Frequent Flyer Programs about Revenue Management not Customer Satisfaction

The ultimate guide to airline customer relationship management and loyalty

Airline Frequent Flier Programs (FFPs) were introduced around three decades back by American Airlines in an attempt to reward repeat customers and build brand loyalty.

Since prior efforts were primarily focused only on attracting new customers, this was considered to be a novel approach to Customer Relationship Management and marketing.

Since then, CRM, which has acquired various hues and shades, is considered to be a crucial component of an airline’s business environment.

Let us attempt to understand both CRM and Frequent Flier programs. A generic definition of CRM is given below:

“Customer Relationship Management is the disciplined application of Customer Information to build customer relationships through.”

  • Continually refining insights into customer needs, habits and economics
  • Developing targeted and tailored value propositions based on those inputs
  • Strategically focusing business resources on activities that build long term customer and economic value.

A typical CRM initiative will therefore involve:

  • Collection and refinement of customer related information and
  • Gathering actionable insights along with their subsequent application

Thus developing an understanding of how customer expectations can be fulfilled by bringing about changes in service offerings and service delivery.

The ultimate objective of any CRM initiative, as for any other marketing exercise is a positive impact to the bottom line. However, CRM is different because it is a market pull strategy – ie. it identifies expressed consumer needs and acts on them, whereas generic marketing is a service push approach which tries to convince consumers about the need for a particular service.

Frequent flier programs are, by a subtle contrast, largely push based as suggested by the following definition:

“Loyalty programs are structured marketing efforts that reward, and therefore encourage, loyal buying behavior – which is potentially beneficial to the firm.”

However, FFPs have always promoted the CRM cause by essentially allowing airlines to access a minefield of valuable data about the customer’s travel patterns and behavior which is happily shared in lieu of reward miles and other fringe benefits. Therefore, they have retained their position as the staple CRM initiative for most brands.

Frequent flier programs: Long term loyalty, or short term cash?

Over a course of time, the utility of FFPs for airlines has increased exponentially, albeit in different ways. From a weapon in the marketing arsenal which required investment, they are now being operated as independent profit centers.

Let us see how this model works.

As compared to earlier loyalty programs which used to offer reward miles on air travel alone, contemporary programs reward buying behavior at retail outlets, financial firms and brokerages and through credit cards as well.

These firms purchase reward miles from airlines at a certain price, thus offering them revenue and free mass marketing while the customers get an attractive incentive to spend.

This means that miles are earned mostly outside the airline, and incentivize frequent buyers rather than frequent fliers.

Nevertheless, airlines love their loyalty programs due to their high cash generation ability which can be gauged from the following facts:

  • American, Delta and United’s FFPs each generate more than $1 billion annually.
  • Qantas reported more than $1.1 billion in partner payments for 1 year ending June 2011.
  • In 2002, when United filed for bankruptcy, the only part of its operations that was making money was its FFP.
  • In December 2011, Air Canada’s Frequent flier program Aeroplan announced that it will pay Air Canada Can$70 million ($60 million) ahead of schedule for reward tickets issued to help ease a cash crunch at Canada’s biggest airline .

However, these developments have led to a shift in focus from “customer satisfaction” to “revenue optimization”, turning FFPs into profit centers.

Moreover, while the existence of an FFP does not always guarantee motivation to use the airline, their unavailability is sure to drive customers away, thus turning them into a hygiene factor for the customer. Moreover, in some cases Customer data is also spread across many external partners and thus cannot be used in data mining or Business Intelligence activity.

While it is essential to invest and sustain a good Frequent Flier Program, it is also necessary to come up with viable improvements/alternate ways to handle Customer Relationships more effectively.

An alternative: Value based segmentation for loyalty programs

Let us go back to the definition of CRM, which states that ‘the disciplined application of customer information to develop continuously refining insights’ is a key aspect. Segmenting customers on the basis of their lifetime value is an effective way to realize this objective.

Value based segmentation is an approach that assesses the total value of a customer to the company.

The total value of a customer over his/her lifetime (known as Customer Lifetime Value or CLV) can be computed on the basis of various parameters and many times over, depending on the requirement.

For example, to justify additional marketing spend on existing customers, it might be required to evaluate their current wallet share vs. wallet size, ancillary service spend and satisfaction level with our services. However, for isolating customers at the risk of attrition, their behavior at touch points and long term versus short term spend will become more important.

Segmenting customers on the basis of their value to the company instead of flown miles (the current method of segmentation for FFPs) can be a better basis to gauge their importance to the Airline.

a. Calculating CLV – a preliminary approach

A Customer’s Lifetime Value is the sum of his/her Current Value and Potential (Future) Value. An elementary way to calculate this value has been formulated on the basis of the RFM Model, the components of which are –

  • R (Recent) – Period since last purchase [A lower value corresponds to a higher probability of the customer making a repeat purchase]
  • F (Frequent) – Number of purchases made within a certain period [Higher frequency, higher loyalty]
  • M (Monetary) – The cumulative profit that the airline acquired from the customer over a given period [Total Revenue – Total Cost]

While M is an indicator of current value, R and F are indicators of potential Value.

