Tag Archives: insurance

Using Payments to Improve the CX

In an industry with infrequent customer touchpoints, like insurance, every policyholder interaction holds a lot of weight. Your organization only has so many opportunities to connect with insureds, which means one negative experience could result in policy cancellations or customer churn.

It’s imperative for insurance organizations to evaluate the one, universal touchpoint every policyholder must engage with: making premium payments. This is one of the few moments your organization has a policyholder’s attention, so evaluating and optimizing your insurance payment experience could be monumental to organizational success.

Because the policyholder payment experience is critical to the success of insurance organizations everywhere, we decided to uncover what this experience is truly like. We conducted an online survey in June 2020, through which we asked policyholders about their recent payment experiences, payment preferences and what contributes to a good user experience.

We discovered a few key points that are affecting the insurance payment experience. Here are a few of the biggest takeaways:

Policyholder retention is key

Combating customer churn and policy cancellations are well-known challenges in the insurance space – and those pain points were represented in our survey results.

To gauge overall satisfaction with their current insurance provider, we asked survey respondents how likely they were to look for a new provider in the next 12 months. In total, 45% of respondents said they are “likely” or “very likely” to search for a new insurance provider in the coming year.

The results was consistent across generations: 50% of respondents under the age of 30 and 52% of respondents ages 30-44 said they are also “likely” or “very likely” to look for a new provider.

The key takeaway: Retention and satisfaction should be major focus areas among insurance organizations.

See also: COVID-19: What Buyers Want Now

Convenience drives online payments

We dug into how respondents felt about their insurance provider’s payment experience. When asked how they chose to make their most recent insurance payment, 77% of respondents said they made an online payment, either through a one-time checkout route or automatic payments (like AutoPay). This response was consistent across all age groups; 87% of respondents under the age of 45 made their most recent insurance payment online.

payment methods

Next, we asked why this majority chose to make a payment online, rather than mailing in a check or calling their insurance provider. Overall, convenience was king.

38% of respondents chose the online option because they felt it was convenient, and a further 39% of respondents were already enrolled in AutoPay. This proclivity toward online payments is a fantastic trend for insurance providers; more insureds opting to make payments through self-service options ultimately means less work for your organization and, likely, a decrease in print and mail costs.

But, before you get too excited about those results, there is another side to that coin we must consider.

While online or AutoPay options appealed to many policyholders, we also found that payment platforms that aren’t user-friendly actually deterred online payments: 28% said they chose not to pay online because their provider’s system was too difficult to use.

So, while many insureds would prefer to make payments online, they may opt for a manual method if the online payment experience offered to them is subpar.

The key takeaway: Optimizing the online payment experience is critical; simply having an online payment option does not mean your organization is automatically providing a positive user experience.

Policyholders expect omni-channel offerings

We also wanted to get a sense of how satisfied insureds are with the omni-channel payment options (i.e. omni-channel capabilities, where you can pay a bill on your phone just as easily as you can on your laptop) their insurance provider offers. Overall, policyholders are satisfied with their options, with 46% responding “very satisfied” and 28% responding “satisfied.”

While it’s encouraging to see satisfied insureds, this feedback means a lack of omni-channel options could be a dealbreaker for your policyholders. If your organization is unable or unwilling to provide the flexibility of omni-channel offerings (which many of your competitors likely are), you could face customer turnover.

The key takeaway: Omni-channel offerings aren’t an option any more; they’re expected for insurance payments.

See also: 3 Tips for Increasing Customer Engagement

Insurance organizations can no longer afford to ignore their online payment channels – the results of our survey made that extremely clear. As one of the most frequent policyholder touchpoints, your organization’s payment experience could be the factor that determines churn rates and overall organizational success.

Simply put, optimizing your online payment channels is the best way to provide a positive policyholder experience and retain your customers.

New Actuarial Model for Unclosed Business


Unclosed business (premium) refers to the premium income from insurance policies that are yet to be processed, but for which the entity is liable at the valuation date. Because the insurer is liable for all unexpired risks irrespective of whether premium has been received, it needs to be stated in accordance with the standards set out for the unclosed business.

