Thursday, September 19, 2013

Network Planning - Mis(t)ery 101

Not so long ago, I was asked by a well known regional carrier in Europe to head their network planning department. Intrigued by the request - I am probably better known to fill positions that require long hours with my feet on my desk and a coffee in my hand rather than doing actual work - I accepted the preliminary discussions with management and found the network planning department in dire straits. This made me wonder, after all the airline was one of the more successful ones in Europe and network planning is an important intermediary in creating the big bucks. What went wrong with this particular airline and with many airlines all over the world in general?

Without doubt, the airline industry is a relatively new one compared to other industries in the modern world. Not that long ago - in fact, my grandfather actually remembers the prevailing headlines of his youth - two brothers hopped a few steps a few meters in the air on an engined crate that we now recognize as the initial predecessor of the B787, F35, and other mighty winged machines of today that are capable of performing incredible endeavours.

It is therefore of little surprise that the airline industry is far from a sophisticated one (but don't tell your passengers that, they might get scared!). Quite rightly, the focus of airlines have lied and still do today on safety, environment, and shareholders equity. Efficiency, on the other hand, is a topic that airlines all over the world, be it in Europe or Africa, still only are trying to get to grips with.

As an example, only in the 1970s was the ever so important topic of yield management developed (or should I say invented?). A lot had to do with regulatory bodies, of course. When IATA tells you that you absolutely cannot fly from A to B for less than 1,000 bucks and you can under no circumstances offer more than a cold, mushy sandwich and a bottle of water on that same route, who are you as airline CEO to worry about yield? Relax, have some coffee. After all, your competitor is bound to the same restrictions.

And network planning is in a similar stage of evolution. Even with large flag carriers of today, network planning is often done by senior managers that have a gut feeling that a certain route might be a good one for the company. Middle management then sets to work on the new route planning, and lower management implements the plan while finding out that the current fleet of aircraft actually is a rubbish one for the new routes! Airlines are high tech? Think again! This article's intent is to give a very brief overview of where the industry stands today with respect to network planning and where I believe it should focus on (and in all fairness, quite a few airlines adhere to the discussed concept, so I am certainly not entitled to that Nobel Price just yet!). I fully realize that I probably kick a few industry seniors against their shins by stating my views but hey, I just return the favour!

When we have a look at the current state of the industry's utilized networks, we quickly can see that airlines all over the world base their routes on two distinct network models:
  1. Hub and Spoke network; and
  2. Point-to-point network.

Of course, as with any model structure, any combination of the individual models is possible, too. We would refer to this combination model as a hybrid model. Since logically a hybrid model comes with all the advantages and disadvantages of the individual models, for the sake of this brief overview, we will not further discuss it in this text.

A Hub and Spoke model is the one that has been in existence since the early days of aviation. A Hub is an airport that is used by an airline to connect revenue traffic from at least two different flights. A Spoke then is an a flight that connects a Hub with a non-Hub airport.

A Point-to-point network is just that. There is no connecting airport, all routes are operated directly.

Both models come with individual advantages and disadvantages. The biggest advantage of the Point-to-point network is that very little network planning is required. This obviously comes with a considerable cost saving making it therefore the ideal network type for low cost (and this time I actually mean low cost as opposed to no-frills) airlines. Hubs come with other disadvantages too. So called banks (a cluster of arriving or departing aircraft to/from the hub) must obviously be scheduled in as short a time window as possible. After all, you do not want your fleet arriving spread-out throughout the day, keeping your resources bound and under-utilized. On the other hand, optimal banks create serious consequences for the hub. Just imagine having twelve A380s arriving within the same hour at the same airport. Apart from stand availability, baggage, catering, and cleaning bottle-necks, the airline would probably lose a considerable number of customers because of long immigration waiting times.

A direct revenue consequence of the Hub and Spoke model is a lower aircraft utilization. It is hardly possible to create equal distance hubs an thus individual banks are difficult to optimize. The latter directly resulting in lower load factors since passengers do not like to wait a longer period of time during a stop-over.

I would argue that this is also the key concept to successful network modelling. Traditionally, network modelling has been done based on individual financial aspects only. Obviously, this has been regarded as the optimum yield generating method. 1+1=2, right? Yet, by optimizing financial numbers on paper, one certainly might be able to please upper management but the most important yield generating parameter of the airline has been completely left out: the customer! An airline is, per definition, a company that renders services. Yet, good services please the customer and therefore generate better revenue overall. As a direct consequence, a low load factor on an individual flight should not necessarily be a reason for immediate termination of this service. If this flight contributes positively to the network as a whole, pro-rated yield could very well exceed individual yields thus creating added value to the network as a whole.

In practice, network planning should therefore consist of three individual major stages:
  1. Network strategy;
  2. Network design;
  3. Network optimization.
The network strategy comes directly from the top. It is derived from the airline's vision substantiated by its mission. For example, FlyHappy might have the vision to become the dominant carrier on Timbuktu to Brasil routes. Or it might want to be the number one feeder to the MilkyWay Alliance in Europe.

The design of the network then comes-up with a number of banks that fulfill the airline's strategy, simulates their respective profitability to the network individually and collectively, and, as a direct consequence, also determines the ideal fleet mix.

The last step fine-tunes, optimizes and monitors in real operations. This stage can also very well be used to evaluate additions to the network before handing over the network to yield management.

Concluding, I would like to reiterate again that network planning should be just that: planning of a complete network. Profitability of individual routes should be valued against the network as a whole. As my financial professor told me a long time ago, the art of financial success is not to turn individual minuses into pluses, it is to combine many minuses to one big plus. I wonder if the guy should have been working for the airline industry?

1 comment:

  1. Totally agree! As a network manager, I face these facts on a daily basis. To upper management, 1+1 always equals 2 and this is just not true in my profession. Good article!