Hazards of Lean Service: Jet Blue’s Valentine’s Day Massacre

The February 2007 debacle at Jet Blue, which could lead to a Passengers Bill of Rights, raises a serious question. When service organizations strive for efficiency goals, do they risk catastrophic results for their effectiveness? Is Lean Service a bad idea for some service industries? If Jet Blue’s experience was an isolated one, we could say it’s an anomaly, but both United and American have had similar problems in the  winter of 2006-2007.

For the past decade we’ve seen manufacturing plants implementing lean principles, based on the Toyota Production System (TPS). TPS has made inroads into all facets of the economy, including hospitals — and goodness knows, we need efficiency gains in health care delivery! Concurrently, Six Sigma philosophies have also made inroads.  In fact, many companies implement a hybrid, “lean six sigma” program. Xerox and Thermo Electron have followed this route. Service organizations, including airlines, have implemented aspects of these programs to reduce cost, whether under the named programs or not. Yet, I’ll argue that a blind adherence to the efficiency side of the argument can lead to the disastrous effectiveness outcomes, which in this case could actually sink Jet Blue.

So what are lean principles?  The heart of lean is the elimination of waste. Seven types of waste — or “muda” as the Japanese call it — exist.

  • Overproduction beyond the customers immediate needs
  • Unneeded transportation of product — think about a poorly designed kitchen
  • Wasted motion
  • Waiting for parts etc. that leads to worker idle time
  • Processing that is unnecessary to the finish product (think of health care billing)
  • Inventory to buffer the system, but which can go bad
  • Defects in outcome, which leads to rework or scrap.

www.gemba.com has a very nice summary table.

Basically, waste can be found is any activities that do not add value to the end product or service. By eliminating waste, material velocity is increased. That is, a product goes from raw material to the customer’s hands very quickly. This delivers huge strategic advantages beyond the cost savings. Bad quality cannot be tolerated, and customers lead times are shortened.

Six sigma gets to the same end through a different path. Six sigma focuses on reducing variability, which is the enemy of efficiency, through various project work focused on specific steps where opportunities have been identified. A six sigma process has variability reduced to the point where there are fewer than 4 failures per million opportunities. Less variability means more predictability, and thus waste can be eliminated.

In both lean and six sigma, the result is a dramatic reduction in “buffers.” We buffer operations with longer lead times to address the uncertainty of production times, and we buffer with inventories to address the uncertainty of demand peaks and spikes. As we reduce uncertainty through more predictable processes, the need for buffers goes away. When you hear the improvement numbers from lean six sigma stories, you’ll think they were made up. They are real.

Waste elimination is a laudable goal in service operations as well as manufacturing operations. But there is a difference, and that lies in the source of the variation that causes the waste. Manufacturing operations, compared to service operations, are far more controllable; it’s almost a laboratory setting in comparison. Uncertainty does result from material and labor inputs, but those can be anticipated and controlled to a great extent. (In the days after 9/11 when the US closed its borders, we did see the uncertainty of raw material inputs dramatically after the car industry, but these types of experiences are rare.) The workers, the design of the product, and the production tools are all under the control of the operations to a very large extent. If sales and marketing are part of the process, the demand uncertainty can also be reduced.

In contrast, services operate in a sea of uncertainty and variability that are much harder to control. Let’s look at these sources.

Uncertainty in task times. The nature of service products is that the execution of each service delivery has some uniqueness. This variability typically leads to a negative exponential distribution of task times. Simply put, this means that while most task executions will fall within some tight range, some executions will take a long time. Consider airplane boarding. There’s uncertainty here, yet Southwest found a way to reduce the uncertainty and achieve faster turnaround times at airports, increasing effective capacity.

Uncertainty in demand. While service demand can be forecasted, no forecast is 100% perfect. Manufacturers can buffer this forecast uncertainty with some finished goods inventory. The simultaneity of production and consumption in services precludes this tactic. The capacity must be available when the demand arises.

Customers’ production roles. Both of the above uncertainties have much to do with customers. We introduce uncertainty. And because customers typically have some role to play in the production of a service, we introduce variability based on how well we perform our roles. Customers almost always have to provide information to service agents to initiate service, and we typically also have tangible tasks to perform.

