Customer Loyalty In Insurance: Technologies and Strategies To Retain and Keep Customers Loyal
Insurance companies find it difficult to attract new customers and retain them, especially when almost every insurer offers the same service or product. It’s also challenging since insurance is a low-touch industry, and insurers seldomly interact with their customers. According to The Independent Insurance Agents of Dallas, the top companies on average have a 93 – 95 percent customer retention rate compared to 84 percent in the insurance industry.
However, with the advancement of technology and the introduction of digital platforms, insurers can now frequently interact with clients, collect their data, provide personalized services and loyalty programs. Check our article on insurance technologies to have a broader overview of innovation in the industry. This article will talk about the ways, and technologies insurers can use to solve their customer acquisition and retention problem.
The customer loyalty problem in insurance
It costs about 7 – 9 times more for an insurer to attract a new client than to retain one. On top of that, the study by IBM shows that the cost of acquiring new clients in the insurance industry is continuously rising. It is set to increase from 15.8 percent of gross written premium (GWP) in 2018 to 17.9 percent in 2022. Insurance companies are actively investing in customer acquisition rather than trying to retain old clients even though they lose about 16 percent of the customer base annually.
New customers get heavy discounts, and then the price starts to increase at each renewal, which instead punishes clients for being loyal. That’s why people would rather shop around for new and better premiums than stay loyal to one insurer. Nonetheless, loyal customers tend to buy more products, stay longer, reduce company expenditure, and promote their carrier company.
In recent years, insurers have been making efforts to increase customers’ loyalty by personalizing quotes and providing incentives. And it’s slowly paying off. According to Earnix, 60 percent of those insurers already achieving advanced levels of personalization are seeing a significant increase in revenue per customer, and 81 percent report that personalization has brought about an increase in customer retention. Now, let’s talk about those strategies in detail.
Customer acquisition strategy in insurance
As we mentioned, the cost for insurers to acquire customers is among the highest across all industries. But there are a number of techniques that help prioritize the right contacts and reduce acquisition budgets. One of the most common techniques is lead scoring — identifying the most prospective customers from a full influx of contacts. The idea here is to work smarter, not harder. While traditional lead scoring uses rather simple techniques of assigning scores based on specific user actions, like subscribing to a newsletter or checking several pages, a more advanced spin to it is predictive lead scoring. It uses machine learning to capture more nuanced customer behavior patterns.
Predictive lead scoring
Predictive lead scoring is a marketing methodology that uses propensity modeling to identify the sales leads that are most likely to convert. Propensity modeling is a statistical technique helping predict clients’ future behavior based on past actions. In this case, ML algorithms use customer insurance data to find attributes indicating that a lead has a high likelihood of being converted. The models are trained to assign higher scores to leads with those attributes. So, businesses can use predictive lead scoring to segment leads into three groups – high quality, medium quality, and low quality based on the likelihood of conversion. With this information, they can then apply different degrees of effort to the customer segments. To implement a predictive lead scoring system, insurers can build it from scratch or buy a ready-made one.
Off-the-shelf lead scoring systems. The ready-made systems such as Maroon.ai and HubSpot are easy to implement and can integrate with your CRM via APIs. Also, these systems come with their own firmographic and demographic data, which can supplement yours. However, the list of software tools that these systems can integrate with is limited, and since you will be sharing data with a third party, there will always be a security concern. Another major challenge with ready-made systems is that you may end up relying heavily on the provider’s software ecosystem. For instance, HubSpot requires you to have their CRM to run a predictive lead scoring feature.
Custom lead scoring. With the custom predictive scoring engine, you can build your infrastructure and carefully choose the data source. Also, while building the system, it’s possible to ensure that it can connect to almost any software you will be using. One of the biggest benefits of building the predictive lead scoring engine from scratch is that you can control your choice of data. However, building your system is a long and expensive process. And stakeholders need to be convinced that it will bring the desired outcome.
But there are successful cases. Indiana Farm Bureau Insurance partnered with BDO Digital to build a predictive analytics model. Once they collected the right data, predictive analytics algorithms and Azure machine learning were used to predict the leads that have a high probability of becoming customers. The models then automatically assign a score to each lead which helps agents to understand where to concentrate their marketing efforts. After the implementation, the company has seen a decrease in lead response time. However, before implementing machine learning and predictive lead scoring, you need data.