Let us see how Current and Potential Value can be calculated using a simplified approach.

Current Value (CV)

Calculating the Current Value is a fairly easy task, which involves summing up all monetary gains from the customer:

Potential Value (CV)

We can calculate the Potential Value of a customer using the AHP (Analytical Hierarchy Process) Approach. This approach comprises of 3 steps:

  • Step 1: Establish the criteria of evaluation and form the hierarchical Structure Model
  • Step 2: Based on judgment, create the weighted coefficient for each range value (C1 – C13) For example; the airline would prefer a customer who travels more frequently than one who travels once a year. Thus, C5 is an extremely preferred when compared with C1.
  • Step 3: Each Customer Potential Value can be calculated by summing up the range values.

A customer travelling seven times an year, and last travelled four months ago will have a PV = C2+C9

b. Interpreting the results

The Customer Lifetime Value score thus comprises of 2 parameters, Current Value and Potential Value. On the basis of these scores, customers can be slotted into four basic quadrants as depicted here:

Improvising the RFM Model

The RFM Model described above looks at the customer’s propensity to make their very next purchase, but it does not automatically follow that they will continue to be the most valuable customers over the long term.

his model is thus suited more to theoretical explanations rather than practical implementations, which will definitely involve several other parameters and acquire an increasing degree of complexity.

A more practical improvisation over the RFM Model has been proposed which uses sales data collected by multiple companies over many years.

  • This data is artificially divided into past, present and future time periods – eg. Considering the date 01/01/2000 as “now”, all data prior to this date becomes “past” and data subsequent to this date becomes “future”.
  • By this reorientation of time, statistical models can be built, tested and extrapolated.
  • Data on customers past behavior and value are fed into statistical regression equations which calculate combinations of behavioral features to see which ones relate to value.
  • The previous tenure of a customer and the duration of the relationship are factored in.
  • After fine tuning the model to relate past behavior to value using the Training Data, the model is fed ‘holdout’ data from a different set of customers of the organization. Drawing on the relationships which have been established earlier, the model predicts future values.
  • Type I (false positives) and Type II errors (false negatives) are determined to assess the percentage of error. These error percentages are factored in while making a CRM decision.

Best practices for a Customer Lifetime Value program

In an ideal scenario, a CLV score would be able to predict all future purchases of a customer, and the costs incurred to service them. It would be flexible and available to business units throughout the organization, who can then use it to evaluate specific facets of a customer’s behavior thus influencing decisions at customer touch points.

However, a CLV framework depends on an organization’s capabilities to capture and analyze data, and if not handled the right way, it can become quite complex to implement. Such situations can be avoided by ensuring that certain best practices are followed throughout the program.

1. Define program objectives clearly during initiation

While laying down objectives might seem to be an easy task, different units amongst the organization will most likely disagree on the parameters to be included in the framework. While the customer service and retention folks would want to focus on the highest value customers, the Marketing unit would like to filter out fickle and risky customers during the acquisition.

Recognizing the fact that different groups will have different requirements from the framework is the first step towards a clear definition.

In such cases, one approach is constructing different Customer Value scores for different business units.

2. Understand business, technology and data constraints

Though it would be really great to compare a customer’s spend at our airline to that with a known competitor, it might not be feasible or realistic to expect that the Customer Value team would be able to source such sensitive data.

Similarly, complex analytics performed on the dataset might help us uncover spectacular insights, but the cost incurred in manpower and software tools for such analysis might be much more than the realized benefits.

It is therefore necessary to understand and plan for such constraints in the initial stages of the project, and avoid setting unrealistic expectations with the business stakeholders beforehand.

3. Demonstrate clear outcomes in the short-medium run

Since the CLV framework references lots of different data types, creates large and complex data models and is integrated within the CRM platform (eg. it acts as an input to the customer service dashboard and the marketing tool), the time-to-market can be unreasonably high.

In order to ensure stakeholder engagement and increase support for the Project, it is prudent to plan simple deliverables in short spurts of time, thus generating quick wins.

For example, the marketing unit may receive a simpler CLV score within 3 months to influence basic business decisions, while more complex parameters keep on getting factored in subsequent iterations.

4. Make evaluation easy

It is imperative that performance metrics for the CLV Framework and success criteria are determined during the initial stages to give the CLV team a clear sense of direction.

Another point to not is that while statistical performance analysis measures (like limiting Type I and Type II errors below 5%) are required, it is also necessary to tie the framework with stated business objectives (eg. increase customer retention by 5%) to ensure smooth senior management buy-in for the program.

Using the CV Score: Multiple implementation possibilities

Value based segmentation can be utilized across various Airline modules in order to achieve business objectives. Traditionally, Frequent Flier Programs allow members to accrue points on the basis of their mileage i.e. the total number of miles flown.

However, there can be significant variations in the profitability of two customers with the same mileage owing to factors like acquisition costs, partner commissions and discounted airfares. In such a scenario, it is prudent to differentiate between two customers on the basis of their net worth to the Airline – an idea which can be implemented using the Customer Value score.

From a service delivery stand point, Customer Value throws up exciting possibilities to perform recoveries and subsequent delight.