think that a successful business plan could be crafted to offer profitability and market penetration within a reasonable industry framework. However, I am glad that I rationally considered the factors that ultimately falsified my various hypotheses. This meant that I didn’t remain “overly fascinated or over-committed to a product idea,” one of the key entrepreneurial pitfalls discussed in Chapter 8 of New Venture Creation.

In preparing for the ‘Venture-palooza’ on March 17, I intend to conduct some research on a specific substitute product for Helios Air: high-speed rail (HSR). This is a fast-growing mode of transportation in various regions such as Germany and Japan, and is currently being proposed in India. The StartupBoeing team doesn’t view HSR as a threat to commercial aviation, since the network of global aviation routes is approximately 4000% larger than that of trains. However, for short-haul routes where regular rail services already operate (such as Pune-Mumbai), HSR could be the service that beats even the fixed-wing airlines in this incredibly competitive travel market. More to come soon.