ZS Associates (ZS) is a global professional services firm that helps companies develop and deliver products for their customers. The Evanston-based company began operations in 1983 and currently has over five thousand consultants in the Americas, Europe, and Asia. ZS experts work side by side with clients, leveraging analytics, technology, and strategy to create solutions that work in the real world. ZS is passionately committed to helping companies and their customers thrive, which drives better business results.

Big data and analytics have become integral to the life sciences industry, where it is recognized as a powerful tool to help companies innovate, transform, and stay ahead of the competition. A big portion of the value of big data for these companies is related to developing better drugs, but there are commercial opportunities as well. In particular, the customers of these companies are becoming savvier with their consumption habits as healthcare information becomes readily available, creating the need for more personalized experiences. This has created a huge opportunity for them to leverage the billions of anonymous health records that they can access to match the individual needs of customers and patients in the healthcare ecosystem.

While the transformative impact of big data for life sciences companies has been widely discussed, an often-overlooked problem is how companies handle, analyze, and secure billions of anonymized patient records to derive meaningful insights. According to ZS, larger pharmaceutical companies have largely built out the infrastructure and expertise to manage their big data needs in-house or have made disruptive moves to the cloud already. ZS has been working with many of these companies on this revolutionary journey.

However, ZS observes that many of the small and mid-sized pharmaceutical companies have yet to attain this level of technological maturity and are still learning the ropes of simplifying their data infrastructure to inform business decisions at scale. One such case was a mid-size pharmaceutical company that ZS has been partnering with for some time to develop strategies to launch a product. To do this, ZS has been harnessing 1.5 billion anonymized patient claims data in conjunction with other data sources to develop market opportunity assessments and generate actionable recommendations for product launch.

For this pharmaceutical company, ZS was working on an on-premise infrastructure, which satisfied its needs for traditional business intelligence reporting to date. However, a set of new drivers forced ZS to think differently for this situation. Firstly, the product launch date was around the corner, giving rise to the need for significantly faster insights. Secondly, data volumes were increasing and configuring an on-premise system set up for peak performance was becoming more expensive. Lastly, ZS wanted to scale up the use of machine learning (ML) algorithms to identify the right customers for launch, which it found limiting with the existing IT infrastructure setup.

After considering these factors and consulting with in-house experts who had been on similar journeys with larger pharmaceutical companies, ZS decided to use the following AWS stack – Amazon Simple Storage Service (Amazon S3) for data storage, Amazon Redshift for data processing, and Amazon Elastic Compute Cloud (Amazon EC2) for machine learning algorithms.

One common, but important question that the pharmaceutical company had in mind was “Who are the right customers to educate at launch?”. To build the predictive algorithms to identify customers likely to adopt the new product, ZS used ML techniques and iterative decision trees to run hypotheses. Whereas in the past, it took 1 month to build this model, Amazon Redshift allowed the ZS team to do this in 1 week due to its snappy performance and ability to perform ad-hoc analyses on data stored in Amazon S3. This allowed them to improve the quality of algorithms and quickly learn from the dynamic changes post launch.

The enhanced processing time from running Amazon EC2 instances also allowed the team to run more hypotheses. Previously, it took the team 9 to 12 hours just to run the model on premises, but with Amazon EC2 instances, this was reduced to an hour. Moreover, the size of the dataset meant that the team previously had to split the data into 4 different sets and run individual models on those data sets, before combining those learnings to the next dataset. With the new set up, the team could run the ML algorithms and decision trees over a single dataset in the cloud.

The ZS team also saw a 50% reduction in costs from moving its data warehouse to the cloud – mainly because of the flexibility of archiving data on S3. Had the team stuck with its on-premises data warehouse, the costs would have doubled due to additional data that needed to be stored. Moreover, the pay-as-you-go approach for pricing meant that ZS no longer had to overprovision its servers. This meant that it was able to spin its Amazon Redshift clusters during the day time and shut them down overnight, which resulted in ZS only having to pay for 1TB of data instead of 10TB. For ZS, the ability to deploy on-demand storage and on-demand processing were the biggest advantages of data warehousing on the cloud.

Big data analytics is here to stay and become even more integrated in the functioning of life sciences companies. The key problem to solve is to get all the data in a single place and make it simple for the right people in the organization to get access to the right data to make the right decisions. Once the right data, technology and people come together, innovation can be expected to accelerate at a transformative pace.