The promise of deep insight at all phases of the insuring process through the microscope of eventuality not only offers the ability to create a strong competitive advantage, but also ensures companies perform their business to the mission statement they articulate.

The very best analytics engagements are, in fact, predictive. These not only guarantee that insurers are on track for what they aspire to be in the markets they serve, but also mitigate future exposures through forward-looking analysis of such things as best of breed business, elimination of risk before losses occur, optimization of business operations, and pricing reflective of the actual exposure insured. 

Beyond applying these powerful techniques to pricing, carriers are now discovering new classes of problems like inspection optimization that can be solved to drive improved risk selection and financial performance.  Traditionally viewed as a cost center, the inspection program presents the opportunity in forward-looking carriers to become a profit center and key strategic competitive advantage.   

Inspection Programs

A majority of property writers today supplement their underwriting and risk assessment process through the use of on-site field inspections. Field inspections create a visual of each risk, an assessment of the quality of the risk against certain exposure variables carriers assess and develop a snapshot of hazard or other conditional variables that increase loss potential. A more detailed assessment will also establish updated insurable values, interior conditions/exposures and violations of such things as code and life safety concerns that help score risk desirability, loss potential and premium adequacy.

Carriers that write homeowners insurance spend a significant percentage of their operating budget in underwriting for home inspection work. Many large carriers allocate one to three percent of annual underwriting expense budget to field inspections. This percentage is usually higher for mid-smaller companies. The majority of insurers outsource this work, so cost management and the efficacy of the program is a high priority.

Working with top property writers, we have discovered that the decision to inspect, as well as the intent of each risk review, is associated with long-standing guidelines or traditions, not necessarily defined on a risk-specific basis. In other words, carriers have "always done it this way", so when certain characteristics of risk (like age, size or geographical location) are identified, they have an inspection done. 

However, this paradigm is changing quickly.

After reviewing nearly two million actual field inspection projects for property writers, MSB identified that traditionally only about 25 percent of inspections performed using today's guidelines are actionable—or produce information that the carrier acts upon to mitigate or properly price a risk. 

In addition, actionable criteria used by carriers aren't always directly tied to loss frequency or severity.  Since the risk management process has never been closely aligned with what makes inspections actionable before the order is made, the industry spends over $140 million per year for work that will add no immediate value to loss experience or the bottom line.

The question at hand would therefore be: Could property writers improve bottom line results if truly actionable risks were isolated in advance of the inspection request, allowing these efforts to target risk identification and mitigation with precision? This is exactly the optimized state that progressive carriers are beginning achieve through the use of proper data and advanced predictive methodology.

In a recent industry survey conducted by Marshall & Swift/Boeckh, 87 percent of the respondents (U.S. and Canadian P&C carriers) have acknowledged homeowner inspection spending as a key business problem that needs to be optimized.  Some forward-looking carriers have recognized this business problem and desire to construct predictive solutions to recommend ideal risks for inspection that are, in fact, tied to predictors of high frequency and/or severity of loss.  These carriers are eliminating wasteful inspection costs and disruption to lower-risk customers.

New Answers

A property writer's homeowners inspection program can be considered to be running in an optimized state when:

• A predictive model is aligned to the front end of the inspection ordering and budget process.
• The model is giving lift to the action ratio of inspections or with an actionable rate of 60 percent or higher on the year's experience.
• Policies are identified for inspections based on the type of inspection needed.
• The best inspection program for each risk is assigned, which in many cases is not an actual field visit but might actually be an entirely different or alternative to traditional mediums, such as phone surveys, mailers or mobile surveys.
• And most importantly, companies see bottom line improvements through reduced loss experience and a closer alignment of premium with exposures insured.

Optimizing the Inspection Program

Data Strategy: Bringing together all of the existing historical inspections data is a good start, but careful consideration should be paid to assemble a comprehensive dataset so all the relevant predictors that will help the model perform well during actual production conditions are identified. Modeling in this problem space is a challenge of data quantity and quality.

It is common for carriers to start developing inspection optimization models with data that is available internally within the organization. This approach of using limited internal data presents serious challenges in developing a robust predictive model. A carrier's trusted business partners as well as non-traditional sources of data can be critical to the right information acquisition solution in this problem space.  An external data acquisition strategy is critical and enough time should be spent in qualifying valuable industry benchmarks, econometric trends, financial data, and location specific information.

About 30 percent of carriers that responded to a national industry survey conducted by MSB, identified the availability of good data as the significant challenge in initiating a project to solve the homeowner inspection optimization problem. These challenges included:

Predictive Model Development: The models should be validated enough to ensure they do not over-fit or suffer from the problem of performing well on the training dataset, but failing on actual production settings. Due to the limited amount of historical inspections data, modeling teams will also encounter problems in handling Selection Bias. It is recommended that carriers utilize an experienced modeling team that has successfully approached this problem space or ones technically similar in nature.

Implementing the Solution: As carriers are working towards optimizing the homeowner's inspection program, it is critical that alternate inspection techniques are thoroughly evaluated. One has to look beyond the traditional exterior drive-by, and exterior/interior inspections.  The new program should seek to optimize the balance between inspection expenses and underwriting accuracy.   In fact, different types of inspections are naturally geared to different kinds of risks and the application of any model predictions should apply this concept.  Good deployment planning, workflow integration design, and continuous model performance validation based on inspection results from the field and underwriter action is critical for the program's success.

Get Started Now

Getting started is imperative and the rewards are substantial. First, a carrier can benchmark the actionable rate of the current homeowner's inspection program and gain a sense of the financial impact of enhancing the efficacy of the inspection program.  Time should be spent in identifying actual drivers of loss by peril, frequency and severity.  Many carriers have this information from their efforts in by-peril rating and the same insights can be applied as a starting point to evaluate whether today's inspection program looks for the right criteria.

With more insight into the actual performance of the current inspection program and fundamental drivers of loss, carriers can begin the process of optimization. Here are some areas to focus on:

• Create a good external data acquisition strategy. Securing the right dataset is directly proportional to the predictive power of your inspection model. 
• Treat the inspection optimization project as a business problem and not purely as a modeling challenge. 
• Evaluate alternate inspection methods in addition to the traditional field inspections.

Since the experience in optimizing business operations with data and predictive solutions is becoming more standardized, many carriers are embarking on information based strategies to better compete in today's evolving market.  Applying a comprehensive data and analytic strategy throughout the solution delivery roadmap, carriers should be able to achieve an inspection actionable rate of 60 percent or higher and an overall inspection program ROI of 5:1.

(Chris Cartwright is CEO of Marshall & Swift/Boeckh, a supplier of local building cost information as well as residential and commercial property valuation technology and services in the United States and Canada.  Services also include MSB's InFocus Analytics, business intelligence, and industry leading "outsourcing" solutions.)