Earlier this year, Qantas announced it had taken a controlling interest in a general insurance consultancy. Hugh Miller reflects on the explosion of interest in analytics across a range of industries
"The complexity of things - the things within things - just seems to be endless. I mean nothing is easy, nothing is simple." Alice Munro
Insurance is an industry with remarkable complexity; pricing is often opaque to the consumer and can be individually tailored. Online quotes have intensified competition. Finally, there has been huge growth in the amount of data available (geographic, meteorological, demographic, telemetric, competitor and historical claims) to understand insurance risk and demand.
This complexity creates significant upside in profits and market share if good pricing and price elasticity models can be fitted. Conversely, models that overfit to spurious trends or ignore the bigger strategic picture can lead to poor results. Thus complex modelling can lead to both a winner's curse or a winner's boon.
The challenge varies across business lines. For instance, injury management schemes such as workers' compensation have huge heterogeneity in injury types and recovery pathways. Better understanding of these pathways via analytics allows optimisation of support for claimants.
Insurance represented much of Taylor Fry's early analytics work. Although the core functions (pricing, reserving, claims management) remain the same, the actuarial support work has been shaped by technological change and advancements in modelling.
Government and welfare
"We should measure welfare's success by how many people leave welfare, not by how many are added." Ronald Reagan
Many governments have recognised the power of data and analytics for maximising the effectiveness of finite fiscal resources. Our company provides an annual actuarial valuation of the New Zealand welfare system, estimating benefit payments that clients are expected to receive over their working age lifetime.
By monitoring how this cost is changing, the effect of policy and operational changes can be measured, enabling better management. Targeted investments can be developed to improve employment outcomes and thus reduce lifetime cost - with measurable return on investment.
The models build in a large variety of risk factors including region, education and age to show how people's circumstances affect their pathways. Over the past three years the government has reduced lifetime welfare costs by over 10%, even after allowance for favourable economic performance. Disciplined measurement and implementation has made this 'investment approach' to welfare a powerful demonstration of effective government analytics, with other countries watching closely.
Another notable feature of this work is that it is one of the company's first supercomputing jobs; calculation is spread across a bank of computers, with full projections quickly creating terabytes of information.
There are many other government activities where analytics use has grown. There is huge potential to add value in areas such as fraud and compliance, health, justice, housing and education.
Airline and loyalty
"If you want to be a millionaire, start with a billion dollars and launch a new airline." Richard Branson
With a new majority shareholder, airline analytics will continue to be a core part of our work. Most of the work is in the loyalty division; Qantas, like many airlines, has a large and profitable frequent flyer programme. Analytics provides the opportunity to help leverage this. First, engagement in the programme can be increased via effective communication and offers. Analytics can maximise the useful information given to customers while minimising ineffective communication; achieving this leads to better engagement and sales.
Second, the airline is expanding its range of business services off the back of its data collection. This is helping third parties undertake digital advertising and market research, all with an emphasis on the scientific analysis of effectiveness. Part of this has involved exploration of web advertising analytics; how to effectively place ads to maximise impact, whether that be sales or brand recognition. Technically, this means the company now competes with companies like Google, which seems daunting. However, access to proprietary data and the chance to offer end-to-end solutions means there are significant opportunities to be developed.
Other industries
We have also been involved in some analytics projects in other industries, including telecommunications, energy and banking. Again, these industries are characterised by plenty of data and good planning, considering the long-term implications of decisions.
Recurring themes
"Twice and thrice over, as they say, good is it to repeat and review what is good." Plato
While every project is different, there are a number of recurring themes in our analytics work:
1. Ensuring the project adds value. This means constantly asking whether a solution will impact on business, and whether it is practically implementable.
2. A bias towards understanding long-term implications. Traditional actuarial work often involves projecting over long terms, and this skill is vital in many contexts.
3. It takes time to move into a new industry.
One must learn the language, dynamics and challenges; it is rare to successfully model without context. This means learning from other experts. It also requires lots of reading.
4. There is lots of competition. In Australasia, traditional actuarial work is performed by a relatively small number of firms. However in the broader analytics space, competitors include economists, technology companies, start-ups, management consultants, software vendors and even a company's internal analytics team. This involves learning humility - recognising that many other people are doing lots of work, much of it of high quality.
Source of success
"Whatever you are, be a good one."
Abraham Lincoln
While all success relies on a mix of planning and luck, it is helpful to reflect on what factors contribute most of the analytics practice.
First, strong technical skills are needed in data analysis and modelling. In a world where better prediction translates directly into revenue or savings, being able to provide a good technical solution to a problem is paramount. This increasingly requires expertise in statistics and computer coding. It also requires making your IT department your new best friend. Keeping on top of developments in academia can also be rewarding, as best practice continues to develop.
Second is business insight. Many analytics projects have suffered from either 'solving the wrong problem' or 'losing the forest for the trees.' Best-practice analytics goes beyond describing current trends; it must recognise how a model can be used to drive change and improvement.
Third is integrity. The actuarial 'brand' is widely respected because people believe that Fellows are appropriately skilled and give advice with honesty. We are fortunate to have inherited this reputation and recognise the need to maintain it in the advice we give.
The analytics industry is wide and fragmented; actuaries are only a part of a landscape that is evolving rapidly. However, our experiences suggest that actuaries have a valuable role to play.
Hugh Miller is a consulting actuary at Taylor Fry. He is also part of Institute of Actuaries of Australia's data analytics working group