Unclosed business, also called “pipeline premium,” can be a significant component of unearned premiums at the balance date. The materiality depends on the distribution channels, such that, if the proportion of business written by brokers and agents were high, the unclosed business would be higher given that brokers can retain premium for up to 90 days after the policy inception date. Broadly, unclosed business can be broken down into these components

  • New business that has been written but not processed
  • Renewals with a date of attachment before the balance date that have neither been paid nor canceled
  • Broker business, where latest information about policies written has not been provided
  • Any others, in situations that lead to non-receivable payments by the company

Most of the actuarial literature gives little attention to models for estimation of premium liabilities. Usually the approach for estimation of unclosed business entails application of chain ladder and methods such as 12-month rolling averages, proportion of premium closed at the end of six months. However, given the lumpy and seasonal nature of unclosed business, some issues can arise with these approaches:

  • Methodology: Most often, methodology for estimation of unclosed premium is inconsistent and unstandardized across business units (BUs). BUs deploy different assumptions to calculate unclosed premium amount. 
  • Process efficiency: Process may be inefficient owing to multiple manual interventions in preparing the final numbers. There could be involvement of multiple people, leading to delays due to lack of coordination and availability. 
  • Accuracy: The estimated numbers suffer from low accuracy.

Given that there is regulatory requirement to report unclosed business, this article presents an actuarial model to estimate unclosed business on a monthly basis overcoming the issues highlighted above.

A Novel Approach for Calculating Unclosed Premium

A new modeling approach was developed with one of the insurance carriers across 10 different business units having 40-plus reporting segments. The data at a policy X term level was requested to start building the actuarial model. Though the model was built at the BU level with aggregated data, it was important to use the policy level details first and then roll up the data to the desired level. Over five years policy level data was received in Excel files, spread over five spreadsheets covering 10 BUs. 

Data Processing:

Data processing is a crucial component of model building, so a thorough approach was undertaken to ensure the robustness of the model. The following data sanitization techniques were performed:

  • Data Quality Assessment — The trends in the data were screened to rule out missing or inconsistent information that could potentially distort the model results. It was observed that for a particular business unit, an erroneous value was present in the data, and the same was highlighted to the client to confirm the veracity of that data point.  
  • Data Gap Assessment — The data requirements were shared with the carrier. Data for some BUs was not complete, and these gaps were called out to the carrier.
  • Data Modification — This entailed changing the formats, as some fields were text. Furthermore, a new field “delay” was imputed, to factor in the processing delay pattern (in months). This was used to prepare the closed gross written premium (GWP) triangle by processing delays over the policy effective months.

For the purpose of a standardized model, a single consolidated source of data was used, and the following data preparation steps were followed:

Figure 1: Data Preparation Steps 

Model Methodology

It was observed from the data that, for most BUs, 90% of the unclosed premium processed within eight months. An assumption was made for the model, that the GWP processes fully within 12 months for any given BU. Given the variability of data across BUs, automating the estimate generation process with a high degree of accuracy allowed for moving beyond point estimation. The model that was finally deployed used a simplified GLM framework, calculating the development in GWP through the chain ladder, volume-weighted, age-to-age ratios and assuming Poisson residuals. Becausee the method used bootstrap simulation, it yielded a distribution around the estimates, giving more information about the results. 

The figure below pictorially depicts the model methodology 

Figure 2: Bootstrap Chain Ladder Methodology

The pivot triangles created as detailed in the section were an input to the model. Using the cumulative triangle for GWP and the volume-weighted, age-to-age factors, fitted values were calculated. 

See also: NPS Scores Provide 3 Keys to Growth

Model Output 

In each simulation, using selected development factors, basis the trend in the pseudo-generated triangle, estimated an unclosed premium accrual figure. The 500 simulations permit the model to pick different trends that may have occurred in the past, and each unclosed figure generated is thus an estimate with a certain probability. Ideally, 10,000 simulations are sufficient to approximate the theoretical distribution as in Monte Carlo simulations. However, the mean and median results of the model varied by less than 1%, and, thus, 500 simulations stood a good approximation, given the constraints imposed by Excel. The trends in data can be quite volatile to estimate a single number with reasonable accuracy in an automated model; thus, we produced a distribution for unclosed business.

The results generated include the mean, 50th, 75th and 95th percentiles to cater to all scenarios possible. The 50th percentile closely resembles the mean value, as 500 simulations will normalize the distribution. The 95th percentile can be interpreted as the tail value in the distribution and less likely to occur but still can occur with non-zero probability (basis the peak in data fed in the model).

