Broker performance analytics "need to be more targeted"

Broker performance analytics “need to be more targeted” | Insurance Business America

Business strategy

Broker performance analytics “need to be more targeted”

Data scientists and agencies must work in tandem for the best results

Business strategy

By
David Saric

Analyzing the performance of brokers and agencies requires more targeted initiatives to pinpoint individualized areas of improvement rather than using more generalized data. To get the best results, data scientists need to closely collaborate with other areas of a company to ensure the correct business insights are present in the final product.

This is according to Justin Milam, associate director of Willis Towers Watson (WTW), who said that “broker performance analytics need to be more targeted, and you need to be asking the right questions in order to get actionable information out of them.”

In a conversation with Insurance Business, Milam detailed what modelling techniques can adopted for greater business insight, the types of questions that should be asked to get useful findings and what challenges may arise when working with a data scientist.

Moving past previous modelling systems

Traditionally, the heuristic approach to analyzing a business and its employees proved to be an accessible means of getting more immediate and digestible information about loss ratios, amount of business being written, conversion rates, and other information.

“This data would then be used to determine bonuses, whether an agent needs auditing or if additional training is required to streamline and bolster productivity, among other things,” Milam said.

“When starting out, some simple one-way or two-way interactions may be the most appropriate model to build until there is comfort with the methodology,” said Milam, emphasizing the usefulness of the heuristic approach as a transition into more sophisticated measures.

These more advanced modelling techniques include a generalized linear model (GLM), with the target variable being loss ratio, or a nonparameterized gradient boosting machine (GBM).

Milam recommends layering both GLM and GBM techniques to pick up on data that may be lost or unaccounted for with each separate process.

Utilizing these methods can give a more nuanced look into a business’s current book and what can be amended for future growth opportunities.

“In the models, you can look at whether your agents are writing multiple lines of business, the credit scores on that business, prior claims, as well as how an agent’s profile can determine whether they’re going to be successful or not,” Milam said.

“You can also look at shifts in business over time. For independent agencies, there could be challenges where if a particular company goes in or out of out of the market, you could see your shift of business really change.”

“Adopting a common starting point to understand what a business wants to accomplish is key, especially if new systems or techniques are being brought on board,” Milam said.

For example, if a company is trying to figure out if a recent hire can perform to its standards, a barometer for success needs to be clearly defined. Whether it is a low loss ratio, high conversion rate, the amount of business being written or the likelihood of longevity within a certain company, each of these will impact a statistical analysis and produce varying results.

Being able to work with a data scientist to delineate a more restricted analytical framework will help generate information that is targeted and will not run the risk of affecting employees or lines of business that may not be relevant.  

Acknowledging the challenges of updated analytics

A WTW report from 2021 found that only 10% of companies were using advanced analytics in their agency or broker management, pointing to widespread skepticism towards data science and a confirmation of the philosophy “if it ain’t broke, don’t fix it.”

“Maturing analytics culture is not something that many want to adopt with open arms, so it is crucial to ease into it in a way that doesn’t seem ominous,” Milam said.

“Data silos really need to be broken down to present relevant information. You want to make sure that the data that you’re using is what the agents are seeking. If you’re coming up with some calculation for a loss ratio that is used by the agency for other diagnostics, there’s going to be skepticism around the use of that.”

Presenting the findings is just as important, as many brokers may not be receptive to an Excel spreadsheet like a data scientist and may find an infographic or pie chart more accessible.

However, certain business considerations may supersede the use of a model as the data scientist may have initially intended. For example, “if a model is built to reduce binding authority for agents with high projected loss ratios, it may be difficult to get buy-in from field executives and general agents if the producer has had a low loss ratio historically,” Milam said.

While this may prove frustrating for some data scientists, adding value in any capacity to a typically neglected area of analytics is worthwhile under any circumstance.

Related Stories

Keep up with the latest news and events

Join our mailing list, it’s free!