Big Data in the Pension Industry Part II: Big Data Today

16 November 2015

Take a moment and think about what happens when you log onto Amazon. One of the first things you experience is that infamous statement that goes along the lines of “People like you bought things like this...” Few people have not marvelled at how accurate Amazon can be in anticipating what products they might be (or are!) interested in. If Amazon can leverage Big Data to serve you better, can you use the same kind of technology to understand and serve your members better?

In Part I I provided a pension professional’s primer on what Big Data is and encouraged you to think of Big Data as the result of data technology increasing in capacity and dropping in price in precisely the same way Moore’s Law has been impacting computer prices. Your smart phone has the same computing power as the most powerful computers costing millions did in the 1980’s.  

The same processes operate on data technology so the ‘big’ in Big Data refers not just to the vast amounts of data that can now be economically produced and processed, but also to the impact the rapidly increasing use of data driven technologies is having on society. The Amazon experience described above is not just a technological experience but also a shared cultural experience. Part II will now discuss how these data driven technologies and methods are already being used in the pensions industry.

Big Data and related technologies have arrived in the pensions industry and are helping tackle some significant issues ranging from improving engagement and savings rates to helping deal with ever increasing regulatory oversight. 

At the centre of all of these challenges are the members and the need to serve them better:

  • Governance – How do I capture the voice of the member consistently and accurately? How do I demonstrate I am operating in the interest of members as defined by the goals I have set and in terms of regulations and best practice? How do I demonstrate that I am managing the scheme in a systematic and cost effective way? 
  • Fund and Scheme Design - How do I demonstrate that my design is fit for purpose? How do I understand who my members are and the needs that I should satisfy? Are my fund and scheme designs appropriate given my membership? 
  • Engagement - How do I keep members engaged? How should I communicate with members effectively and demonstrate ROI on that communication? How can I encourage healthier pension savings amongst my members?

Data driven technologies are now increasingly being used to leverage the data related to, and generated by, members to help deal with these issues. Members are at the centre of the pensions industry and member data is the key to serving them better. But before I go into how this is being done, it is probably best that I explain what I mean by “member data” in the first place.

What is “Member Data”?

Member data is information that can be sourced from HR systems, pensions systems, providers and external sources. 

Member data most commonly available consists of things like fund values, contribution rates, pension options taken (or not taken), savings behaviour, salary, gender, age, years of services and other demographic information. It may not sound like much at first glance, but you can do quite a lot with this information. For example, using the above you can estimate or derive engagement propensities, risk attitude and capacity, education needs and communication opportunities.

The above data together with derived information such as engagement potential and risk tolerance form the building blocks on which a wide range of applications that can be used by trustees/governance committees, HR personnel, service providers and managers to better serve members and meet their obligations more effectively at a lower cost.

Using Member Data to Serve Members Better

In Figure 1 below I have categorised some of the more common examples of how data is being used to serve members. At the centre of the diagram are the members – the focus of the pension industry and therefore the focus of the diagram. Around the members is the data they generate –everything from demographic information to more dynamic data such as savings behaviour. Operating with that data I have categorised data driven pension applications into three groups: Governance, Design and Engagement / Communications. For each of these areas, I’ll briefly describe some examples of how member data is being utilised today.

Figure 1: Data Driven Applications in the Pensions Industry


One of the most common applications of data driven technologies in the pensions industry is in a governance context, helping managers hear the voice of the member in systematic, evidence based and cost effective ways. Automated member analytics can be used to quickly and transparently assess memberships in terms of take up, savings rates, demographics, forecasting and other issues. This assists mangers in understanding where the membership is and where there may be issues.  Management action can be taken and then measured in the future to objectively assess the effectiveness of those actions.

For example, say a trustee committee has decided that improving savings rates amongst its membership is one of its management goals. Savings behaviour can be measured with respect to demographics and other factors to determine where in the membership there are savings hot spots and cold spots. Where savings are lower than would be expected – say amongst middle aged higher earners, action can be taken to increase engagement. Once the actions have been taken, savings rates can be measured again to see if the action has had the desired effect and adjusted as needed.

The appeal of this approach is that it is a way to manage pensions in a proactive, member centric, and diagnostic way that also has the appeal of being transparent and evidence driven. There is a stakeholder management benefit as well. Whether the stakeholders be member representatives or regulators, discussing scheme management executed via evidence driven, measurable  and objective methodologies is a much better conversation to be having.

Fund and Scheme Design

Fund design is a growing area where data driven methods are being utilised. The process begins by assessing how the membership can be segmented in terms of propensity to engage, risk tolerance and work/life cycle. For example, Figure 2 below shows how a membership can be grouped or segmented that represents different levels of engagement and risk tolerance. These segments are derived from data taken from pension systems and are based on a couple decades of academic work indicating that propensity to engage and risk tolerances can be estimated with reasonable accuracy from such data. 

The fund designer can then assess the needs of each of these segments and use it as inputs to the overall design. The final design is then tested against each of the segments to ensure that each segment is being served appropriately.  

The appeal of this approach is that it is very empirical, transparent and member centric. There is a regulatory angle here too as using data driven methods can help demonstrate that appropriate analysis of member needs were incorporated into the fund and scheme design.

Figure 2: Engagement and risk segments for fund design.

Engagement and Communications

Like the Amazon example given above, data driven methods for increasing awareness and encouraging action have been with us for decades and data driven methods for doing this are a staple of web retailers and media sites. The technology has matured and reduced in price to the point where the pension industry is beginning to leverage it to improve member engagement and savings rates.

Member segmentation can be used as a cost effective way to target communications to meet the individual needs of specific groups within the membership and the membership as a whole. Content, tone, channel and frequency of communication can be adjusted and targeted to meet the needs of specific segments or even down to the individual in more sophisticated systems. The diverse needs of different member segments, such as those just starting out vs those nearing retirement vs those receptive to communications vs those less receptive, can be catered to in a systematic and managed way.

Communications programs can be analysed in real time so that areas of relative success and areas needing improvement or adjustment can be identified and managed as needed. ROI on take up, education and communication response can be measured systematically, automatically and based on user data to determine where that ROI is positive and where efforts should be modified.

* * *

The above is just a brief snap shot of what’s happening with data driven technologies and the pensions industry. There is a perfect storm brewing where on the one hand we have an industry in a significant state of change coupled with a technology revolution advancing at ever increasing speeds. The need for change and technology are familiar bedfellows and what’s coming just around the bend has the potential to create huge impacts on the pension industry. 

In Part III we’ll explore the latest in “roboadvice” and discuss why a computer that can beat TV quiz champions at their own game would also make a decent pensions adviser. We’ll also meet ‘Big Mother’ – Big Data’s little sister and learn how she may also have a significant role to play in the near future of the pensions industry. You’ll be able to find this article on the Capita Employee Benefit website in the coming weeks.

Connecting with you

Get in touch if you have a question or would like to explore how our solutions could help your organisation.

Get in touch