Automation & Standardization

To improve the process efficiency, the steps were automated through VBA macros to make the whole modeling process seamless. A single person can run the model, with minimal need for training. With few user inputs like the model month and location details to store the results, the macros cater to these tasks:

  • Data preparation steps 
  • Creating empty workbooks pertaining to each of the combinations of BUs and product classes. This is where the results get populated and stored automatically
  • Executing models as per the steps laid out in section above

It takes around 40 minutes to run the models for 40-plus reporting segments. This enables simplifying the process so that the actuarial resources could just focus on reviewing the numbers and efficiently use their time. As opposed to the existing process, which involved multiple people using different Excel models and disparate data sources, the proposed model reduced the average time by approximately 50%.


For benchmarking purposes, the model was run for each month from February 2017 to November 2018, because it was possible to calculate the actual unclosed business or unearned premiums for these months. The model outputs were compared with the actual unclosed and the unclosed premium booked by the client in general ledger for these months.

It was observed that the new model does better in terms of accuracy for all BUs. It was observed that the 50th percentile was a good prediction for all BUs for most months. In the peak months, like June and December, the 95th percentile was better at predicting the unclosed premium. In addition, for BUs with a very low volume of business in certain months, the 5th percentile fared well.


The advantage of the method proposed in this article over the traditional chain ladder is that it will automatically take into consideration the seasonality in data as resampling of residuals happens. Unlike traditional approaches like chain ladder, 12-month rolling averages require minimal intervention when selecting development trends. Its design allows for the addition of more BUs/reporting segments in the future. It just requires the data in the set format, allowing greater process efficiency. This approach is a win-win solution with respect to standardization, efficiency and accuracy, as was demonstrated with our client’s experience.

A final caveat is that one does not discount the need for review. The estimates produced by the model must be used in conjunction with new information that the business/underwriters possess, because the model takes historical data as an input.

How to Outperform on Innovation

Innovations will not be discovered, developed and scaled by groups of like-minded people of similar background. Solutions will demand diversity and inclusion leapfrogs. Leaders will have to take a fresh look at how they:

  1. Define what diversity means for their organization, redesigning processes and metrics to manage progress,
  2. Act and engage personally to create a more inclusive culture, with clear incentives to motivate change, and
  3. See and leverage the linkages between a diverse and inclusive culture and innovation.

Too many conversations about innovation prioritize technology as the major driver. Technology is one element of the complex innovation execution puzzle. Technology is abundant. Diversity (of thought and perspective, not simply gender and racial representation) and inclusion )the culture and environment where all members feel respected, valued and heard) are core to building and sustaining an effective innovation pipeline. Diversity and inclusion must be created and nurtured organically.

Diversity Must Be Broadly Defined, Sought and Measured

Reality is that while increasing gender and racial representation are essential, these are insufficient drivers of diversity. A gender and racially diverse organization will not be assured of innovation success. Also required in the composite profile: diverse life experiences and education backgrounds, and people who bring different perspectives and knowledge to problem-solving. An organization embracing this definition is set up to make brisk headway on innovation priorities.

Executives Must Personally Engage

Korn Ferry has identified five qualities of the inclusive leader. They:

  1.  Are open and aware, and able to adapt their behaviors to the needs of others
  2.  Advocate for diversity
  3.  Create a psychologically safe environment
  4.  Leverage differences, seeing difference as a source of greater insight
  5.  Drive results by fostering a diverse and inclusive environment

What else is required to build an inclusive culture? Especially for cultures in transition, being inclusive means rooting out and showing zero tolerance for exclusionary behaviors and rewarding inclusive behaviors that support the target culture. It means, for many organizations, innovating how diversity and inclusion efforts are designed and led.

See also: The New Shape of Innovation

Recognize the Linkages Between Diversity and Inclusion, and Innovation

It is unlikely any of us can name a CEO who will say he or she is not customer-centric. But many organizations fall short. Why? Only organizations that are diverse and inclusive can maximize their abilities to:

  • Understand customers as human beings
  • Develop empathy to build enduring and mutually valuable relationships
  •  Anticipate customer needs with speed and depth of insight
  • Address diversity risks, e.g., the loss of mid-career women with children or aging parents, emerging as a consequence of the pandemic.

