Using analytics to keep afloat

The board of directors of ABC Insurance, a property writer, has just finished reviewing its results and is ecstatic about what it sees. A healthy loss ratio of 58%, together with a combined return of 13%, has everyone beaming and congratulating each other. The directors have no doubt their shareholders will be pleased and reward them for their remarkable leadership. But below the surface, there’s a problem here. The problem is this: ABC’s managers are driven more by the bottom line than by the true dynamics of the business – and these dynamics remain largely unknown to them.

HIDDEN TRENDS

Statistics is one of the insurance industry’s most effective tools. It provides a high level – or, more accurately stated – an ‘averaged’ view of a business’ performance. The statistics process swallows seemingly unmanageable volumes of information and spits out convenient averages at the other end – averages that, at first blush, seem to provide the information needed to make effective business decisions.

However, it is often said that the truth is hidden in the details. For this very reason, high-level statistical averaging can be misleading and especially dangerous if used to make competitive business decisions. Bottom line averages are better used as a general indicator of overall performance. In the example above, we see a healthy loss ratio, but the averaging required to obtain that result, by definition, removes the highs and lows from view. The only thing the ratio really tells us is that, generally, more things are going right than wrong.

ABC Insurance’s problem has to do with the details hidden deep within its book and claims experience. Significant, yet unknown trends exist that could threaten the ongoing performance of the business. If the insurer could somehow discover what those trends are, however, it could mitigate any dangers and start taking advantage of hidden opportunities within its business. Even though most insurers do some detailed analysis, we know from real-life analytics engagements that most are barely scratching the surface. Many unknown opportunities exist just waiting to be uncovered.

ANALYTICS TO THE RESCUE

Analytics is a technique by which experts dig deep into the data of an insurer to find hidden trends. Analytics combines the best aspects of science and art. It uses statistical tools, experience-based, trial-and-error methods and best practice methodology. In a sense, it could be viewed as a forensic science that looks for hidden truths about a business.

Here are some hidden trends that analytics uncovered for ABC Insurance. While some are rather troublesome, many others show hidden strengths and growth opportunities.

* Despite a healthy average loss ratio, a detailed analysis of their book would show that the loss ratio for residences between 1,000 and 1,500 square feet is actually 76%, indicating that re-pricing or market share decisions should be made.

* On the other hand, properties between 2,500 and 4,000 sq ft have a loss ratio below 50%, uncovering a market segment that could be aggressively targeted.

* Analytics would also show that bi- and tri-levels have a higher claim frequency, as well as a higher-than-average claim payout. Plus, the percentage of these styles within their book is much higher than the industry average. Why is this happening and what can be done about it?

* They would also be surprised to learn that sewer backup is a huge problem in a couple of small communities where sewer systems are old and subject to failure. The insurer doesn’t know that because their current risks in those communities have no reported loss experience yet – with “yet” being the operative word.

* Further analysis would show a high percentage of their insured properties in a certain region are in high risk brush fire hazard zone. That’s trouble waiting to happen.

* Accurate valuations of each residential and commercial policy would show that virtually all policies are underinsured, presenting both a premium opportunity and a business exposure. The overall situation shows that 80% of their book is underinsured by 25%. This problem is never addressed because conventional wisdom states that: (1) total losses rarely happen, (2) there is an overwhelming need to stay competitive, so pressure dictates that limits remain low and (3) guaranteed replacement cost coverage will avoid uncomfortable litigation. Again, more trouble waiting to happen.

* In certain territories, new business for a particular commercial occupancy type is booming because of adverse selection and they don’t even know it. Other carriers have already walked away from that market segment because of the high risk.

* Chronic underinsurance in commercial property policies has led to a high frequency of co-insurance penalties against clients, resulting in depressed premium revenue, legal exposure and a diminishing reputation.

* One surprising discovery is that agricultural business for dairy and poultry occupancies turn out to be extremely profitable despite a common misconception that they were problematic.

* On the claims end, they don’t know 40-foot dumpsters are being used for debris removal much more frequently than necessary.

* They also don’t know that while 80% of claims paid include acceptable painted drywall per sq ft costs, 10% of the drywall costs average 23% above acceptable norms.

This list can go on forever so I will stop here as the point has been made.

Claims analytics can uncover individual items and groups of items that fall outside of the norm and then uncover common characteristics that permit the insurer to better control costs and avoid claims leakage.

HOW DOES IT WORK?

The analytics process starts with as much information as the insurer can provide. Ideally, information is stored electronically at a very granular level. We as the analytics service provider can then go ahead with the analytics investigation using our own additional information including:

* historical cost data on materials, components and labour for all postal codes up to the present time. This is needed to accurately calculate reconstruction costs and to estimate correctly partial loss repair costs at a local level;

* additional industry information on each property;

* historical loss data on the property and on surrounding properties for underwriting purposes;

* detailed historical claims records, which are used to identify claims severity drivers and to aid underwriters with risk selection;

* geo-coding postal addresses into latitude and longitude coordinates;

* geo-coded Canadian hazard data such as brush fire zones and distance to coast;

* industry wide property econometrics and demographic information to serve as “norms” for comparison purposes; and

* industry wide underwriting and claims best practices to serve as “norms.”

Not all of this information may be available at the outset of an analytics study. But it will become more readily available as information collection and archiving practices improve. At the same time, the accuracy and content of industry-wide databases is increasing in concert with today’s information explosion. Such data can be used to properly assess and underwrite the risk, manage losses and uncover hidden trends that may feed back into the underwriting and claims processes.

Once the carrier’s detailed information reaches a statistically valid level, it can then perform analytics to uncover hidden trends. Underwriting opportunities include:

* Uncovering individual policies and groups of policies with common traits and that are seriously underinsured.

* Identifying policies or groups of policies that have a high hazard exposure previously unknown to the carrier.

* Identifying policies that have incorrect or missing information. By example, many policies are placed in the wrong rating territories. Geo-coding can identify these.

* Performing market segmentation analysis based on property size, style, age and location to uncover strengths and exposures by segment.

* Comparing against industry norms to see how the carrier stacks up against the competition.

* Comparing point of sale performance in various ways to improve productivity and to better balance the book.

Within the claims area, opportunities include:

* Identifying individual claims cost items that deviate from validated estimates and from industry norms.

* Being able to compare individual claims adjusters and offices against each other.

* Identifying claims severity drivers that consistently escalate the cost of claims.

* Summarizing losses by market segment and feeding this information back to underwriting so that pricing, products and marketing can be adjusted.

* Identifying best practice improvements to ensure claims are settled accurately first time.

CONTINUOUS IMPROVEMENT

Analytics and the improvement projects that result from it do not really have an end. Best practices help guide us from one opportunity area to the next, allowing an insurer to improve the business constantly. At the same time, the new information uncovered by analytics can be fed back into the underwriting and claims process. This continuous exchange of findings, allows both sides of the business to improve in concert with each other and because of each other.

FLASH FORWARD ONE YEAR

It is one year later. The board of directors of ABC Insurance has once again reviewed its year-end financials and is now justifiably pleased. After a year of using analytics as a guide, the loss ratio is down to 56% and the combined return has increased to 15%. But the really good news is that the board knows these trends will continue; they now have a much better understanding of the details of their business. And as a consequence, they are able to make sound business decisions through analytics. Congratulations are truly in order.