Data clearly matters in pharma. For the fourth year in a row artificial intelligence (AI) and Big Data ranked as the most impactful technologies in the industry, according to GlobalData’s “The State of the Biopharmaceutical Industry – 2023” report, which includes feedback from 198 companies. On top of that, the life sciences analytic market is expected to surpass $45.1 billion by 2030, according to ReportLinker. But what does all of this investment in data mean to marketers?
“The axiom ‘Most people use statistics like a drunk man uses a lamp post; more for support than illumination’ is especially applicable in today’s era of nearly unlimited data,” says David Marks, PhD, MSc, VP, Medical Director, Elevate Healthcare. “Increasingly powerful computing and AI can explore the relationships between hundreds of thousands of variables simultaneously. However, correlation does not always equal causation between variables. A statistically significant association of two factors does not necessarily imply real-world applicability.”
So, how can marketers get the most out of all of this data? And how can they determine what data will lead to the most meaningful change for their campaigns, healthcare professionals (HCPs), and ultimately patients? Dr. Marks says it helps to have the right people in place.
“Having people who are hypothesis-driven thinkers is the fundamental beginning of data analytics for marketing,” he explains. “People of this type will create some of the most innovative analytics and approaches to accept or reject propositions. To improve or even discover new analytical data metrics, begins with people who are hypothesis-driven.”
Fortunately, PM360 has access to such people—our readers—and we asked them about the hidden gems that can be mined from data they have access to or how to better find key insights.
Data with Unrealized Potential
“Data that identifies HCPs who recently saw patients who visited a brand’s website is an impactful source that is often overlooked,” offers Sarah Caldwell, General Manager of Veeva Crossix Analytics. “This can help improve biopharma brands’ marketing approach as they can use this privacy-safe data and reach out to HCPs with support materials and timely information when they are more receptive. This not only strengthens the relationship with the HCP, it also can help connect patients to the information they are already searching for.”
Another way marketers can better understand their audience is through customer sentiment analysis, according to Alex Kooluris, SVP, Management Supervisor, ConcentricLife.
“By mining unstructured text data from sources such as customer reviews, social media interactions, and chat logs, companies can gain invaluable insights into customer emotions, preferences, and pain points,” Kooluris says. “This emotional intelligence can revolutionize marketing strategies and product development, enabling companies to tailor their offerings more effectively and empathetically.”
It can also help companies understand what factors outside of a clinic are most impacting a person’s health, which is something Christine Lee, Head of Health Partnerships at AnalyticsIQ, believes can be especially valuable.
“Predictive consumer and social determinants of health (SDOH) data are a treasure trove of insights,” Lee explains. “This unique blend of data sources is often overlooked in conventional marketing, but when combined with patient or medical information, SDOH data offers a granular understanding of an individual’s behaviors, motivations, and potential future actions. By tapping into these insights, healthcare organizations can craft campaigns, treatments, and messaging that resonate deeply with their populations’ real-life challenges and aspirations, fostering genuine connections.”
Myla Maloney, Chief Growth Officer, Premier, Inc., also sees value in other unstructured data sources such as clinicians’ notes and pathology reports. While this type of data used to be hard to analyze, natural language processing (NLP) is making it easier to quickly identify potential risk factors and clinical signs and symptoms that could predict subsequent disease development.
“One life science corporation leveraged real-world data to conduct a study based on a hypothesis that people were getting lung scans (e.g., chest X-rays) that may have shown incidental pulmonary nodules (IPNs) but were not followed up on for a variety of reasons,” Maloney offers as one example. “After applying NLP technology, the corporation was able to identify 152,000 patients with IPNs in a population of six million patients in New York and flag them for follow up, bringing them back into the health system before potential lung cancer progression.”
Uncovering More Insights
While Nataraj Dasgupta, VP, Analytics, RxDataScience, a Syneos Health company, agrees that NLP has advanced in recent years, he says that most enterprise search solutions used to explore datasets are still largely superficial. But, that is now changing thanks to OpenAI’s release of GPT3.5/4.
“Using reward models, proximal policy optimization, and other advanced AI techniques, OpenAI showed that language models can semantically interpret text and respond with human-like intelligence,” Dasgupta says. “Additionally, newly launched vector databases made it simple to store semantic representation of data known as embeddings, which coupled with retrieval-augmented generation made it easy to find documents relevant to a query from massive data assets in seconds. Marketers can now use generative AI models to mine a treasure trove of multi-modal data—ranging from text to video—to reveal key insights that have remained out of reach.”
Even in a world that is overloaded with data, to get the right insights marketers may need to collect more to answer their most pressing questions, such as: Where are most customers dropping out across the customer journey?
“To answer this, marketers first need to ensure they have data at every step of the journey, from clicking an ad to completing a purchase,” explains Jeremy Helgeson, Senior Analyst, Tagging & Visualization, Spectrum Science. “It’s easy to focus on the key points in the funnel (impression > click > conversion), but by collecting and mapping data at every step, marketers can leverage more nuanced data that sheds light on user experience or audience preferences, which can have significant impacts on a brand’s overall marketing strategy.”
It’s also crucial that marketers—and organizations—are de-siloing their data and making sure it is all working together. As Jenny Baban, SVP, Customer Experience Management, CMI Media Group, explains, “marketing campaigns do not operate in silos, and similarly, datasets should not either.”
She continues: “Approaching data more holistically starts with a strong taxonomy consistently applied across all datasets. This will ensure variables from many sources can be brought together for analysis. Secondly, it’s imperative to develop a clear measurement strategy in advance and define a test/learn plan which outlines how available data sources will be used together to answer key questions. This holistic perspective enables us to unlock data’s hidden potential and provides a more comprehensive understanding of performance.”
