When the topic of text analysis arises, what comes to mind firstare claim notes — those long, yet unstructured records that containraw materials for claims insights.

But insurers are rife with text across the board — medical notesfor life and disability policies; property descriptions for realestate underwriting; press releases, legal notices and newsarticles for commercial policies.

The relatively recent addition of comments, complaints, praiseand opinions from social media gives unprecedented access to thevoices of customers.

This mass of raw information — by most estimates at least 75percent of an organization's data — floats under the waterline,beneath the structured tip of the data iceberg.

If you could tap into that mass, you could:

• Process claims and applications faster, by automating the mosttime-consuming portions of the review process.
• Improve your organization's ability to spot suspect claims orfraudulent activity.
• Quantify key elements of policyholders' experiences andcontinuously monitor drivers of customer satisfaction anddissatisfaction.
• Reduce risk by detecting potential noncompliance withregulations or corporate standards.

Utilizing unstructured data 

Insurance companies generate and collect thousands of pieces offree-form textual content each day: call center notes, claimsadjuster comments, emails, open-ended responses in customer surveysand social media communications. But while organizations arestoring these documents, many are not harnessing the full potentialthese rich, yet complex, data sources can provide.

Unlike the neatly structured data that sits in our warehousesand data marts, text-based insights are buried in free-form fieldsthat are often challenging to analyze at scale. However, with morestorage and processing options than ever before, and withincreasingly sophisticated analytical tools, the time is right toseize on the benefits of this resource.

Related: As insurance fraudsters get smarter, so doinvestigators and their methods

It's easy to see how much qualitative and descriptive valueunstructured data sources offer. Often, they contain the critical"why" factor that explains someone's past actions or future intent — a factor that may be missing from thetransactional or structured variables. Why did this policyholderdecide to cancel their policy? What were the conditions leading upto the claim? How was their customer service experience?

With text analytics and natural language processing, answeringthese types of questions becomes possible  — andmore importantly, on a scale far beyond what we could accomplishwith manual review. But how can we replace human intuition with amachine? There are many techniques we can apply to these textdocuments. Which we choose depends on the type and characteristicsof the data, the use case and the goals of the analysis.

Unstructured data sources can provide the "why" factor thatexplain a policyholder's past actions or future intent. (Photo:Thinkstock)

Practical examples of text analytics

One goal may be to shorten the claims and application reviewprocess. We can apply categorization to sort the claims intological bins based on the injury or accident descriptions from theclaims notes. These categories can be used to more finely focus theexisting routing patterns. Contextual extraction can pull out keyentities and data elements such as dates, names, vehicles, dollaramounts or body parts injured.

For example, the names of vendors, attorneys and other thirdparties sometimes exist only in claim notes. Extract these from thetext to create new structured fields, which can be used for bothoperational efficiencies as well as analytical pursuits. These newstructured fields become potential predictors for modelsidentifying claims which may be fraudulent or might result in highpayouts.

Or let's say you're trying to use free-form comments from acustomer survey or call center agent notes to predict an outcome(e.g., likelihood to recommend, policy cancellation, future claim).If you have sufficient training data, you might employ some of textanalytics' statistical and machine learning techniques. 

Oftentimes, incorporating text topics or clusters into aconventional data mining flow yields better model performance thanthe structured variables alone. One major health insurance carriertold us, "The verbatim are a treasure trove of useful information — we find them more valuable than all the rest of thesurvey combined."

Customer sentiment, emerging issues

With the advent of sophisticated, automated text analytics, goneare the days of waiting weeks for results to come back from an adhoc analysis; we can easily trend customer sentiment or detectemerging issues with almost no latency. Adding that element of timegives additional context to your analysis.

Are customers promoting our brand more or less than lastquarter? How about versus last week? Do those trends correlate tospecific website changes, advertising or other marketingactivities? Is there an emerging pattern in the text that may be anindication of systemic fraud?

In this spring's inaugural "Forrester Wave: Big Data Text AnalyticsPlatform," Forrester asserted that "only the enterprises thatare obsessed with winning, serving, and retaining customers willthrive in this highly competitive, customer-centric economy." Textanalytics provides a compelling opportunity for insurance carriersto listen to and serve customers in a personalized, proactive waythat was unattainable in years past. Its potential has barely begunto be tapped within the insurance industry.

For carriers and payers who want to gain a competitive advantageand achieve deeper insights into their customers, the possibilitiesare vast. The tools are available, and the storage and processingoptions are cheaper than ever. The time is now.

Related: Analytics dynamic evolving between insurers andagencies

Christina Engelhardt is a text analyticsconsultant within the Global Technology Practice at SASwith a passion for helping customers extract the maximum value andinsights from their unstructured data. 

Elizabeth Dykstra is a principal systemsengineer with SAS. For17 years she has worked with insurance companies to adopt, embraceand capitalize on the benefits of advanced analytics.