We all see the severity of the challenges arising this year, many of them with long-tail effects. As leaders, innovators and change makers, it is up to all of us to advance diversity and inclusion in whatever organizations we lead or influence. It’s the morally right thing to do, and it’s the only commercially smart answer.

P&C Distribution: What’s Old Is New

There is a great deal of activity afoot in the P&C distribution space. New models are being explored. Old models are being upgraded for the digital era. In a recent SMA research report, I identified eight different models or options for insurers to consider. However, even though I positioned this as a revolution and an explosion of new activity, it is fair to ask if these distribution models are really new. About 3,000 years ago, it was probably King Solomon of Israel who said, “There is nothing new under the sun.” But the thought still holds true: The ideas remain the same; they just get reinvented and modernized.

This is very much the case in insurance distribution. Let me explain.

The new models to consider include: establishing digital brands, new affinity relationships, bundling insurance with the underlying product, digital marketplaces, worksite marketing for P&C, selling through new ecosystem partners and insurtech distributors. Although these certainly provide some interesting and important options for insurers seeking to expand distribution, it is not wholly accurate to say that they are new. In fact, (and now I’m dating myself) I distinctly remember being part of an industry research study in 1995 identifying scenarios for the future of insurance where we discussed options like bundling, affinity, worksite and selling through non-traditional partners in other industries. So, the logical question becomes – what’s different? Why are these types of options becoming popular and part of the strategy picture for many insurers?

There are some fundamental reasons why this whole range of channel options are important now.

  • The digital connected world: The world is undergoing a rapid digital transformation, which drives customers’ expectations. Every other industry is reaching customers via digital and mobile channels and technology have advanced to the point where it is relatively easy to do so.  
  • The API revolution: Connecting with new partners is significantly easier than in the past due to APIs being built into virtually every software solution. The ability to “plug and play” to connect to new partners for information exchange and transactions enables more dynamic partnering.
  • New competitive pressures: Insurers are seeking new ways to reach specific customer segments. As more insurers expand their channel options, they exert pressure on those that stick purely to traditional channels.

Thus far in the blog, I have not said the words agent or broker. But it would be a grave error to think that all the new channels will dominate and leave human intermediaries with a dwindling market share. In fact, agents, brokers, MGAs and other traditional distribution partners are leveraging advanced technologies. Insurers are providing many digital options for intermediaries to conduct business with them. And tech companies are providing innovative solutions for the agent/broker market. In addition, these traditional distribution players are also leveraging some of the other channel options to create hybrid models. Some are creating their own digital brands. Others are expanding their distribution through new affinity relationships, partnering with or acquiring insurtechs or connecting to customers through non-traditional partners.

See also: P&C Distribution: Blending Models

All the participants in the insurance distribution area enjoy many options. Thus, it comes down to selecting the channel mix that best aligns to business strategies and customer segments. The most important consideration is finding the right blend of the old and new. And it is evident that labeling some of the channels as old is a misnomer, given the innovation that is occurring across all the channel options.

In addition to the new SMA Research report, you can find more insights on P&C distribution in our Digital Distribution Virtual Experience on Dec. 16. This event is part of our Insights to Solutions Series.

How COVID Alters Consumer Demands

The insurance industry has long been suspended in a paradoxical state whereby state-of-the-art risk assessment actuarial models are used to project future losses, while being supported by notoriously lagging infrastructure and shared services. The need to leverage the ever-improving efficiencies offered by new technologies has become increasingly urgent, as have the variety and efficacy of those new technologies. Nevertheless, rapid innovation has been hard to achieve in the industry, and, while digital transformation has been happening in certain areas, the pace has historically been slow and reactive.

The 2020 COVID pandemic has served as a watershed moment for the industry. Prior to the COVID era, many carriers focused their digital transformation efforts on back-office areas:

  • Automating processing and servicing activities; including document ingestion, intrafirm data flows, insured notice generation and mailing, etc.
  • Modernizing the tech stack to facilitate internal operations; building workflows to connect distribution teams to underwriters, underwriters to support teams and so on. Additionally, underwriting platforms have been fit into multiple carriers’ process flows to get underwriters more existing risk details from brokers up-front, and the corresponding model inputs from actuarial teams, to improve risk assessment
  • Reducing middle- and back-office headcount, targeted at improving the expense ratio, a byproduct of the aforementioned focus areas

Now, as the world looks toward a post-COVID world, that focus has quickly shifted from back-office efficiency to front-office, customer-facing interaction. Seeing the shift that other areas, for example clothing retailers and retail banking, have seen toward always available, short interactions for customers to engage with services and acquire products, insurance carriers are moving their resources and strategic future planning toward areas where insurtech offerings can meet this new demand. 

