As an industry, Insurance is traditionally less progressive when it comes to technology and innovation. It is still common to see carriers using antiquated systems that lack enterprise data and functional integration capabilities. In fact, only recently have major Insurance carriers taken into consideration the fact that there is a steady shift of consumer preference to online channels. Some would argue that due to the ‘traditional’ nature of insurance companies, innovating the way business is done is less common than in other industries.
With advances in technology, development of new services, and evolution of conventional business models, there are many areas Insurance companies can capitalize on to operate more efficiently, decrease risk, and improve both the top and bottom lines. Through my experience in the Insurance sector, I observed an area in particular that would benefit from leveraging both technology and services in developing innovative solutions for a traditional Insurance practice: property inspections for Homeowners Insurance.
Typically, all properties being submitted for coverage are inspected by the agent prior to submitting a policy application. Agents usually have little or no training in performing inspections and utilize very basic criteria in doing so. The Homeowners Insurance inspection process is an area in which Insurance companies can make vast improvements.
Agents should focus on revenue generating activities.
Today, many carriers depend on their agents to perform and evaluate properties on policy applications. This poses a few different issues:
- Opportunity cost: An agent’s job is to sell and service insurance policies; when time is spent inspecting properties, it is at the opportunity cost of potential revenue.
- Lack of expertise: An agent typically is not highly trained for property inspections, thus risks or potential hazards that would be captured by trained inspectors may go unnoticed.
- Conflict of interest: Agents are primarily compensated based on premiums for policies they write and consequently, there is an inherent bias towards passing the property inspection.
Third parties specializing in property inspections could be utilized to easily mitigate these pitfalls and provide immediate value in the quality of applications and therefore the quality of policies written.
Property Inspections should utilize Predictive Modeling to select inspections.
For most Insurance carriers the criteria for selecting proprieties to inspect are ambiguous and inefficient, which leaves agents inspecting properties that do not result in actionable findings. This wastes time and money – to the tune of $140 million each year. The Claims Journal reported in 20111 the Insurance industry spends about $190 million every year on property inspections; however, 75% of those inspections do not produce information that would mitigate future claims loss.
Many companies collect and analyze data to predict future outcomes in order to mitigate and act on risks today, a method called Predictive Modeling. Some examples include:
- Credit Bureaus use Predictive Modeling to determine credit scores
- Hospitals analyze medical records of large populations to forecast flu epidemics
- Some government programs use Predictive Modeling to predict and prevent crime
- Auto insurance companies use statistics to predict the probability of incidents
Currently, underwriters use specific data points and probability to price policies appropriately based on individual risk; however, this logic should also be applied upfront to decrease operational cost of conducting inconsequential inspections and increase the quality of information captured. This cost savings can be significant as each inspection can range from $25 to over $150 (depending on the type of inspection needed).
Although the use of statistical analysis and Predictive Modeling to identify risk seems like a sound approach, the technology is not mature enough to be a turnkey solution, providing immediate value in property inspections. The shortcomings are as follows:
- Valuable information is not always readily available and accessible to carriers
- Industry best practices on how to leverage information once gathered is not existent
- Which data points are important and how to interpret them is not clearly defined
- There are a few vendors that provide Predictive Modeling services and access to data such as Lexis Nexis and Verisk; however, these services are far from perfect and use protected proprietary data models, limiting their ability to adjust for each carrier’s specific needs
As a long term solution, an alternative to utilizing vendor Predictive Models is to build a data model in-house. Although most carriers do not own much of the data needed to build an accurate model, they should begin to collect and store it such that a model can be developed over time. Vendor solutions can be used in the interim and best practices should be collected in order to eventually develop a viable internal solution.
There is a need for change in the industry, the options are new and should be explored.
Leveraging Predictive Modeling and third party vendors can greatly increase the quality and decrease the cost of policies written. Only 37% of Insurance carriers leverage Predictive Modeling in the homeowners space (which is not specific to the inspections function), thus there is a lot of ground to cover to improve these operations. As property and casualty insurance carriers have experienced massive claims losses in recent years due to natural disasters, there is a need to take a second look at the system and find ways to mitigate losses instead of simply raising premiums.
1 How to Optimize the Value of Inspections, Claims Journal – http://www.claimsjournal.com/news/national/2011/10/17/193127.htm
2 Use of Predictive Models Widespread in P/C Insurance: Survey, Insurance Journal – http://www.insurancejournal.com/news/national/2013/11/07/310601.htm
3 The Challenge of Property Inspection Optimization, Verisk Review – http://www.verisk.com/Verisk-Review/Articles/The-Challenge-of-Property-Inspection-Optimization.html