The Boeing 787 Development Story

Employees cheering 1st 787/10 in 2/2017

Today, the Boeing 787 is a successful product, with production rates at 12 units/month, and a total of 521 flying just over 5 years from launch. By comparison, in 49 years of production, Boeing built 1,528 units of the 747. And, having just flown in a 787 from San Francisco to Paris and back, I can attest that it was for me less tiring than in any other plane, which I attribute to the higher air pressure. It is close to that of Lake Tahoe (6225′) while other planes are closer to Squaw Valley High Camp (8200′).

Back in 2008-2011, however, the news coverage of the 787 was not so positive, as the plane’s product launch accumulated a delay of more than three years, with analysts pondering what had gone wrong. To keep this event in perspective, we should remember that multiyear delays in product launches have recently been the rule rather than the exception in commercial aircraft, worldwide. In Europe, the Airbus A380 was 2 years late and, in Russia, so was the regional Superjet 100. But the question remained of how Boeing, an organization with 100 years of experience in designing and building airplanes, could not have done better.

I would like to present here a few explanations that have been proposed, without passing judgment as to whether any or all of them are accurate.

What happened with the 787 launch?

The following timeline explains it:

What was special about the 787?

The following summarizes what was special about the design of this plane and of its supply chain:

  1. It used more composite materials than any earlier plane, as shown in the following picture:
  2. In manufacturing, it used a concept similar to what the car industry calls modular assembly, with major subsystems built by “strategic partners” and only final assembly done by Boeing. The sources of the different modules for the body are as follows:

    Sources of modules

  3. I recently heard a car industry executive explain that, on a map of the world, his supply chain looked like the route maps you find in the back of airline magazines, and the same is true for the 787:
    Looking at the geography of this supply chain, however, it is clear that the choices were not driven by the search for cheap labor. In the 21st century, Japan, Sweden, Australia, or the UK are not places you go to for this purpose.

Explanation 1: All Projects Are Late

Such a delay is nothing special. Product development projects are always late, particularly when they involve new technology. As described in The Project Game, in most corporations, managers shun project leaders who give realistic estimations and give the jobs to those who promise the shortest lead times compatible with the laws of physics. Once they fail to deliver on time, these leaders work extra hard to atone for it and, once promoted, repeat the pattern with the next generation.

Where such a system is in place, no amount of engineering ingenuity and no project management tool will make much of a difference. Chronic delays are baked into the culture and there is not much point in looking for causes that are specific to a given project. In Deming terms, there is no sense in trying to assign a special cause to a common-cause phenomenon.

Explanation 2: Management Obsession With Miscalculated Costs

Debra and Chad Smith devoted Chapter 9, pp. 155-182 of their 2014 book Demand Driven Performance to a case study of the Boeing 787, the “Dreamliner.” They view the choices made by Boeing as driven by the pursuit of lower unit costs, as calculated using the simplistic approach of the Generally Accepted Accounting Practices (GAAP), which means treating overhead as if it were variable.

The authors trace the problem back to Boeing’s acquisition of McDonnell-Douglas in 1997, as a result of which McDonnell-Douglas executive Harry Stonecipher became Boeing CEO. The Smiths specifically blame him for imposing on Boeing the excessive outsourcing that had sunk McDonnell-Douglas’s commercial aircraft business.

According to them, the goal was to increase Boeing’s Return On Net Assets (RONA) by reducing net assets. The search for a clear, consistent, and generally accepted meaning for “RONA” is a wild goose chase. “Net Assets,” for example is described by some authors as the same as shareholders’ equity and others as the sum of fixed assets and working capital. In most balance sheets, “Fixed Assets” include intangibles like intellectual property or goodwill, but some exclude these in RONA calculations. That product development is the creation of intellectual property makes this point particularly salient.

All we need to know in the case of the 787, however, is that, according to the Smiths, the focus on RONA made Boeing want the strategic partners to bear part of the development costs, which many insiders warned against, as leading to poor fit between subsystems and, in the long term, loss of intellectual property and know-how.

