6.3. A Variation on the Diffusion Model: The Rise of Low-cost Air Travel in Europe
The Bass model is a particular instance of the more general conceptual diffusion model introduced in Figure 6.2.[] In this section, we examine another diffusion model with similar feedback loops but entirely different equation formulations written to fit the early growth strategy of easyJet, one of the UK's most successful no-frills airlines, at the dawn of low-cost flights in Europe. The model shows an alternative way to represent word-of-mouth and advertising that nevertheless retains the essential combination of balancing and reinforcing loops associated with adoption and S-shaped growth. It is recognisably a firm-level model where diffusion of low-cost flights happens in competition with established airlines. It is a small model of some 27 equations in total.
[] Another diffusion model, with explicit and quite detailed representation of firms' influence on adoption, is reported in Morecroft 1984. The modelling project investigated the migration strategy pursuedby one of the Bell Operating Companies in the early 1980s as it attempted to upgrade the private telephone switches used by business customers from traditional electromechanical technology to electronic. Thousands of customers generating several hundred million dollars per year of lease income needed to be persuaded to adopt electronic switches. The model included formulations for product pricing, sales targets, incentives, sales force hiring, turnover and time allocation to represent the influences of the firm on market migration. For such a technically complex product, adoption was driven principally by sales force persuasion. Word-of-mouth sales were virtually non-existent.
Figure 6.10. Muted word-of-mouth
To appreciate the model's boundary and scope, it is important to imagine the European airline industry not as it is today but as it was back in the mid-1990s when full-service air travel was the norm and low-cost flights were a new and unproven business concept.
6.3.1. easyJet - A Bright Idea, But Will it Work?
The historical situation is described in an article called 'easyJet's $500 Million Gamble' (Sull, 1999). The opening paragraph sets the scene:
This case study details the rapid growth of easyJet which started operations in November 1995 from London's Luton airport. In two years, it was widely regarded as the model low-cost European airline and a strong competitor to flag carriers. The company has clearly identifiable operational and marketing characteristics, e.g. one type of aircraft, point-to-point short-haul travel, no in-flight meals, rapid turnaround time, very high aircraft utilisation, direct sales, cost-conscious customer segments and extensive sub-contracting. easyJet's managers identified three of its nearest low-cost competitors and the strategy of each of these airlines is detailed in the case study. But easyJet also experienced direct retaliation from large flag carriers like KLM and British Airways (Go). These challenges faced easyJet's owner, Stelios Haji-ioannou, as he signed a $500m contract with Boeing in July 1997 to purchase 12 brand new 737s.
Imagine yourself now in Mr Haji-ioannou's role. Is it really going to be feasible to fill those expensive new planes? In his mind is a bright new business idea, a creative new segmentation of the air travel market to be achieved through cost leadership and aimed at customers who are interested in 'jeans not business routines'. Feasibility checks of strategy are natural territory for business simulators, especially dynamic, time-dependent, strategy problems, such as rapid growth in a competitive industry. At the time, there were differences of opinion within the industry and even among easyJet's management team. Some industry experts had a dismal view of easyJet's prospects (in stark contrast to the founder's optimism), dismissing the fledgling airline with statements such as 'Europe is not ready for the peanut flight'.
To bring modelling and simulation into this debate, we have to visualise the dynamic tasks that face Mr Haji-ioannou and his team in creating customer awareness (how do you attract enough fliers to fill 12 planes?), and dealing with retaliation by rivals (what if British Airways or KLM engage in a price war, could they sustain such a war, what would provoke such a response?). The starting point is a map of the business, a picture created with the management team, to think with some precision about the task of attracting and retaining passengers and the factors that might drive competitor retaliation.
6.3.2. Visualising the Business: Winning Customers in a New Segment
Figure 6.11 shows how a start-up airline attracts new passengers and communicates its new low-cost, no-frills service to the flying public. The diffusion task is far from trivial, because when you think about it (and modelling really forces you to think hard about the practical details that underpin strategy), the company has to spread the word to millions of people if it is to fill 12 brand new 737s day after day.
Potential passengers represent the cumulative number of fliers who have formed a favourable impression of the start-up airline. Note that these passengers have not necessarily flown with easyJet, but would if they could.[] This rather abstract way of thinking about passengers is a convenient simplifying assumption that enables us to look at growth of interest in low-cost flights independent of the firm's capacity to offer such a service. Such simplification is justified for our limited purpose of exploring the feasibility of filling 12 planes. Bear in mind, however, that the scope of a model always depends on its purpose. For example, a model to study the growth of the whole airline (rather than simply growth of potential passengers) would include the company's internal operations such as hiring and training of staff and investment in planes, as in Sterman's (1988) well-known People Express management flight simulator.
