by Charles Sybert –
When you walk down a city street, you get bombarded with billboards, storefront adverts, bus shelter posters, and all sorts of gimmickry enticing you to open your wallet and buy all sorts of products. Until recently, the marketing messages are focused on the general public and are very generic. Things are changing and technology has evolved to provide a one-on-one marketing experience based on the user’s precise location. Nowadays it is not unusual for a message to appear on your cellphone alerting you to use your reward points to buy your favorite coffee as you walk close to a Starbucks. That message is made possible using hyper localization technology powered by hyper-personalization. Using this technology, insurance companies too can provide notifications with a local texture and feel on an individual basis and tailored to their insureds and prospects.
Hyper localization is delivering specific content or data based on the users’ precise location. Most marketing and pricing are based on a geographical sizing of a zip code or perhaps a subdivision. Hyper localization delivers the user’s actual location or interaction based on a cell phone or other GPS-based device instead of a general area. To borrow an analogy, it’s like Norm walking into Cheers, where everyone knows your name. The only difference is, technology is used to enable a large company to make you feel welcome.
In today’s marketing, insurance companies can target markets based on the reach of various television stations and a more granular breakdown by cable or satellite television providers like Comcast and DISH Network. Through hyper localization, the message can go the most refined grain, the market of one. Using the exact location, messaging can be tailored to the user’s current experience. The messages can range from:
These messages are tailored to the exact experience the user is having at that moment with a unique local flavor. The goal is to deliver messaging that makes it sound like the companies know you and familiar with your locale.
Some of the fundamental principles to hyper marketing include:
The goal is to improve the insurance company’s image from that of a large company where a customer is only a number, to one that the insurance company is a community member.
It is commonplace for insurance companies to use location as a part of the rating algorithms. With current technology, the location is generally at the zip code or maybe the neighborhood level if using sophisticated location services. The exact location is usually very specific for coastal properties, but the more you move inland, the accuracy reduces. This can provide a false awareness of the location-based risks such as vehicle density, crime statistics, or other impactful areas. Usually, the gap between estimated risk and actual risk is believed to be small enough and will balance to a net-zero.
Perr & Knight performed a detailed location-based study on home policies in Florida, comparing street-level data with master location data. Perr & Knight found 5.7% of the book of business would require a premium adjustment, with 3.8% being underpriced. The premium adjustment ranged from an increase of 86.7% (or approximately $2,800) to a decrease of 45.4% (or approximately $2,100).
While the mispriced policies are a small percentage of the book, this can cause the book’s quality to erode. The overpriced policies could migrate to a competitor with a more accurate premium estimator, while the underpriced policies remain, causing an increase in risk exposure.
Essential items to keep in mind while reviewing the need for hyper localization pricing:
Using hyper localization efforts will help drive an ever increasingly more accurate premium to match the exact risks.
According to the FBI statistics, Insurance fraud costs the average US family between $400 to $700 of additional insurance premiums or about 40 billion dollars a year. This fraud can be broken down into three major types:
A group of individuals working together to file a group of claims across multiple insurance companies. For example, staging a multiple car accident or vandalism of numerous homes in different areas.
An individual purposefully performs an act of damage or submits a false claim that did not occur. For example, setting fire to a kitchen or submitting a worker’s compensation claim for an accident and treatment that did not happen.
When a claimant reports the value of a legitimate claim for more than it’s actually worth. For example, in a property fire, the claimant reports more electronics, furs, and other personal items when the claimant owned none.
Insurance companies are continually battling fraud through Artificial Intelligence and Machine learning for payment discrepancies, increasing scrutiny of the loss reports, and other statistical efforts. Insurance companies are now adding location intelligence and analytics to the tools to help identify and fight fraud based on hyper localization.
Location intelligence provides the capability to gather, organize, and visualize various pieces of data for a given incident and store the data for future trend analysis. At the claim level, the adjuster would visualize the geotags of all evidence, such as comparing the pictures location from the police report (if available), the claimant, and the insured. If there is a discrepancy, then the claim becomes flagged for fraud.
The FBI estimated up to $6 Billion of the $80 Billion government Hurricane Katrina relief was fraudulent. While Katrina was an apparent loss, smaller weather-related issues can be challenging to assess if the claimant was affected. For example, a small hailstorm strikes a subdivision, and homeowners make hail loss claims. Through hyper localization, it is possible to overlay the storm’s damage map received from the weather service with the home’s actual location received from satellite view with the insured loss reports. This will provide a precise area of actual losses versus fraudulent claims.
Using machine learning based models, trends can be identified based on the type of loss and location of the loss. This would identify organized or premediated losses as the concentration of a particular kind of loss is out of the ordinary. For example, one restaurant may have a claim for spoilage due to a freezer malfunction. A plethora of similar spoilage claims are logged in the same general vicinity over a period of 6 months. A human would have issues identifying this trend but using the hyper localization data and machine learning, the trend would be identified and flagged earlier in the process as possible fraud.
The fraud trend identification can impact the loss ratio and help underwriting focus on high-quality risks. When claims identify individuals or an area with a large percentage of fraud, underwriting can increase the scrutiny of applications received from that area, reducing future fraudulent behavior.
Using hyper localization capabilities sounds like it would be just adding a metadata element to existing systems, but it can be more complicated than adding a new source. The following is a list of critical steps to utilize hyper localization data:
When including hyper localization data, it is essential to focus on both the technology and business impacts.
Hyper localization is a powerful capability used by more and more insurance companies to interact with prospects and insureds. From a sales perspective, insurance companies can provide sales messages with a local texture or even the actual agent. During underwriting, hyper localization enables improved pricing based on proximity to known risks or density. Fraud detection provides the most significant opportunity as location intelligence with augmented analysis offers new ways to detect fraud not previously identified. The location analysis must be based upon a strong data foundation with a proper acquisition, transformation, and access enabling business decisions. Adding the tool of hyper localization will ensure your company remains at the forefront of the digital experience.