Still, pharmaceutical companies have only scratched the surface. Siloed use cases and impact stories abound, but investments have rarely led to profound organizational changes. Few have yet succeeded in deeply embedding digital and analytics throughout their organizations. Executives are growing frustrated by the lack of broader, transformational impact.
There is a way forward. Detailed analysis of existing industry leaders suggests five proven actions pharmaceutical CEOs and CDTOs can take to advance from small-scale experimentation to industrialization of digital and analytics in the next 1-2 years.
Action 1: Rethink operating models and reorient around outcomes
Companies that have begun to accrue large-scale transformational change from their digital and analytics investments have shifted to product- and platform-oriented operating models. This reallocation of resources empowers and equips high-performing, cross-functional, and autonomous teams to work continuously toward specific overarching goals rather than individual products.
Companies that are effective in making this shift structure their organizations around business outcomes and users—organizing cross-functional teams against parts of the patient or healthcare provider experience, for example—rather than around discrete technologies such as mobile apps or chatbots. Product and platform-oriented operating models have been shown to better orient cross-functional teams around desired impact and tend to result in fewer hand-offs, greater productivity, faster time to market, and improved user experiences. A forthcoming McKinsey survey of more than 50 product teams found that this way of working improves delivery speed and employee satisfaction by an average of 20 percent and customer satisfaction by 15 percent.
Early adopters of the approach have realized significant value as they reorient from delivering projects to delivering business results. A large global pharma company, for example, replaced a siloed approach to developing and delivering technology projects with cross-functional teams that were aligned around outcomes and empowered to pursue those outcomes as they saw fit. The new operating model reduced testing time by 50 percent, improved product delivery speed by more than 20 percent, and boosted customer satisfaction by 50 points in some cases. It also created the foundation for greater agility in developing new digital products.
In conjunction with the shift to a product or platform operating model, companies can see benefits from making complementary changes to their vendor agreements and technology choices. Many, for example, consolidate their vendor portfolio to include a few key strategic vendors and move their commercial vendor arrangements from activity-based to outcome-based pricing and performance management. They might shift their application support and maintenance deals from pricing based on volume (for example, tickets serviced) to pricing based on outcomes (for instance, uptime), thereby bolstering application performance over time. We've observed that pharmaceutical companies can cut costs by an average of 10 percent with vendor consolidation, while outcome-based contracts can orient the strategic interests of vendors and customers around more strategic aims, enabling greater alignment, flexibility, and innovation.
Action 2: Dive into DataOps to drive innovation
Data access and quality have been major inhibitors of digital and analytics transformation in the pharma industry. It's difficult to identify what data exists, access it quickly, monitor its use, and ferret out duplicate or siloed data sources. In many companies, manual data processes persist, making any advanced analytics at scale or speed impossible. Before pharmaceutical companies can achieve digital and analytics transformation, they must embrace data transformation.
Enter DataOps, a set of collaborative practices, capabilities, and tools that can standardize and automate data use to improve quality and reduce the cycle time of advanced analytics. A DataOps approach can enable pharmaceutical companies to extract more value from their data and more quickly advance and scale their digital and analytics initiatives. In a company that embraces DataOps, business users can more readily find, access, integrate, and republish data across the enterprise, a critical foundational capability for analytics applications.
Here's how it works: Cross-functional teams build a data stack that can integrate data from multiple sources. They then define standards for data governance and ownership, access management, development priorities, and data products. A DataOps approach requires new skill sets, such as data stewardship, to manage data products and ensure their quality. But once pharma companies have those capabilities, they can develop a set of automated tools, such as anomaly or data drift detection, to streamline data-engineering efforts. A DataOps engine can be the driver of ongoing digital and analytics innovation.
GlaxoSmithKline has embarked on a DataOps transformation with Onyx, a group of platforms GSK intends to use to "build a comprehensive data and machine learning ecosystem," according to the company's website, that will support its scientists in creating next-generation medicines.
