This is the fourth in a series of articles by Valen Analytics looking at the hurdles that insurers must overcome to effectively implement and gain value from data analytics programs.

Every tool has its ideal purpose and a line at which it’s overused. It has been said, “To a man with a hammer, everything looks like a nail.” We all rely on spreadsheets to provide critical information, but in predictive analytics, controlling versions of spreadsheets can create a minefield.

A staggering 50 percent of insurers use spreadsheets to deploy predictive models to their workforce. When that happens, formulas can easily break and there’s no easy way to incorporate models into normal workflows, which means little to no insight is available on how or if they were used in decision-making. Spreadsheets are also ubiquitous, inexpensive, and almost everyone has some knowledge of how to use them. In an insurance environment where IT backlogs are constantly being reprioritized, insurance industry leaders understandably rely on spreadsheets to handle the task at hand.

Across the insurance industry, there are a number of reasons that spreadsheet use has become a crutch, and they must be overcome in order for carriers to get the most value out of their predictive analytics initiatives.

Good Is the Enemy of Great

Spreadsheets can be effective as a bridge between tools and can also be great for prototyping predictive models. However, simple prototyping shouldn’t be confused with pushing a model into production. In many instances, insurers continue to rely on spreadsheets simply because the solution is “good enough.”

Many analytics professionals have built a spreadsheet-based solution that works for them as an individual. However, as documents get shared out, issues ranging from human error to version control become more common. After all, while one person has created something that aligns with their specific role and approach to work, individualistic shortcuts don’t scale.

The Expense of Replacing Spreadsheet-based Workarounds

Many insurers seem to believe the cost of implementing a new system outweigh the benefits it will offer. However, they’re often not taking the full cost of their incumbent approach into consideration.

In the context of spreadsheets, insurers should quantify the amount of time employees lose moving between systems and manually inputting data. Manual data inputs in a spreadsheet should sound off alarm bells, as those inputs can’t be checked for errors or inconsistencies among users. Nor will you ever know the number of times a formula in Excel has become corrupt, and the downtime or loss of productivity that follow.

If this is a spreadsheet used to inform underwriters about risk quality and pricing, how can you have confidence in the accuracy of decisions being made?

Growing insurers must also factor how long it takes to get a new employee comfortable with an existing workflow. Since Excel isn’t a standalone application, it is often overlooked when creating appropriate documentation for specific uses. This presents significant opportunity for trial and error. Using tools that are easily available can seem like the best way to move forward, but it can also be incredibly expensive in time, both for training and maintenance.

Why Add to a Backlog?

Many IT departments across the insurance industry have a considerable backlog of projects and development, causing insurance leaders to question whether they should prioritize technology workflow improvements. After all, initiatives like improving user experience for policyholders should, theoretically, have a bigger effect on the bottom line than reworking internal processes.

Many managers even believe that their use of spreadsheets is alleviating some of the backlog and allowing the IT team to focus on matters that will have a bigger impact. Again, this group is failing to take into consideration all of the costs and business risks associated with spreadsheets.

Overcoming These Objections

For insurers that are looking to move beyond their reliance on spreadsheets, it’s important to quantify the overall ROI for embedding tech-based solutions into the typical workflow. Many have come to realize that building an entire solution in house will ultimately take too much time and effort, which has created an opportunity for insurers to improve their agility through the use of an external partner.

APIs also offer the ability to build data initiatives into workflows that seamlessly offer relevant information and actionable insights at the point of decision. This increases the likeliness that employees will actually use the information they’re being presented. It can even include tracking mechanisms to ensure usage, whereas there is no way for insurance leaders to know how often formulas in spreadsheets are being utilized.

Finally, because API-based solutions are lightweight, they can fit in without disrupting the IT team, helping to overcome the backlog.

Beyond APIs and workflow, carriers should consider data governance and model monitoring, both of which are critical in obtaining and maintaining regulatory approvals. There should also be an awareness of performance. After all, a solution that is suffering from downtime obviously can’t be used.

Spreadsheets have their uses. From organizing and moving data to back-of-the-envelope model testing, their low financial cost and near ubiquity ensure that they aren’t going away any time soon. However, spreadsheets don’t work well for deploying analytics initiatives at scale, and they prevent insurers from gaining real, tangible and measurable value. Insurance leaders must create a vision, help IT prioritize and overcome backlogs, and find the right modern partners and tools.