Human Error – and How to Avoid It During the Pricing Process

Photo by Poppy Thomas Hill from Pexels

One would think that with all the digital advancements of the past few years, human error as such has been obliterated from the rating/pricing process. But when it comes to pricing, it still plays a powerful role. One that carries a significant business cost. The ambitious goal to minimize (or, ideally, get rid of) such mistakes, must start with identifying their cause.

Common Errors in the Pricing Process

The most common errors in the pricing process boil down to technical inefficiencies and data inconsistencies that surface when rating modifications are made.

Ideally, rating modifications should carry over across a company’s systems. A change made in one system should carry downstream and be represented elsewhere. Similarly, rating modifications need to be traceable – right back to one clear source.

Not knowing who made a change, where, and when, opens up a smorgasbord of issues for professionals all over the enterprise. From future modifications to governance and compliance, challenges to reconstruct pricing decisions and why and how they were made harms the whole organization.

Behind Errors: Disjointed Systems

Many insurers and banks use multiple systems to generate rates, often involving multiple people across departments and several disjointed processes. Moreover, oftentimes, these processes are, at least in part, still manual.

And the problem is not only with mistakes and the steep price insurers pay for them. Disjointed systems are rigid by nature, their upkeeping takes effort and they are not scalable.

Rate sheets developed in one system, then transferred manually to an Excel spreadsheet, edited, and then uploaded (again, manually) to a front-end or quoting system are a breeding ground of errors. Even under ideal conditions.

The Business Cost of Errors

Errors like the above always come with a price. Many times, an actual price tag. Other times a long-term setback originating from an error-ridden system where decisions cannot be traced back and audited.

Insurers might end up with quality assurance issues, leading to fraught decision making and delayed time to market. Things look different when the entire pricing process is covered by one system.

What Happens When Pricing Happens Under One Roof?

Imagine a price committee meeting. A price model is presented, and a committee member asks, “what if instead of 20%, we reduce this figure to 15%?” or, “what happens if we increase the rate for segment by 5%?”

If the company in question used a disjointed, manual pricing process, it would take a long time to answer these questions. Regardless of how long exactly, committee members would not have answers to their questions during the meeting itself. At least not without cutting corners and compromising quality.

But today’s business environment does not offer the luxury of time. Pricing decisions are urgently needed, and it is possible to answer these questions on the spot, with an end-to-end rating system. When a company uses one single system for pricing, an end-to-end rating engine, the types of checks, changes, and challenges that arose during our imagined pricing meeting can be done instantaneously—and with minimal errors.

The Analytical Rating Engine

A singular analytical rating engine enables the seamless incorporation of analytical & ML models and business rules and takes care of all the pricing decisions within the enterprise.

Analytical rating engines enable data management, modeling, pricing sophistication, and deployment of rates in real-time—all of which allow companies to dynamically speed up their time to market by implementing rate changes in a quick and agile manner to any point of customer interaction.

Analytical Model Execution – Without the Mistakes?

Selecting the correct pricing strategies and then executing them efficiently is key to rating engine success. An ideal rating engine should allow insurers to define rating parameters—for instance, quote type, segment, and channel—and easily apply rating rules and rating logic.

After developing or uploading the right model, insurers should be able to test it before deployment.

Once the model is ready for production, it goes out for approvals and is then deployed to the online system, ready for large-scale distribution.
Finally, all of this should be auditable and strictly governed, to meet local compliance standards.

How Analytical Rating Engines Minimize Errors

Analytical rating engines work on two levels to minimize error: governance and controls, and automation. Together, these provide the insurer with full visibility over their pricing process.

Governance and Controls

Working on a single system means that all actions taking place within that system comply with its rules and can be tracked and overseen. A well-designed analytical rating engine will, for example, block a user from deleting data tied to an existing model, enables sophisticated feedback loop systems for more sophisticated pricing decisions. But it also makes sure, that a single change in a model or rating sheet will have ripple through the entire system and automatically update every calculation, every quote it is connected to.

Automation

If any action needs to be systematically repeated throughout the enterprise or to facilitate the pricing process– e.g. a systematic, enterprise-wide pricing update following the change in one of the rating factors -, integrated automation allows users to apply it throughout the system.

These automated processes – like a sophisticated system of conveyor belts – relieve humans of manually applying changes throughout rating sheets, systems, and more.

A Thing of the Past

An end-to-end analytical rating engine consolidates all the information insurers and banks need to price more effectively in one place. Moreover, a well-structured system that enables the operationalization of analytics & ML models gives way to more sophisticated pricing decisions, that can be delivered in real-time to the right touchpoint. And most importantly, human error—which we now have the technology to overcome—can become a thing of the past.

As the company’s Chief Analytics Officer, Reuven Shnaps leads the Earnix Analytics organization. In his role Reuven oversees Earnix product roadmap management and directs the company’s analytics strategy, as well as research and delivery for Earnix customers worldwide. Reuven has spent more than 20 years developing and analyzing advanced statistical and economic models. Over his 10+ years at Earnix, he has gained extensive experience in crafting analytical solutions, and managing and implementing pricing & analytical projects for leading financial institutions around the globe. Prior to joining Earnix, Reuven worked in the economic and statistical consulting group at Deloitte & Touche in the United States, serving corporate clients across multiple industries. He holds a PhD and a master of arts in economics from The University of Pennsylvania and a master of arts in business economics and a bachelor of arts in economics and statistics from Bar Ilan University.