[] We are drawinga distinction between wanting a product or service and actually buying it. The distinction is important in practice because customers often need time and further persuasion to buy (even though they are interested) and firms need to build capacity to supply. Interestingly, the Bass model makes no such distinction, an assumption that is not unreasonable for new product adoption at the industry level when one might assume that the adoption fraction subsumes aggregate product availability (i.e. if there is chronic capacity shortage during the rapid growth phase then the adoption fraction is lower than it would otherwise have been). For explicit stock-and-flow modelling of the stages of consumer awareness in new product adoption, see Finsgud 2004.
Figure 6.11. Creating awareness of low-cost flights among potential passengers: word-of-mouth and marketing
Source: From Supporting Strategy: Frameworks, Methods and Models, Edited by Frances O'Brien and Robert Dyson, 2007 © John Wiley & Son Limited. Reproduced with permission.
The number of potential passengers starts very small (just 5 000 in the model) and grows over time. How does growth take place? The remaining parts of the figure show the factors that determine both the increase and loss of passengers. In practice, this information comes from the management team, coaxed-out by a facilitator who is helping the team to visualise the business.
Passengers increase through a combination of marketing spend (posters, advertisements on the web, television, radio and newspapers) and word-of-mouth. The influences are similar to the Bass model, but the formulations are less abstract and not as elegant. Here, I provide a brief summary of the main concepts and equations. More information is provided in the Appendix of this chapter.
The increase of potential passengers depends on a host of influences. The first term in the equation represents word-of-mouth, formulated as the product of potential passengers and the conversion ratio. The strength of word-of-mouth is captured in the conversion ratio, which is itself a function of relative fare. The lower easyJet's fare relative to established rivals, the more potent is word-of mouth. For example, at a dramatically low relative fare of 0.3 (i.e. 30% of rivals' fare) the conversion rate is assumed to be 2.5 meaning that each potential passenger on average converts two-and-a-half more potential passengers per year. When the fare is as low as this, it becomes a talking point among the travelling public, just as happened in real life.[] At 50% of rivals' fare the conversion rate is reduced to 1.5 and at 70% it is only 0.3. Eventually, if easyJet's fare were to equal rivals' then the conversion rate would be zero because a standard fare cannot sustain word-of-mouth.
[] In some cases, very low fares may deter passengers due to concerns about safety, but in this particular case easyJet was flying a fleet of brand new 737s, which instilled confidence.
The second term represents the separate effect of marketing on the increase of potential passengers, formulated as the product of marketing spend and marketing effectiveness. Marketing spend is set at a default value of £2.5 million per year. Marketing effectiveness represents the number of passengers wooed per marketing £ spent. It is set at 0.05 passengers per £, so marketing brings 125 000 potential passengers per year (2.5 million * 0.05).
Finally, the effect of route saturation represents the limits to growth arising from the routes operated by the 12 new planes. It is formulated as a graph function that multiplies the previous two effects. When very few potential passengers have heard of the new service (people who live in the catchment area of Luton airport, 30 miles north of London, who might use these routes and planes), then the effect of saturation takes a neutral value of one and has no detrimental impact on word-of-mouth or marketing. As more and more would-be fliers in the region form a favourable impression of the new service, the effect of route saturation falls below 1 and begins to curtail word-of-mouth and marketing effort. When all fliers in the region are aware of the new service, there is no one else left to win over and the effect of route saturation takes a value of 0. The formulations for route saturation are explained in more detail in the Appendix of this chapter and can be viewed by browsing the model that will be introduced shortly.
The loss of potential passengers depends on service reputation. The lower the reputation, the greater the loss. Industry specialists say that service reputation depends on ease-of-booking, punctuality, safety, on-board service and quality of meals. For short-haul flights, punctuality is often the dominant factor. The model does not represent all these factors explicitly but simply defines service reputation on a scale between 0.5 (very poor) and 1.5 (very good). If reputation is very good then fliers retain a favourable impression of the airline, so the annual loss of potential passengers is very small, just 2.5 per cent per year. If reputation is poor then the loss of potential passengers per year is damagingly high, up to 100 per cent per year.
6.3.3. Visualising Retaliation and Rivalry
Figure 6.12 shows one possible way to visualise the retaliatory response of powerful European flag carriers to low-cost airlines in the early years. I must emphasise here the phrase one possible way, because there are many ways that a management team such as easyJet's might think about competitors. Part of the team's model-building task is to achieve the simplest possible shared image, drawing on the sophisticated (and sometimes conflicting) knowledge of the team members. A fundamental question is whether there is a need to model, in-depth, one or more competing firms. Do you need a detailed representation of British Airways or KLM to understand the threat such rivals might pose to the feasibility of easyJet's growth strategy?