Digital transformations that would historically take three to five years are now happening in under six months, spurred by software-as-a-service availability, ever-increasing service offerings for insurers and acquisitions. Customers simply do not want to interact with insurers the way that they have historically, and this is serving as the driving force for this digital transformation acceleration. Carriers that do not follow suit will quickly fall behind in this new digital reality.

See also: Re-engineering Claims Payments

Many high-impact insurtech capabilities are seeing massive growth in carrier interest and activity to address these new customer needs:

  • Virtual agents have gained popularity as a way to serve customers with specific needs who want to quickly complete their transactions. IPSoft and RozieAI both have virtual agent technologies that serve as omnichannel communication mediums between customers and carriers; no longer relying on preconfigured call routing or static online forms, these “humanized” virtual agents are able to parse and understand natural language, both text and voice, and either address customer wants completely virtually or route to a human agent quickly. This technology has driven increased volumes as well as higher customer experience reviews, both big wins in a highly competitive market.
  • Gig economy coverage is quickly becoming a necessary offering across major carriers to serve customer needs that have been growing rapidly both pre- and post-onset of the COVID pandemic, including food delivery, ride shares, home shares, etc. The digital nature of gig services requires a digital solution for the insurance covering them, and customers who seek coverage in the gig economy will always demand a more innovative model than the traditional insured-broker-carrier interaction paradigm.
  • Claims are being fully automated, both in adjustments and processing, in a customer-focused way that promotes speed and accessibility while still working to mitigate losses. Insurtechs such as OCTO Telematics and HOVER offer services that collect and process enough data to no longer require on-site adjusters for many auto or home incidents. With pre-installed data collection devices in commercial and personal vehicles, data about a car’s performance in an accident can be gathered at the primary source, rather than during an adjuster visit. Similarly, roof damage can now be adequately assessed via a series of customer-captured smartphone photos that are submitted to adjusters virtually, aided by technology that helps the adjuster properly assess the incident.

Industry disruptors serve as proof of the efficacy of focusing on these areas. Lemonade’s revolutionary customer experience has led to triple-digit year-over-year growth in gross written premiums, and it’s not difficult to see the popularity of its value proposition. A fully virtual chatbot, given a human name and face, quickly guides a potential insured through qualification questions for renter’s or home insurance (with more coverages promised in the future). Claims are often settled within minutes or seconds using propriety technology to assess customer reports. Lemonade has made it so easy to purchase insurance and file claims on a beautifully designed mobile user experience (UX), that customers are happy to never have human interaction.

In the post-COVID future, a differentiated, digital-driven customer experience and enhanced product offerings will determine who wins and who loses.  

Even with this new focus on front office engagement, there are still hundreds of insurtechs offering back-office solutions to further enhance risk analysis and processing, and the accelerating capabilities of these new services can have material effects on carrier performance. Data services increasingly rely on the Internet of Things (IoT) to collect data at the source and provide both risk-specific histories as well as aggregated datasets for new markets. Home monitoring systems and pre-installed vehicle sensors continuously collect massive amounts of data every day, and carriers must be ready to ingest and leverage it in their pricing and actuarial models, as well as use it to drive more robust loss prevention strategies.

To leverage these new technologies, carriers will need to become more comfortable using services operated on the cloud versus the traditional on-premises model. Of course, the sensitivity of customer data transmitted to an external service is of primary concern, and insurtechs are offering top-of-market security to address carrier requirements. The future of big data ingestion and efficient processing will require cloud solutions, and carriers who seize these new partnership opportunities will set themselves up for a better outlook in performance.

See also: 6 Megatrends Shaping Life Insurance

As carriers consider the new reality of customer-facing demands, as well as the new possibilities of internal process efficiency, potential partnerships have become more attractive, and even necessary, for success. As a result, clear opportunities for acquisition in these areas will likely emerge in the near future.