The right data analytics solution is not a collection of individual applications reserved for just a few ‘expert’ individuals within a company.

Data analytics is essential to survival in today’s high-speed and competitive R&D and manufacturing world. Effectively managing manufacturing process data, along with the full portfolio of data collected during the pilot to commercial transition, and post-launch, is critical for launching a successful and profitable product.

A facile data analytics solution can support the required prevention, detection and response, risk reduction and regulatory stewardship workflows to support product evolution. This requires effective connectivity to the supporting data sets, necessary when performing tech transfer and ultimately product launch.

Figure 1: This manufacturing scenario illustrates how data analytics delivers real-time monitoring and assessment of process equipment reliability and maintenance requirements to improve productivity. (Credit: Alkemy Innovation Inc.)

Examples of scenarios utilizing the preferred data analytics approach include:

Scenario A: Real-time monitoring and assessment of process equipment reliability and maintenance requirements (Figure 1). In this example, facility operation is being monitored to enable improved diagnostics and drive improved manufacturing productivity for:

  • Defining a real-time diagnostics assessment plan to provide early detection and rapid engineering assessment during investigations, in addition to establishing a baseline for continuous improvement initiatives
  • Development of business documentation required to address future process improvements
  • Reducing costs by improving operational efficiency

Supporting secure communications to facilitate a company culture of team and customer collaboration.

Figure 2: This tech transfer scenario illustrates how to utilize a data analytics solution to look at and compare all relevant data for a specific reactor, bioreactor or downstream purification process. (Credit: Alkemy Innovation Inc.)

Scenario B: Pharmaceutical tech transfer for a specific reactor, bioreactor or downstream purification process (Figure 2). In this example, frontline scientists defined a golden batch reference, which served as a baseline for rapid investigations going forward post-launch. This enabled them to:

  • Use data analytics pattern search to quickly overlay, analyze, and evaluate historical and current pilot and commercial bioreactor and chromatography, performance from a variety of data sources (e.g., process data historian, analytical data, etc.)
  • Provide real-time monitoring capabilities to enable frontline users to quickly document and identify deviation causality to implement a solution at pilot scale
  • Support a culture of collaboration to create and share process knowledge.

Having access to consistent and well-linked data is one of the biggest issues for pharmaceutical R&D1, because without it, tech transfer and product launch activities will be hindered. However, a comprehensive data analytics solution doesn’t have to require an expensive overhaul of the entire data-integration system.

New data analytics applications can be used by a wider audience, and not just by a small group of IT and statistical gurus sequestered from the frontline users of data. These solutions include applications enabling frontline users of data to perform data analytics without complicated spreadsheets or writing custom code. 

So, what are the first steps? It begins with the recognition that company culture must embrace a new way of thinking. This new way of thinking — coupled with the advances in process analytical technology (PAT), cloud-based solutions, and analysis applications — can support product tech transfer and launch. This article walks through key elements for consideration when implementing these data analytics concepts. 

Data Analytics in Action

Begin by considering how front-line scientific staff and engineers utilize data to make decisions. Their know-how and experience, along with business drivers, are the foundation upon which to base a data analytics solution. Their activities and use of data, along with their collaboration, are required to create transformational linkages among research, tech transfer and commercialization — providing the cornerstone to a successful data analytics solution.

But that is just the beginning, because for a data analytics solution to be effective, it needs to be embedded in the culture. This requires infusing analytics throughout the organization to permeate the way people think and communicate with each other. Senior management leadership is crucial as this group holds the keys to drive cultural change and make required investments. Data must be analyzed and presented so it is vertically integrated, understandable, and useful to everyone — including non-technically trained staff and management.

True integration requires access to relevant data sources, cross-linkages among data sources, and workflow management. Affordable storage options like high-powered databases and cloud-based offerings have now made it possible to aggregate and manage vast amounts of data.

PAT and process control software can generate and track data through every step in the manufacturing process. This technical end-to-end integration provides scientific leaders with open-minded, competent resources to create an accessible and suitable IT infrastructure.

As shared by the FDA, using a scientific framework to mitigate risk, while facilitating continuous improvement and innovation in pharmaceutical manufacturing, is a key public health objective2.

To meet all these objectives, a data analytics solution must provide:

  • Greater visibility into every aspect of the tech transfer process to support current and former R&D learnings about the product
  • Real-time insight into the facility operation to quickly spot when a process deviates from defined key performance indicators (KPIs)
  • A facile platform to easily search historical and current data to see what happened when and why.
Figure 3: Effective data analytics requires a completely different approach than traditional empirical analysis. (Credit: Alkemy Innovation Inc.)
On the Business Front

Figure 3 outlines an effective process for determining the data required to drive business decisions. This process requires management and scientific leaders to be involved from the very beginning. Business decisions should be based on the data needed, not the data that happens to be available.

In all stages of development and manufacturing, the following questions must be answered with certainty:

  • Are we effective at accessing and analyzing our data?
  • Can we quickly respond to client and/or FDA queries?
  • Are we working as collaboratively as we could be between pilot and commercial scale?

It is imperative to identify quantifiable KPIs, and integrate these into the data analytics solution. The KPIs can be easily revisited and reassessed to ensure the PAT solution meets the business needs. In other words, users should build a foundation of analytics into the company culture to drive efficient, effective, and ultimately profitable workflows.

Figure 4: To have a holistic process view, one needs to effectively capture the full set of disparate data sets: contextual data, time-series process data and time-series analytical data. (Credit: Alkemy Innovation Inc.)

A unique way to look at the organization of data from R&D to manufacturing is to compare the core similarity of all frontline users of data, regardless of their location in the product development lifecycle (Figure 4).  Whether the focus is on a screen, an assay, a batch, a run, or other data layers — each activity has a “handle” collectively referring to contextual time-series process data. A data analytics solution should allow frontline users of data to easily access the holistic view of a data set, and easily compare it to others.  

On the Technical Front

Finding, analyzing, and visualizing data from multiple machines, processes, and databases across multiple site locations can be a daunting task without the right software. What’s needed is data analytics application software to access disparate data sources, and to grow within an organization as required to flex the company-wide data asset management solution.

Figure 5: Data analytics should begin with the business drivers, and then leverage a single application capable of connecting to the disparate data sets, eliminating the need for complicated data warehouses or duplication of data. (Credit: Alkemy Innovation Inc.)

Frontline users can now leverage existing applications to link data from multiple sources, gain insight, and communicate key findings (Figure 5). The Seeq application, for example, greatly simplifies the extraction of value by sitting above the data sources.

Seeq provides quick access to anyone on the team for searching current and historical data assets, adding context, cleansing, modeling, finding patterns, establishing boundaries, monitoring assets, and collaborating in real time. While databases can provide final reports or individual analytical data, they do not provide the leverage required to overlay and then further analyze data comparatively across batches. 


The right data analytics solution is not a collection of individual applications reserved for just a few “expert” individuals within a company. Instead, it is a single application designed in accordance with the organization’s business drivers, and fully integrated into the company culture.

As discussed, there are applications available today that can be applied immediately to demonstrate an ROI and build confidence in the plan, which can then be easily scaled across an organization. With the right mindset and leadership, companies can get the most out of their data assets to not only develop a facile data analytics solution, but also change company culture to stay competitive.


1J., Cattell et al., McKinsey & Company, “How Big Data Can Revolutionize Pharmaceutical R&D”, McKinsey Center for Government, October 2013.

2“Pharmaceutical CGMPs For The 21st Century—A Risk-Based Approach Final Report,” Department of Health and Human Services, U.S. Food and Drug Administration, September 2004.