Explanation 3: Commercial and Political Pressure

One issue the Smiths never brought up in their case study is the political quid pro quos in the design of commercial aircraft supply chains. To let their airlines buy aircraft, some governments pressure aircraft makers to let local manufacturers build parts. Even inside the US, aircraft manufacturers are prone to spread their manufacturing facilities through many states and congressional districts to maximize the number of legislators with a stake in their prosperity.

To the extent that this happens, suppliers are not chosen exclusively on the basis of technical excellence, ability to deliver, and ease of communication. If spreading the work around is an unavoidable cost of being in this business, perhaps the risk can be mitigated by contracting out individual parts like machined wing components rather than elaborate subsystems like fuselage sections, especially when they are to be made of new materials.

The following table shows how 473 pre-orders received by 2010 broke down by country ranked by decreasing total fleet size with, in green the countries that have made the 787 a particularly large fraction of their total fleet, and in red the ones that didn’t buy any:

I don’t know of any relationship between the Japanese airlines that bought the planes — ANA and JAL — and the companies that make the modules,  but the conjunction of a high volume of orders and a large number of modules made in Japan suggests some linkage. Italy, however, is a counter-example. Italy’s Alenia is building the center fuselage while no Italian airline had bought a single 787 by 2010. Of course, just because there was no quid pro quo in Italy doesn’t mean there wasn’t one in Japan.

Explanation 4: A Failed Attempt at Copying Airbus

Another explanation I heard recently from an industry insider is that Boeing attempted to emulate Airbus’s product development approach but did not understand it well enough to pull it off. Why would Boeing do this? Boeing is one single company that grew in the airplane business since 1916; Airbus started out in 1967 as a government-sponsored consortium of companies in France, Germany, the UK, and later the Netherlands, and Spain.

All of these companies had a history in aeronautics but none had the critical mass to challenge Boeing on its own. From the beginning, each company contributed modules like cockpits, flight controls, wings, or fuselage sections to final assembly, a pattern that endured after the member companies merged into the Airbus company.

On its face, Airbus seemed unlikely to succeed. Only 25 years before the consortium started, the member companies had been making planes that were shooting at each other. They were based in countries with a centuries-old history of hostility and operated in different languages. How could such a team compete with one single company, based in a single country, that had dominated the commercial aircraft industry for decades?

Yet, against all these odds, Airbus did succeed and, according to my source, Boeing tried to apply lessons from Airbus to the 787. As it didn’t work out, Boeing reverted to its own approach, bringing in-house some of the work that had been contracted out, for example by acquiring facilities from Vought Aircraft Industries in 2009, and rebuilding internal capabilities that had been hollowed out.

Perhaps the Beluga is a fitting symbol of the Airbus approach, a cargo plane specially designed to move aircraft modules between factories. For the 787, Boeing’s answer was the Dreamlifter, a 747 modified for the same purpose.

Conclusions

Designing and launching a new, innovative commercial aircraft model on time is not impossible. Boeing pulled it off with the 777 in the 1990s, which raises yet another question: why couldn’t they do it again?

#CommercialAircraft, #GlobalSupplyChain, #SupplyChain, #Boeing, #Airbus, #ProductDevelopment, #Dreamliner, #787

 

 

2 comments on “The Boeing 787 Development Story

  1. Great writeup Michel on a complex challenge which human ingenuity is often greatly underrated:

    aerospace related projects.

    For me the most impressive example is Whirlwind (the first commercially used digital computer) and the project SAGE (Semiautomated Ground Environment) led by system dynamics creator Jay W. Forrester and his team at MIT Lincoln Lab, IBM being the contractor (with little experience back then with complex projects as this); well described in Thomas P. Hughes’ “Rescuing Prometheus”.

    The essence of such projects like the Boeing 787 is that the gained knowledge gets the chance to diffuse across the boundaries of the firm outside into the world enabling people to understand that complex projects have their constraints, yet proper project management and communication can make the real difference towards excellence.

    • Whirwind I was completed in 1951, and the MIT page on it does not mention commercial use. The first commercially used computer was actually the Ferranti Mark I built in the UK, a cleaned up version of the 1949 Manchester Mark I, the first true Turing machine, for which Alan Turing himself wrote programs.

      Manchester Mark I

      Ferranti Mark I

Leave a Reply

Your email address will not be published. Required fields are marked *