Articles - Active portfolio management: Why active insurers are fitter

 

They can predict potentially adverse consequences, plan for a range of different scenarios, and change direction at speed as the circumstances require.

 

To achieve that, active portfolio management requires insurers to maintain good-quality data, but also to make it straightforward for business leaders – the insurer’s active portfolio managers – to visualise and analyse this intelligence. Relatively simple dashboards are an excellent way to achieve this, but producing meaningful outputs often requires collaboration across the business.

 

For example, a dashboard providing a breakdown of yesterday’s new business could include data from the sales team such as the number of new policies sold, average and total premiums, and the average discount applied. However, for leaders to assess the quality of that business, they will also require data on expected claims and on costs. Finance may be able to provide the latter, but more forward-looking data will require input from an analytics team with predictive models.

 

Once this foundation is established, insurers can scale these analyses to identify long-term patterns. By comparing performance across different periods, leaders can clearly distinguish meaningful trends from daily, weekly or quarterly fluctuations.

 

In practice, active portfolio managers will also want to be able to drill down into the headline data. Perhaps they want to understand how discounts varied on policy sales in different parts of the country, or the varying level of profitability generated by different agencies. Maybe it would be useful to understand how premiums achieved varies across different risk groups.

 

All of which sounds like a challenging ask of the insurer’s data teams. And it’s certainly true that in the past, building such dashboards and updating them each day required significant resources. Today, however, modern analytics technologies allow insurers to automate much of the work. Combining sales data and the insurer’s predictive models, it is possible to use a tool like Radar combined with Power BI to automatically update the dashboards each day.

In which case, business leaders need only to decide what information they want to see in their dashboards. That can be challenging, given the amount of data potentially available, but four items are critical for structured portfolio management:

 

For example, new business written, existing business, renewals and quotes issued but not yet converted.

 

KPIs could include number of policies written, premiums achieved, conversion ratios, loss ratios, based on expected claims, and combined ratios that include cost.

 

The periods that dashboards should cover, whether daily, weekly, quarterly or something else, and comparison periods.

 

These could include data on age, gender, region and distribution channel, for example.

 

So far, so good. But how should active portfolio managers use the information these dashboards provide? To put these insights into practice, it is helpful to think in terms of three strategic layers:

 

First, the “scoring layer” incorporates the insurer’s modelling outputs – from all the models that produce some kind of score for the likelihood of particular outcomes. That could include risk models, demand models and market models, but also scores related to issues such as capital commitments or anti-fraud reserving. The idea is to gather all the predictions that might inform active portfolio management.

 

Second, the “decision-making layer” is the stage at which the insurer maps its scoring against factors such as its appetite for risk, its operational and strategic targets, and its long-term business plan. It incorporates activity such as scenario planning and simulations designed to test the likely results of key decisions.

 

Third, in the “production layer”, the insurer executes the decisions made and then monitors the actual results on an ongoing basis. Those results can then be fed back into the scoring layer, so that the process continually iterates and improves over time as the cycle repeats.

 

Critically, within the scoring layer, active portfolio management requires insurers to think hard about those models – to return to the debate about which datapoints and dashboards are critical. It’s possible to look at the data in multiple dimensions, using different KPIs to drive different types of analysis. For example, measures such as actual to technical premium (AP/TP) ratios and retention rates can help insurers identify profitable clients – and therefore to guide future retention activity.

 

Interrogating the data can also help insurers to understand what they’re seeing in more detail – why are some client profiles loss-making, for example? Securing this understanding is crucial if improvements are to be made. Equally, the impact of potential improvements can be modelled before they are implemented.

 

All of this data processing and analysis is time-consuming. Insurers’ efforts to date have often been compromised by the manual workloads required. But the good news is that emerging technologies have huge potential to reduce those workloads; the advent of artificial intelligence (AI) tools therefore represents a potential opportunity to step up active portfolio management.

 

The process won’t change – from measuring and identifying key data to isolating and understanding the information. But by allocating much of that work to embedded AI, insurers will be able to accelerate decision-making, while freeing up resources for other value-additive tasks.

 

The ultimate objective should be to secure competitive advantage through speed to market. Insurers able to build the dashboards and models required for active portfolio management, and then to harness those resources to inform decision-making, will be in a far stronger position to hit their commercial targets and achieve their strategic goals.