Action 3: Industrialize AI with MLOps
With DataOps in place, the pharmaceutical industry has a multitude of applications for AI and machine learning, from automated indication finding to risk estimators to predicting the improvement in patient outcomes from digital health care solutions. However, like many companies in other sectors, pharmaceutical companies have typically treated their AI pilots and projects as one-off initiatives, with business units requesting advanced analyses and model building from their digital and analytics teams as needed. It's an inefficient and, more importantly, unrepeatable approach. And while it may deliver some success in specific business cases, it will never yield transformation at scale or speed. What's more, this approach can open the organization to unnecessary risk.
To advance from singular, siloed efforts to building a standardized engine for continual AI development, pharma companies can implement machine learning operations (MLOps). Rather than developing models from scratch with no consistent mechanisms for deployment or monitoring, MLOps standards and processes enable repeatable, factorylike development, deployment, and monitoring of AI capabilities.
In an MLOps environment, machine learning models and applications can be built by drawing on a catalog of reusable components, or chunks of code, that perform specific tasks and can be rapidly combined. Individual products can still be managed by product owners who understand and can advocate for the user. But on the back end, an AI production line is hard at work delivering reliable, risk-compliant, and ready-to-use models at scale, deploying them through a constant process of integration that automatically tests code for conformity with development standards.
By adopting MLOps, pharmaceutical companies can scale AI and machine learning across the organization, integrating it into core business processes and workflows. Digital teams, supported by MLOps and DataOps, can advance from simple regressions to foundational models and generative AI that can radically change how work is done throughout the pharma value chain, including R&D, medical, and commercial functions.
Action 4: Rapidly transform talent strategies
Scaling digital and analytics delivery and outcomes requires more than new operating models, data practices, and AI/machine learning approaches. Having the right talent is critical. Pharmaceutical companies can offer competitive salaries, compelling missions, and innovation cultures—all key to attracting high-performing digital professionals. Yet, the struggle to attract and retain the tech talent they need, from data scientists to product managers to experience designers, remains very real.
With digital tech talent vital to their transformation initiatives, pharmaceutical companies would benefit from sharpening their hiring and retention efforts. A recent wave of tech-company layoffs offers the pharma industry an opportunity to scoop up critical skills and talent to drive their digital and analytics ambitions. Two approaches can help: the talent "win room" and hiring by acquisition.
A talent win room is a hub for prioritizing and meeting an organization's most urgent talent needs. To set one up, the company can pool resources in a center of excellence. Taking a page from the product-operating model described above, a cross-functional product team—comprising tech-talent specialists and a product leader—focuses on the talent lifecycle, from planning and hiring to management and retention. A talent win room offers a fast, flexible approach, using an agile test-and-learn model, to developing new ideas. It may develop better talent planning approaches, new processes for hiring and onboarding, innovations in talent management, more meaningful development opportunities, and new career paths that prioritize diversity, equity, and inclusion.
Action 5: Define a digital health strategy
US venture funding in digital health peaked at almost $30 billion in 2021.2 Many pharmaceutical firms are eager to participate in this digital health space. As firms' portfolios shift toward scientific breakthroughs with substantial budget impact (for example, metabolic or cardiovascular R&D) and/or administration complexity (for example, cell and gene therapy), digital health solutions can provide an attractive adjacent value pool. In relation to their digital and analytics ambitions specifically, partnering with or investing in digital health companies creates opportunities for pharmaceutical companies to introduce new digital health applications more rapidly and to more directly engage with patients and in their therapeutic areas of focus. In the past three years, the top five pharma companies have announced more than 50 investments in digital health companies and assets and have formed twice as many digital health partnerships.
Most pharma companies are eager to realize the kinds of business outcomes that embedding digital and analytics capabilities throughout their organizations can deliver. But that will require them to make the transition from dozens of small projects with limited impact to larger digital transformation plans in line with their business strategies.
Pharmaceutical firms seeking to jump-start their initiatives can begin by aligning key stakeholders around their digital and analytics transformation goals and developing a road map to guide their efforts. Then they can invest in the kinds of fundamental changes outlined above—new operating models, DataOps, MLOps, new talent strategies, digital health partnerships and investments—to extract the full value of their data and realize the outcomes they seek.