I have never met Dominic Cummings and I don’t suppose I ever will. But the thought occurred to me this morning that I don’t need to. I have met his type in my professional life multiple times over the last forty years. He is an iconoclast. He is the person who has developed expertise in a small field and has persuaded himself that it is applicable on a much wider stage. He is convinced that his special magic dust can take the place of tried and tested processes which have outlived their usefulness. And he is intelligent enough to convince people who should know better that he is right.
What has this to do with the technology that underpins the operation of commercial airlines? Pull up a chair and I’ll tell you.
Airlines were one of the first types of business to make extensive use of computers in delivering their products to the market. The basic standards for how the commercial environment around booking, ticketing and delivery of flight services were defined in the late 1950s by American Airlines and IBM working together to create a system that was eventually called Sabre. Other companies were working on the challenge at the same time but AA/IBM was the key partnership. By the way, this is a big subject and I am using a broad brush so necessarily I am glossing over a lot of the detail.
Sabre was built to allow real-time control of the inventory of seats available for sale on American’s flights and it was very successful. It used the most powerful computer systems available in its day and it represented a huge investment for both the airline and its technology partner. It stretched the capabilities of the technology available at the time to the very limit which meant that its programming was geared towards efficiency rather than comprehensibility or maintainability.
Over the next several decades Sabre, and the similar systems it inspired, added functionality to manage pricing, ticket-printing, seat assignment and many more aspects of the airline business. Because most of the systems followed similar standards, and all of them were constrained to follow government regulation around the world, they were able to work effectively together. Airlines could book their passengers on connecting flights operated by other airlines. A passenger’s bags could be safely checked in at the start of a connecting journey with reasonable confidence that they would make it to the last airport even if they had to be carried on multiple flights operated by different airlines. It was possible for travel agents to book flights directly in the airline’s reservations system without needing to wait for an operator to answer the phone and make the transaction. All of this was in place by the mid-1970s at a time when other businesses were just dipping their toes in the water of automated billing systems.
A huge proportion of what an airline reservations system was doing by 1978 when I started in the industry was mandated either by government regulation or by the need to maintain compatibility with hundreds of other systems. The scope for innovation was constrained and yet innovation continued. Airlines introduced revenue management from the 1980s onwards. This ability to vary prices according to forecast demand has boosted airline finances even if it has not always been popular with customers. Towards the end of the century airline systems were connected to the newfangled World Wide Web to allow customers to make bookings directly without need of travel agents or long phone calls. All the time though the basic structures of seat inventory, fare structures and ticketing standards continued more or less as they were in the 1960s[1].
Of course eventually other industries got on the IT bandwagon and by starting much later when technology was more advanced and cheaper they were able to advance rapidly. Business like Amazon, Facebook and Google could not have been created any earlier than they were because the technology was just not available to support them. Entrepreneurs who knew only the systems of the Internet age would look at the airlines industry’s huge spend on mainframe computer systems and enormously complex applications and scoff.
“How can this industry be so primitive?” they would ask.
“I could build a better replacement in 30 days in my garage” is a direct quote that I was in the room to hear circa1997.
And so inevitably the technological whizz kids who had fuelled the great dot com boom of the late 90s and the later Web 2.0 explosion convinced enough financial backers to provide the paltry funds that would be needed to build replacement systems using modern technology. I have personally witnessed a least three waves of this activity and I’m sure that there were many more.
The thing is though – not a single one of them ever replaced the decades old systems that they considered beneath their contempt. The very best of them managed to create something of value that could be added to the existing systems to enhance their operation. Most of them failed completely to do even that. Even the one that Google bought and failed to deploy at Air Canada. The point is that all that archaic functionality is doing stuff that needs to be done.
Around 25 years ago a few airlines realised that they could dispense with quite a lot of the complexity but only if they reduced the level of service provided to their customers. Low cost airlines like Southwest and Ryanair didn’t try to sell through travel agents or offer connecting services. They didn’t allow reservations to be changed or refunds given for unused tickets. And in general they only operated short flights of up to three or four hours duration. By stripping out a lot of things that are hard and complicated to do they were able to keep costs and hence fares down for the segments of the market that could live with those restrictions. The technology needed to support LCCs was simpler than the traditional systems that are used by the majority of airlines. The mistake that many of the new tech evangelists made was to imagine that LCC systems were cheaper and faster because of newer technology. In fact it was because they needed to do much less work and didn’t need to plug into a huge and complex global network.
So what has this got to do with Dominic Cummings?
Dominic Cummings is undoubtedly a person with a high degree of intelligence – albeit an intelligence that appears to be extremely narrowly-based. He has gained the attention of the money men and persuaded them that he can take over the running of government without all the complexities that have built up over the centuries since the principles of democracy began to be established. His crowning achievement was to engineer a victory in the Brexit referendum and then to follow up on that by pushing though the most draconian implementation of Britain’s exit from the EU. Just like the computer whizz kids who declared that existing systems were obsolete he has persuaded enough people that the European Union and Britain’s participation in it are over-complex, expensive and obsolete and so should be abandoned completely as quickly as possible. He is impervious to the argument that in the world of the 21st century the EU provides services that we as a country cannot do without. If we don’t get them from the EU then we will have to get them from somewhere else. But the current policy of driving out those public servants who may have the expertise to manage this does not bode well.
Similarly the ability to manage a novel pandemic disease depends on structures and processes that have been built up over literally hundreds of years. The current Cummings administration clearly does not believe in expertise other than political and financial manipulation. As a result the toll of Covid-19 in the UK, and England in particular, has been much higher than in other comparable countries.
In the airline world those companies that drank the Kool-Aid and were persuaded to commit to the new generation of technology discovered that it couldn’t deliver. Fortunately they still had the option of falling back to the obsolete systems that actually worked. Sometimes there was a financial loss involved in the adventure but at least the businesses were able to survive. Time will tell whether the damage being done to Britain’s governance and financial prospects by the Cummings scorched earth policies will be as survivable.
[1] In the noughties paper tickets were phased out but the underlying data structures continue in use in airlines systems and still form the basis for control of payments and entitlement to fly.
