Big data: why the pensions industry needs to know more about its members
It would be very difficult to design the layout of a supermarket if you had no idea what products people bought and in what order.
It is at least as difficult to design a default fund without knowing what age people are likely to retire and how they will want to take the money when they do.
That’s why the pension industry, from providers and lawyers to trustees, needs to better understand its consumers - the members.
How can it do this? According to a conference panel at the National Association of Pension Funds’ annual conference, the answer lies in big data.
What is big data?
According to Dr Eric Tyree, chief data scientist at Capita Employee Benefits, the best way to explain big data is to apply Moore’s Law to Ferraris. Moore’s law is the observation that the power of computer chips has doubled approximately every two years while the cost has halved.
If you applied it to Ferraris back in the 1980s - we would now have super-cars that cost less than a pound and travelled at nearly the speed of light.
Retail organisations like Tesco have been using big data for nearly two decades to help improve its Tesco clubcard. It is now so intrinsic to the organisation that they spend nearly a billion pounds a year giving people clubcard rewards.
How does it work?
Big data programmes follow a standard procedure, explained Giles Pavey, chief data scientist at Dunnhumby, the data science marketing consultancy responsible for the Clubcard project.
The first step is to gather your data. This can come from a number of sources such as research, online forms and behavioural habits.
Once you have pulled all the data together, you can use it to generate insights. These can be insights on an individual level - this person drinks semi-skimmed milk, or on a group level - people who buy expensive wines often also buy cheese.
From there, you then need to act on those insights - for example put the cheese near the wine in your supermarket. Pavey believes this is often where the problems come in, either because the insights generated aren’t actionable, or because the organisation is inherently resistant to change.
The latter problem is one that often applies to pensions. The industry generally has not been very open to listening to its members and then making changes according to their interests. Certainly for a long time providers only thought of their customers (schemes) rather than their end consumers (members).
How can we use big data in pensions
All that is now changing, at least in part because of freedom and choice. Now the majority of members are no longer defaulted into an annuity product, forward-thinking providers are starting to try and learn more about their consumers, to understand how they can keep them within their suite of products.
At the scheme level, trustees are trying to predict what members might do, and therefore how they might redesign their products.
Both of these are problems that can be solved by applying the same big data techniques that retailers are already using.
However, if the industry is to take advantage of big data it needs to be more open-minded, explained Richard Butcher, managing director of PTL. And while big data isn’t yet perfect - we need to act now, he concluded.
Otherwise - we’ll be designing a supermarket where the baked beans are kept in the freezer and the pizzas are found in the vegetable aisle.