Data required for predictive lead scoring
Machine learning is only as accurate as the data fed to the models. This means that if you do not prepare and provide the right data, you will obviously receive inaccurate results.
So, what data is needed for predictive lead scoring?
- Demographic data. Some examples of demographic data are a job title, location, education, income, age, etc. They can be powerful conversion predictors. For example, Google claims that the number of searches for the term “insurance near me” has increased by more than 100 percent over the last two years.
- Subscription history. It gives you information on the type of policies your leads are interested in and likely to buy.
- Engagement and activity. This information is collected from consumer actions and behavior on the insurer’s website, such as the number of pages viewed, time spent on the website, interactions with support agents, etc.
Using this data, you can train the models for them to accurately choose the best leads. Once you have acquired the customers, another problem arises, which is how to retain them.
Customer retention strategies in insurance
The insurance industry is plagued with high rates of customer dissatisfaction which makes customer retention one of their biggest problems. For instance, according to the J.D.Power auto insurance 2020 report, only about 50 percent of customers can be considered brand advocates of their carriers, about a third are somewhat satisfied. And the dissatisfaction keeps growing. This is also confirmed by ACSI Finance, Insurance, and Health Care Report 2019 – 2020. Even though one should revisit these statistics after the pandemic is over, general dissatisfaction levels aren’t something uncommon in the industry.
Since carriers operate in a highly commoditized market, customers can hardly differentiate between providers. Therefore, people turn to make their choice purely based on the price difference. Once they find a better price, customers are likely to switch.
However, insurers are working hard to change customer perceptions of them. With the introduction of digital platforms, they can communicate more with consumers and provide personalized quotes and incentives programs.
Personalized quotes can’t be created from imagination. That is why customer data holds great value in the insurance industry. It provides carriers with the information needed to personalize their services and provide more value to customers. According to Businesswire, over 75 percent of consumers would share their personal data to get cheaper insurance premiums and personalized services. However, to personalize insurance, you need first to collect data.
Data collection and monitoring. Carriers can obtain new customer data using forms on their digital platforms during registration. This data can then be used to create a starting quote for the client. However, with this customer information, they are unable to create personalization for subscription renewals.
So, the walk-around for this problem is to offer clients an upfront discount to encourage them to opt-in for monitoring. The monitoring is usually done using telematics devices on cars; Fitbit devices that the users wear; fire alarms, security cameras, and other smart devices on property. Clients will also receive discounts for subsequent subscription renewals if they stay safe. This provides incentives for customers who opt-in for monitoring, live healthily, drive safely, and protect their property. For example, a monitored driver who is 40 percent safer than expected may be offered a 10 percent discount by the carrier and be confident the person will not churn. They often start by collecting data such as car information, driving history, personal information, current condition and amenities of the property, health history, and current health and activities.
Behavior tracking mechanism. All devices that monitor user behavior collect data that the insurer then uses to provide personalization. For example, cars that have telematics devices generate data based on the state of the vehicle. This data is then sent to the carriers for analysis to generate insights to help with claim processing and personalization. Note that the data is only sent to the insurers if the user authorizes the manufacturer to do so.
Data analysis and use. Once the provider receives the data, it’s able to analyze it using machine learning to understand the current state of its customers and act when necessary. For example, a warning notification is automatically generated and sent to the user if the driver is going above the speed limit. Or, if the client is overweight, then the carrier can provide them with health coaching services. Providers can also use this data to generate policies based on user risk levels, car make, model, mileage, and more.
Since clients are in the game of finding the best prices for their premiums, insurers should be able to play that game. For example, some insurance companies provide customers with the ability to adjust the price of premiums based on their financial situation. A good example is Trov — a world-leading insurance technology platform that provides its clients with the option to manage their premiums. Users can turn off insurance protection they do not need to reduce the overall cost through the mobile app. To prevent people from canceling their subscriptions due to financial constraints, it is better to give them the option to turn off the protection of some items. Personalization is a good way to retain your customers and keep them loyal.