Lean & six sigma work best in repeatable, standardized operations. While many services are repeatable, how well can they be truly standardized given the above? Let’s look at Jet Blue’s experiences.

Jet Blue has a very efficient operation, and they have achieved very high ratings in customer satisfaction — until Valentine’s Day. The disaster all stemmed from the horrible storm that struck the eastern third of the US. Weather is one of those uncertainties that airline service simply can not control. Flights will be cancelled, and passengers will need to be rebooked. This is an extreme example of demand uncertainty that leads to huge demand spikes. But to eliminate the muda (waste) of inventory going bad, airlines have cut capacity. I seldom fly on a plane that isn’t 90 to 100% booked.  With little bandwidth, it can take days to rebook all passengers on canceled flights. (Consider what happened to United after the Denver blizzards in December 2006.)

Airlines’ bandwidth to handle surges is further strained by other external factors. Government regulations restrict a pilot’s on-duty time. Time sitting in delays on tarmacs counts against this time window. Thus, even if they have the planes and the crew in the right place, the crew may not be able to fly, necessitating bringing in a new crew — but from where? How?

Jet Blue exacerbated the problem by having too lean an infrastructure in its central operations. Apparently, crews couldn’t contact central operations to get flight orders. 99.99% of the time that infrastructure was just muda. Not now.

In manufacturing, waste elimination in one category tends to reinforce other waste elimination. For example, wasted motion or transportation helps reduce defects. In services, the wastes more correctly can be trade-offs. To avoid the defects of poor customer service, some buffers are inevitably needed for those sources of uncertainty that cannot be controlled or contained.

So what’s the answer? (If I knew it in full, I’d be a rich man.) The airline service industry needs buffers against uncertainties caused by Acts of God. Jet Blue, being a young airline, had never experienced a need for an extensive contingency plan. Unfortunately for them, they didn’t get a gentle wake up call to this fact. The legacy carriers — for all that we hate them — do have these contingency plans. But contingency planning doesn’t eliminate the need for buffers; it just uses the available buffer more effectively. Passengers can still be affected terribly, like the folks United abandoned in Cheyenne, but part of the dissatisfaction results from lack of explanations. (But that staff is just muda, right?)

However, the financial structure of the airline industry is such that they can not afford extensive buffers. While the fault lies in many corners — the government-structured bankruptcy system that hasn’t let the industry rationalize, management mistakes in labor contracts, etc. — a big part of the blame lies with us, the consumer. We have become so price conscious in our airfare purchase behavior that we are blind to a better value.

Imagine an airline said, “We are going to deliver a higher level of service to you. Bags will be handled better. Cabin comfort will be improved. We’ll serve food. Waiting lines will be reduced. And in the event of weather disruptions, we are 90% certain we will rebook you within 24 hours of the end of the weather disruption.” Would that appeal to you?  Of course it would. But would you pay $100 a ticket for that extra level of service? $50? $25? I doubt many of us would opt to patronize this airline. (Airlines that have tried a value approach have failed.)

Much of the blame for our price consciousness goes to the airlines. Their yield management systems mean that prices seem capricious and bear no relationship to value. But even if pricing were more transparent, understandable, and value related, would consumers pay for the value? The government used to dictate airline fares. This meant that airlines competed on service. It also meant much higher fares. While we grouse — with the aid of the media — about airlines, air travel is a bargain these days. But wouldn’t it be nice to have real options in a value framework?

The disaster at Jet Blue is particularly surprising given the company’s background and guiding principles. When the founders were outlining their service design in 1999, they didn’t start with some operating model, e.g., hub-and-spoke versus direct connections or type of airplane. Rather, they started with a values statement: Safety, Caring, Integrity, Fun, and Passion. Yes, a cultural foundation lies at the heart of the company. Their operational design has reflected the values of the company — until now. But if they are true to their values and become a little less lean where expenditures are justified, they have a chance. (Can they sell the value of the “insurance policy” of contingency planning to customers? Talk about a tough sell!) If any airline could come back from a disaster of this magnitude, I’d bet on Jet Blue.