The Age of Big Data and Agencies – Not

This is the age of big data. The concept is to collect all kinds of data and then use various statistical techniques, usually some sophisticated form of regression analysis, to identify correlations between the data, trends and opportunities. The data has to be big because searchers are seeking data relationships that exceed a human’s ability to identify, because either the data point is minute, the relationship defies common sense, or there is so much data it just overwhelms a brain. The information can then be used to entice people to buy things, vote a certain way, or perform better, etc., through a variety of means. It is absolutely fascinating.

Of course, the key is having quality data, and this is a major issue for agencies for many reasons.

The first reason is simple: If poor or wrong data is input into an agency’s management system, something is likely to go wrong.

Examples include CSRs having to take extra time to understand what is happening on an account, fixing the data, backing out data, and sharing the fixes with the producer, accounting, carriers, and/or clients, all of which reduces productivity and efficiency.

Another example is that bad data can cost agency money. The wrong commission amounts either input from the carrier (i.e., entering 11 percent instead of 12 percent) or paying producers too much. Every agency owner always tells me that paying producers too much is impossible, but my firm has been collecting information on producer compensation for 20 years and comparing their compensation to their supposed commission splits. At least 50 percent of the time, the producers’ actual compensation differs from what they’re supposed to be paid. The majority of time, they’re paid more.

Another example where bad data costs an agency is an errors and omissions (E&O claim. Bad data is a common cause of agencies losing E&O claims.

A common example that involves extra work, less commission and E&O exposures is when CSRs enter the wrong writing company on brokered business. I see this mistake often, and not just once or twice when looking at a book of business.

Bad data in agency management systems is significant and expensive, and I don’t see it getting better based on the data I am seeing from agencies around the country.

Bad Data and Benchmarks

Bad data also negatively affects agencies through the benchmarks that consulting firms use. These firms are given information by agents, and if the data in the agents’ systems is poor, then the benchmarks will be lacking. (The bad data provided does not include the cooked-up data some agents give to these firms so they can look better and “win” recognition.) This is one reason – although not the only reason by any means – that agents often look at these benchmarks and wonder if they are correct.

Benchmarks can be useful if they are accurate representations, although the best benchmarks by far are milestones on a journey of constant improvement.

I have seen agencies wrecked trying to meet these faulty benchmarks. If agency owners/executives understood the quality of data going into the benchmarks they would never use them or they would use them more generally. It is not even that the ones using the benchmarks have better data and think everyone is up to their standard. Most often, they do not understand how poor their own data is.

Poor Data Threats

Managing an agency using poor data is an awfully brave or foolish endeavor, and yet most every agency owner I meet thinks their data is far better than it is. Likely close to 50 percent of the agencies that hire my firm cannot provide an accurate, even within 10 percent, count of their policies and accounts for the past three years. This is an extremely basic and important metric, and almost always the owner thinks he or she can provide this data. But owners can’t, because they don’t use their agency management systems correctly. Account and policy counts are simple examples but exemplify how endemic the data problem is.

The opportunity here is significant, and the threat if agents don’t join the times is significant. The opportunity to use big data is fantastic because agencies have big data. Almost no agency owners have enough mathematical education or knowledge to complete data analysis themselves, but hiring someone to do it for them using their data will provide the results they will know how to use. But it can’t be done without quality data.

The threats of not fixing the data problem are significant.

Big firms, insurance and non-insurance, are already mining higher quality insurance databases, and they’ll use this data against agents in the near future. Other firms with high quality data may have built a new revenue stream by selling their data en masse.

Bad data also usually encompasses data agencies are not supposed to keep or need to keep. Because agencies do not have governmental immunity for data breaches and can instead be prosecuted for a breach (makes sense that a crook breaks into your property and you can be prosecuted for inadequate barriers, right? Only attorneys and government employees can think this stuff up.), the less data an agency has, the better within reason.

This means purging old data. Purging old data also improves the quality of big data because the data is more current and the agency has more resources with which to manage the current data.

Another threat is cost. My studies show that agencies with better data are also more efficient, often by 10 percent to 30 percent more efficient. The opportunity is that with so much more efficiency, agencies can invest much more in organic growth and acquisitions. The threat is that if it cost your agency 10 percent to 30 percent more to service accounts, the ability to compete in a soft market will be minimal.

Another threat is thinking you can do this yourself.

Big data and regression analysis may seem easy on the surface, especially if you see “Moneyball.” But the work and expertise required is significant. I find some agencies going through the motions thinking they’re being adequately thorough, but agencies, even big ones, rarely employ people with the specific education required to use this data correctly.

A common theme throughout is quality of data over quantity of data, even in the age of big data. Junk in is still junk out, and even quality in but used incorrectly is still junk out.

The opportunity to use big data is here. The only question is if you and your agency will join the era.

Topics Agencies Data Driven Professional Liability