Advancing Your Advanced Analytics Capabilities

Use cases for advanced analytics

Now that we’ve talked about what advanced analytics is and why it’s important, let’s look at three examples of it being put to use. One in finance, one in healthcare, and one in retail.

Finance

Every minute of every day, millions of credit cards are being swiped around the world. In order to identify potential cases of fraud, credit card providers deploy ML models to help determine irregularities based on previous purchasing habits, locations of purchases, even times of day purchases are made.

Using the output of these models, credit card providers are able to know when a customer has had their information stolen before the customer does and can take preventative measures — all without slowing down the flow of transactions.

Healthcare

Redapt client Zelis is a healthcare technology company that, among other things, processes out-of-network claims between doctors and insurance providers. Given the sheer amount of claims, the company needed the ability to predict and isolate claims that were potentially fraud, waste, or abuse (FWA).

To help them with this problem, Redapt built out a framework for the company to efficiently sift through data to identify and fingerprint historic costs related to procedures. This framework was flexible enough to be modified with new information as it arrived.

By leveraging the public cloud, Zelis’ data scientists were able to run advanced analytics models to not only flag potential FWA claims, but accelerate the processing of non-fraudulent ones.

Retail

Online personal styling service Stitch Fix connects customers to clothes by utilizing advanced analytics to develop algorithms based on popular combinations and styles.

These algorithms then intelligently style entire outfits based on customer interests, seasonality, fit, and so much more. This way, all customers have to do is simply tell the company a little about their likes and dislikes. Then, before they know it, they’ll receive outfits based on that information.

The benefits of this approach are twofold:

• Stitch Fix is able to serve more customers than traditional, in-person styling services.
• Customers are able to have an experience normally reserved for the more affluent.

The most common challenges in adopting advanced analytics

It has been estimated that nearly 90% of advanced analytics projects like ML fail to reach production.

One of the common driving factors behind this failure rate is the disconnect between IT and data science.

IT tends to stay focused on making things available and stable. They want uptime at all costs. Data scientists, on the other hand, are focused on iteration and experimentation. They want flexibility and performance.

Beyond internal disconnects, many of the roadblocks Redapt has seen enterprises encounter on the way to adopting advanced analytics come down to data. Specifically:

Traditional organizational power structures

In reality, technology is hardly ever the core blocker to transformative analytical change. Most of the friction comes from antiquated organization parent-child structures that make cross-functional and agile analytical transformation almost impossible to implement effectively.

Data quality

Results from AI and ML models are only as good as the data used to arrive at them. Without proper data processing and pipelines, actionable data will be hard to come by and your models will fail.

Data organization

AI and ML need to know where to look within data sets. A failure to properly organize data with sourcing and tagging will only send advanced analytics models into a sea of information with no sense of direction.

Data storage

Advanced analytics requires access to a substantial amount of data to be effective. While the public cloud has made storing the necessary information more cost-effective, if your organization relies on cloud-native tech stacks on premises or at a co-location, you can easily misjudge the amount of storage you need.

Data governance and security

The more data you make available throughout your organization, the more business objectives you are able to achieve.

This availability requires following proper governance over your data in order to keep it safe and ensure only specific data is available to specific teams that need it. The challenge, however, is putting governance rules in place that every team in your organization follows.