The leader of a team modelling project should not impose a rigid answer on this question of how much detail to include. The modeller should be sensitive to the opinions of the management team while always striving for parsimony. After all, to achieve buy-in the model must capture their understanding of their world in their own vocabulary. In these situations, it is useful to bear in mind that experienced business leaders themselves simplify their complex world. If they didn't they couldn't communicate their plans. Good business modelling, like good business communication, is the art of leaving things out, and focusing only on those features of reality most pertinent to the problem at hand.
Figure 6.12 shows just enough about competitors to indicate how, collectively, they could stall easyJet's growth ambitions.[] Recall that word-of-mouth feedback relies for its contagion on the start-up's fare being much lower than rivals. But what if competing firms try to match the start-up's low price? The figure shows how such price equalisation might take place. At the heart of the formulation is a balancing loop labelled 'Restructuring'. Rivals' fare is shown as a stock accumulation that takes time and effort to change. Competitors cannot reduce fares until they cut costs and a flag carrier like BA may take years to achieve cost parity with a low-cost start-up.[]
[] I have chosen a high level of aggregation for rivals. The purpose is to capture in broad, but dynamically accurate, terms how rival airlines respond to price competition.
[] Large carriers will match low seat prices regardless of cost by providing some seats at a discount. Price cuts can be implemented very quickly through on-line yield management systems that allow dynamic pricing according to load factors. But it is a limited option. For example, out of 150 seats there may be 15 cheap ones. For very popular flights there are no cheap seats. Hence, the industry-wide effect of discounting is merely a partial adjustment to fully competitive prices. Only cost parity can deliver competitive prices that are profitable in the long-term for a firm catering to a growing population of price-conscious fliers.
Figure 6.12. Rivals and relative fare
Source: From Supporting Strategy: Frameworks, Methods and Models, Edited by Frances O'Brien and Robert Dyson, 2007 © John Wiley & Sons Limited. Reproduced with permission.
To understand the formulation, suppose that rivals set themselves a target fare equal to the fare set by the start-up (in this case, £0.09 per passenger mile). The magnitude of the underlying cost equalisation task is now clear – it is the 64 per cent difference between rivals' initial fare of 25 pence (£0.25) and easyJet's fare of nine pence. Such an enormous change can only be achieved through major restructuring of the business. The pace of restructuring depends on the time to change costs. Normally, one would expect this adjustment time to be several years, and in the model it is set at four years. The resulting equation formulation for the rivals' fare boils down to a standard asset stock adjustment equation where the change in the rivals' fare is equal to the difference between the fare set by the start-up and rivals' fare, divided by the time to adjust costs.[]
[] Sometimes, firms may seek ways to shortcut painful and slow cost-reduction programmes. For example, when easyJet first appeared British Airways launched a new semi-autonomous airline called 'Go' to compete, free from the service traditions and cost constraints of the parent airline. Go offered reduced in-flight service and much lower fares than regular BA flights, to a variety of European destinations, from London's Stansted airport. Ultimately, this move proved to be a short-term fix. Go was subsequently sold and BA continued to implement cost-reduction programmes for many years.
Figure 6.13. Feedback loops for the launch of a low-cost airline, a variation on the diffusion model
Source: From Supporting Strategy: Frameworks, Methods and Models, Edited by Frances O'Brien and Robert Dyson, 2007 © John Wiley & Sons Limited. Reproduced with permission.
6.3.4. Feedback Loops in the easyJet Model
Figure 6.13 shows the main feedback loops in the complete model. In the centre is the growth engine, involving potential passengers and conversion from word-of-mouth. Above is the limiting process in which route saturation eventually restricts the passenger conversion rate. These two loops are the equivalent of the reinforcing and balancing loops in the Bass model. In addition, there are two new loops. In the bottom right of the figure is a balancing loop involving rivals' fare and cost cutting. The dynamic significance of this loop is that it tends to equalise rivals' fare with the start-up's fare. Relative fare rises gradually to parity, thereby reducing the strength of word-of-mouth. Finally, on the right is a balancing loop involving potential passengers and loss rate that captures churn in potential passengers arising from the start-up's service reputation.
Of course, this brief model of passengers and fares is a sketch of a more complex reality. Nevertheless, it contains sufficient detail to fuel team discussion about passenger growth and price retaliation. When simulated, it contains sufficient dynamic complexity to create thought-provoking growth scenarios.