Carriers like InsureApp provide their customers with custom quotes and policies based on their car age, model, annual mileage, user driving records, and lifestyle. This is made possible with Telematics devices and smartphone sensors that collect data such as driving speed, location, distance covered, daily activities, and more. The algorithms can predict accidents and health issues and take action when needed by providing timely and accurate notifications with lifestyle or driver coaching recommendations. Also, users who respect all the driving rules and live healthily will pay less, while those who don’t will pay more.
InsureApp provides customized quotes based on lifestyle.
Incentive programs in insurance
Another way to keep your clients loyal is by rewarding them with incentives. These rewards can come in different forms and motivate the consumers to interact more with the insurer. If your incentive programs are well-structured and captivating, clients may spend more or get expensive premiums just to be rewarded.
Incentives in car insurance. An auto insurance company called Allstate uses telematics programs called Drivewise and Milewise to monitor driver behavior. Users will receive incentive points every six months for safe driving hours, safe stops, safe speeds, and low mileage. They can then use these points for savings on travel, gift cards, brand-name merchandise, and more.
Allstate’s Drivewise app monitors driver behavior.
Another way in which insurers can grant incentives is through accident forgiveness benefits. Normally, when you have an accident, your insurance rates increase. However, drivers with good driving records enrolled in the accident forgiveness program will not pay more. Belairdirect is an insurance company that provides this program to its customers. However, to be eligible for their program, you need to be at-fault claims-free for more than 7 years before they can forgive an accident where you were at fault.
Incentives in health insurance. Rewarding clients for a healthy lifestyle also keeps them engaged, happy, and loyal. For example, Vitality is a UK-based insurance company that rewards its users for a healthy lifestyle. Customers can link their activity trackers to Vitality and earn points for being active. The more points you get, the bigger the rewards. Some of the rewards include discounts on hotel bookings with Expedia and Mr & Mrs. Smith and discounts on Apple Watches and Fitbit devices.
Vitality rewards users with points
Another good example of a health insurer actively providing its clients with incentives is Oscar. Using their digital platform, they provide users with health advice through a concierge team of medical professionals. They also set aside a dedicated health insurance advice team that can be contacted at any time through the app.
Oscar Insurance app
Incentives in home and property insurance. Some carriers have also taken the initiative to provide customers with discounts on their home insurance based on location, age, security, and upgrades. For example, StateFarm – offers home insurance premium discounts for customers that purchase various insurance policies from them. They also provide discounts on houses with fire, smoke, or burglar alarms and other home monitoring systems.
Predictive technologies to make retention strategies more efficient
For an insurer to retain clients, it is also in their best interest to identify users with a high propensity to churn and proactively re-engage them with personalization and incentive offers. To do so, they will need to first collect customer information and monitor their actions. They can use the same data which is collected for lead scoring and personalization. This data is used to train machine learning models to predict which clients are most likely to churn. Once the carriers have identified those users, they can also use machine learning to determine the best way to approach the clients and keep them loyal.
Customer churn predictions in insurance
Carriers can apply machine learning and data analytics to reduce customer churn rates using the following steps:
Detecting customers at risk of churn helps take measures in advance.
Collecting user data. To predict if a user will churn, you first need historical subscription figures. The data includes users who had historically renewed their premiums and a subset of those who had canceled. To help with the predictions, data related to the clients’ type of policy held, location, engagement, feedback, education level, and income can be used. Now, data scientists compile a set of attributes that are common to users who have churned.
Segmentation. The users are then placed into segments according to their level of engagement, monetary value, satisfaction, type of insurance, and more. Now, ML models are trained using datasets from each of the segments.
Observation. Once you have collected data, selected attributes, and segmented the customers, it’s now time to observe user behavior within a specific timeframe. To determine the period for observation, you need first to know when your users normally churn. It will ensure that you leave a longer window for re-engagement. However, it’s up to the company and data scientists to balance out the observation and re-engagement windows.
Building the churn prediction model. Now, it is time to develop the churn prediction model. Usually, the data scientists will train and test many models to find the one with the most accuracy in detecting potential churners. The common ML models used for churn prediction are decision trees, logistic regression, and random forest. The selected models are then deployed in a production environment and constantly tracked to increase the accuracy of predictions. With this information, insurers can invest resources efficiently and determine their next best action to help improve retention.
