At Tamr, we have the privilege of working with the world’s most innovative companies to solve the most difficult data management challenges that prevent them from maximizing the value of their data – and their business. Their success not only helps their bottom line but, often times, also contributes to a greater good. A perfect example of this is a recent article from Forbes on “Biting The Data Management Bullet At GlaxoSmithKline”.
GSK has long been a Tamr client, and we believe their success is what occurs when a company truly focuses on using its data as an asset and has the foresight (as well as courage) to adopt new approaches to old data problems.
As the article mentions, GSK recognized in early 2015 that they needed to rethink their R&D data environment to remain competitive with their peers. After making Mark Ramsey the first Chief Data Officer of R&D, GSK solidified their commitment to competing on their analytical capabilities, which could not only grow their business but also provide better outcomes for patients. Mark’s vision for data within GSK R&D was to “make it easier to access and use for exploratory analysis and decision-making about new medicines”. The team wanted to find a way to to “provide analytics-ready data of all kinds across R&D in a timely manner.”
The issue plaguing the R&D organization was that years of “data was kept within silos created for particular scientists, experiments, or clinical trials”, rendering secondary analysis of it nearly impossible with existing approaches. This is the most common challenge that large organizations across all industries run into when aspiring to compete on analytics. They have no easy, effective way to integrate the data for analysis – halting their ambitions immediately.
Determined to serve up timely data across a variety of domains, the R&D team first selected the top 10 most important use cases to focus on and then rethought the traditional approach to data management. GSK knew that methods like master data management and ETL would have taken too much time and effort, and consequently sought new, machine learning-based approaches to solving their data integration challenges. Their belief that innovative approaches to data curation would be needed to succeed in architecting this Data Centre of Excellence led them to Tamr.
GSK employed Tamr’s probabilistic matching approach to combine data across the organization and across three different initial domains (assays, clinical trial data, and genetic data) into a single Hadoop-based data within 3 months – “an unheard-of objective using traditional data management approaches.”
Tamr’s product and services teams went to work using the technology to map raw data, much of which was in GSK internal formats, to industry standard formats using machine learning. “The team would feed in the source trial data, and what the target format should look like—and then let the machine go to work. Outcomes initially had 50/60% accuracy levels, and now in some domains they are at 100% accuracy”. Moreover, Tamr completed the project on time.
Tamr’s success in completing this integration is not only a testament to the power of Tamr’s machine learning and field team, but exemplifies the type of outcomes that can be derived when a customer is willing to invest in their data environment and leverage new, innovative data management approaches. GSK is benefiting from this new Data CoE and “use cases have expanded from 10 to 250” as the business is seeing significant reductions in times to get an answer to an ad hoc question. More importantly, the work should ultimately contribute to better medicine and better outcomes for those that GSK ultimately serve.