A majority of my conversations with appraisal supervisors fromauto insurance carriers include a discussion on appraisereffectiveness, which makes sense since these are the individualsthat can have a significant impact on claim cost, cycle time andcustomer satisfaction. Most frequently, I hear that the obstacle tounderstanding the true effectiveness of appraisers is that the datacan become too cumbersome — some insurers can review 20measures or more to gain the insight needed to track and measureindividual performance. Scaling this activity across theirappraisers to get a peer comparison can be an extremelytime-consuming exercise, which means it doesn't always get done,resulting in organizational blind spots and missedopportunities.

Working for CCC, which contains a vast warehouse of claims dataand provides software and services to more than 350 auto physicaldamage (APD) insurance carriers, my colleagues and I started toresearch how we could help appraisal supervisors increase theirvisibility into staff performance. Our goal was to help enable anactionable snapshot through the smart and simple application of keyclaims-related data based on important metrics identified by theinsurance carrier. The result of this work is a data model thathelps identify individual and peer-to-peer appraiser performancewithin an organization using a set of carrier-specific metrics.

What can a simple data model really tell you?

When we started thinking about this, we didn't have a set numberof measures in mind; our guiding principles were to offer somethingsimple and actionable. We researched other similar models to seewhat lessons had already been learned in taking this type ofapproach. The NFL passer rating, which compares quarterbacksweek-to-week, over the course of the season, and over their careers— caught our attention.

In place for more than 30 years, the NFL passer rating uses fourvariables: completion percentage, yards per attempt, touchdowns perattempt and interceptions per attempt. Each is weighted differentlyand an average is established and a rating is assigned. The QBs,coaches and general managers all know this system and can thereforecommunicate using a standard language and make informed decisionsbased upon individual objectives and strategies.

The simplicity of this formula has drawn some critics, but ithas stood the test of time and has been referenced as being among the most indicativemeasures of wins and losses.

Can a similar concept be applied to APD insurancecarriers?

We think so.

Just as each NFL team has its own strategies and playbooks, weknow each claims organization approaches claims differently.Through our work, however, my colleagues and I have found that theexpectations for appraisers within an enterprise tend to beuniform. Sure, there may be regional differences, such as vehiclemix, but a management team usually has a universal philosophy onhow the field staff should approach assigned claims.

Since these uniform expectations exist, we started analyzing thedata in CCC's warehouse, which includes more than 170M claims worthof data. My colleagues and I focused on data that could helpsupervisors quickly gauge the performance of individual appraiserson a standard scale within an organization for the purpose ofdriving more effectiveness across their teams. Staying with oursimplicity theme, we observed that just a few data elements canprovide meaningful insight into performance. While thesecharacteristics could be applied to a broad set of data, we chosefour broad measures that a hypothetical insurance carrier could useto get an effective snapshot of their staff's performance based onmetrics that it cares about — speed and accuracy. We focusedour analysis on those activities that were directly in theappraisers' control:

• Repair percentage of labor amount.
• Parts utilization, amount and count.
• Cycle time (assignment sent to estimate sent daysaverage).
• Supplement percentage and average supplement amount for claimswith supplement.

Different measures could easily be substituted for the ones inthe hypothetical insurance carrier's model based on the insurancecarrier's specific needs. Depending on the insurance carrier, theremay be priorities that require the use of different metrics. Thekey is to keep it simple so that once the model is set up, the datacan easily be refreshed and results reviewed.

For this hypothetical insurance carrier, each measure isweighted differently but similarly, and the average of thesemeasures becomes the appraiser rating. We decided on a scale of 1to 100, with a rating of 50 reflecting average performance. Thisnumber rating makes it a simple exercise to identify top and bottomperformers, so focus can be placed on areas where improvement maybe required.

From a practical application perspective, and using the measuresoutlined above, a manager may find that certain appraisers have asignificantly larger average supplement amount than their peers.Assuming that appraiser sees similar vehicles to the rest of thestaff, this may be a coaching opportunity.

Looking at the data

For example, let's say the hypothetical insurance carrier hadthe following averages for their key performance metrics:

• 36% of repair percentage labor amount
• $1,027 in parts cost per claim
• 6.7 parts replaced per repairable vehicle
• 2.6 days cycle time (assignment to estimate complete)
• 37% for supplements
• $1,400 average supplement amount for claims withsupplements

Let us assume that an appraiser for the hypothetical insurancecarrier had the above numbers. Their score would be a 50 out of100. Depending on how this appraiser's numbers compared to hispeers at the insurance carrier, the manager would identify areasfor this appraiser to focus on in order to align with the insurancecarrier's expectations. If the overall score for this appraiser wasin the top half of his peers, the managers may decide to put moreemphasis on staff members that ranked in the bottom quartile of theappraisers. Alternatively, the manager may decide to focus onindividual measures amongst the appraisers. For example, thisappraiser's supplement frequency may be high compared to his peers(45% versus 37%). The manager may focus more on what is leading tothe reopening of the file and examine the appraisers daily workhabits to suggest changes.

In today's data-driven world, measuring performance isessential. Direct Repair Program scorecards have been effective inidentifying how repair shops are preforming, and this similarapplication can allow for a simple and quick look at your ownstaff.

Interested in learning more or setting up your own simple model?Contact me at [email protected].

Bart Mazurek is director of Solutions and Consulting atCCC Information Services Inc.