Dutch health insurance company CZ is a good example of an insurer that uses churn prediction to solve its retention problem. They developed and tested four models – logistics regression, decision tree, neural networks, and support vector machine. For them, the most important attributes for identifying churners were age, the number of times the customer is insured at CZ, and total health consumption. Out of all the models used, logistics regression and neural networks were the most accurate after performance evaluation. Using this data, the marketing team concentrated their efforts on the customers with a high probability of churning.
Churn prediction may be enough to act. However, suppose the carrier operates across several digital channels and has a variety of incentives up its sleeves. In that case, another predictive technique can help pinpoint the best strategy to address those churning customers.
Next best action predictions
Next Best Action is a customer-centric marketing technique that finds the best action a business can take, increasing their chances of gaining or retaining a customer. The strategy helps choose the best message, channel, and service to proactively engage the customers who are more likely to cancel their subscription. So, without machine learning and the next best action predictions, carriers could miss warning signs.
Identifying the customers likely to churn. That should not be a problem if the insurer had already built churn prediction models.
The next best action. Carriers can then use predictive analytics to market to those consumers based on their behavior. They can build machine learning models that use data such as income, gender, and age to predict the customer’s actions. There are two approaches for enabling the next best action – rule-based and ML-based.
With rule-based recommendations, rules are written in anticipation of users’ actions. For example, offer a 5 percent discount when the user engagement drops below a certain level. This is an incentive that may keep the consumer loyal to you. However, the rule-based approach is limited because it cannot act outside of set rules. So, if a user performs an action that has not been set, the system will not act. Therefore, you will constantly need to add and edit the rules.
With the machine learning approach, algorithms are created to find patterns in data and learn from them. Once the algorithm has learned the patterns, you should use it to select the next best action. Instruct the algorithm on which touchpoint you want to use, and it will tell you your chances of retaining a customer. For example, for people likely to cancel their subscription due to financial reasons, the models may tell you that providing a discount will increase the chances of retention by 65 percent. If you select a different touchpoint, then the probability will change since customers have different priorities and preferences. Your “next best action” should be the touchpoint with the highest probability of retention. So, using data like customer behavior and previous interactions with the brand, the models can create recommendations of the next best action.
This approach can be enhanced if the models are constantly learning and self-updating to ensure that they can make accurate predictions, given that customer behavior is not stable. This data from the models will help carriers know when to reach out and provide personalized attention and solve potential issues.
Focus on customer experience and their sentiment
To break the mold of commoditization, insurance companies have to focus on customers and leverage technology already available to advance experience. Besides the technologies we discussed above, there are several opportunities to look at. Let’s mention them in this concluding section.
Understanding customer emotion. Sentiment analysis is a technique that employs natural language processing (NLP) to understand customer emotional patterns in written or spoken texts. Carriers can apply sentiment analysis to any kind of text that they capture during interactions. The same IBM report we mentioned gives a case of a Japanese insurer that reduced customer complaints by 20 percent over the year of experimenting with the technology. It helped the company pinpoint the areas of improvement and act on these insights.
You may read our story on designing a sentiment analysis tool in the hospitality industry.
Reducing settlement time. A number of techniques are available to speed up claim processing and settlement period from hours or days to mere minutes. One of the rising strategies comes from auto insurance, where several companies use image recognition to evaluate car damage on the spot. Customers can take several pictures of their damaged vehicles with smartphones. You may learn more about this approach in our dedicated article on automated claims processing. Another powerful tool here is using the combination of robotic process automation and intelligent document processing to streamline paperwork.
Communication enhanced by digital assistants. When it comes to customer interactions, it’s hard to enable fast and responsive 24/7 first-line support. Digital assistants or chatbots have been around for several years. In insurance, they can address customers’ most common questions and requests, allowing the human team to handle specific cases. A success story comes from Finland’s financial service company OP Financial. Their strategy was to engage customers by providing financial and insurance advice on the website via a messaging widget. The portal experienced a 1,000 percent growth in a matter of months. But with the influx of visitors, the waiting time increased significantly. In response to that, the company designed two chatbots, both for financial and insurance communication. They managed to cover 44 and 50 percent of all digital conversations with customers, greatly reducing the wait time.
At the end of the day, customer loyalty and satisfaction revolve around a holistic experience with the brand. So, proactive reengagement with personalized incentives should not be the only focus but contribute to that overall experience.