The race is on among life science companies to use artificial intelligence (AI) to turn big data into big competitive advantages. Early applications suggest that these tools can help decision makers find the right audience for their therapies, forecast future sales more accurately and run what-if scenarios to discern the optimal paths to market.
Yet there are several hurdles that stand in the way of immediate results. First, the increasing volumes, velocities and varieties of data from both internal and external data sources are often housed in multiple systems and applications throughout the enterprise. Second, data models that are most useful to AI require extensive experience and access to historical customer and industry data to properly build.
Another significant roadblock is very familiar to data scientists yet often overlooked by executive decision makers: preparation. According to a recent survey by Forbes, data scientists spend nearly 80 percent of their time preparing data for use in modeling. 1 That’s astounding, considering (1) the expense in time and human resources these processes represent and (2) the urgency companies have to generate actionable insights that might have a positive impact for their customers and shareholders alike.
Awareness has been growing about the need for data warehouse solutions that unify previously siloed data. When data is scattered across different applications, companies don’t have a complete view of the domains in which they operate. Without a process or system in place to unify it, that fragmented data can frustrate decision making and impede key business goals such as sales productivity, marketing campaign effectiveness and compliance. It is important to note the extent to which companies have been investing in data – from both internal sources and trusted third parties – to build the models needed.
An edge can be gained, however, by ensuring that the data warehouse solution a company invests in is able to cut preparation time down significantly, freeing up data scientists to engage in higher-value tasks. The way in which data is structured and stored can significantly reduce preparation time and speed time to analysis and insight. A data warehouse solution that is purpose-built for the life science industry will combine prebuilt models based on commonly used data sets with the flexibility to easily customize new models to fit unique needs and the continuously changing landscape.
The more data AI tools are fed, the more powerful they become. So a data warehouse must be easily and highly scalable to accommodate more information from additional sources without the need to disrupt ongoing work. Premium data warehouse solutions will even decouple the ability to compute from the ability to store data. This results in zero impact on computation capability when adding memory. An added benefit is an increase in productivity of a company’s data science teams while leveraging the advances of secure cloud technology.
There is good reason to anticipate that adopting artificial intelligence tools will have a measurable impact on business. To be most effective, these tools must be built on the foundation of a purpose-built data warehouse solution that can speed up the time to value of the data being prepared, modeled and mined for actionable insights.
1. Gil Press, “Cleaning Big Data: Most Time-Consuming, Least Enjoyable Data Science Task, Survey Says,” Forbes.com, March 23, 2016, https://www.forbes.com/sites/gilpress/2016/03/23/data-preparation-most-time-consuming-least-enjoyable-data-science-task-survey-says/#24c462456f63.