We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.

Applying Real-Time Analytics to the Life Sciences Industry

Originally published March 24, 2009

In recent history, life sciences companies’ ability to gain or maintain their competitive advantage, direct their sales activities and adhere to legal compliance has become increasingly reliant on their business data. As such, business intelligence (BI), the ability for an organization to turn its business data into information, has become a key initiative in data asset management strategy.

Throughout its development, business intelligence has moved away from traditional transaction-side reporting, mainly “show-me-list” reports against a transactional OLTP/CRM/SFA database, to “show-me-KPI” reports against a transformed intelligence-optimized database. Such intelligence data warehouses centralize a company’s key performance indicators and focus on quantifying the relationships between related business entities to provide – in a much more robust manner – the ability to answer the business questions that drive performance.

Broadly defined, analytics is the evolution of the above-described BI system’s query execution and key performance metrics monitoring, and combines it with the capability to interact with this information through executive dashboards, alerts and analytical navigation at the speed of business. Successful analytics deployments should include the capacity to analyze, discover and draw actionable business conclusions or modify operational behavior.

To the life science organization, real-time analytics enables BI reporting access to business data not yet transformed in the data warehouse. This article seeks to outline some of the business applications where real-time analytics can create greater business value.


Figure 1

The Case for Business Intelligence

Enterprise online transactional processing (OLTP) systems and databases are optimized for storing business-related transactions, entities and operations such as calls and samples for each targeted physician. Standard entity-based transactional reporting tools may be used for quick data retrieval and publishing static list reports, but provide limited visibility into key performance indicators, rollup aggregations and the ability to actively measure and analyze metrics against sales activities, marketing campaigns and other business operations.


Figure 2:
Sample OLTP ERD Diagram that Maps to a Transactional List Report

Conversely, the business intelligence and analytics systems are optimized for analyzing the key performance indicators (KPIs) between business entities in fact-centered star schema relationships, usually stored in the data warehouse. Examples of business intelligence reports might be TRx, call reach and frequency, and # adverse events against one or more business entities at various levels of aggregation.

Analytic applications are leveraged at all levels of the organization, from sales operations to marketing to the field sales force, to generate dashboard reports, alerts, drill-down metric analysis and more. Common dashboards could be regional and national product sales or related treatments.

With an increasing reliance on data to monitor and drive the life science organization’s business decisions, the ability to analyze and dissect this data has prompted many organizations to adopt small to extensive implementations of BI systems to complement their existing library of transactional reporting and decision-making strategy.


Figure 3:
Multiple Star Schemas that Map to Corresponding Dashboard Reports

Real-Time and OLAP Business Intelligence Considerations

With business operations increasingly data-driven, it makes sense that a greater emphasis has been placed on the desire for ready access to business data at all times. More timely decisions may be made and customers may be more efficiently attended to when an organization is more responsive to its business operations.

Generally, OLAP systems leverage business data in the warehouse that has been transferred from the OLTP system or legacy data source through some form of data extraction, transformation and load (ETL) process into star schema table structures or data cubes optimized for aggregating key facts to measure the business. This ETL refresh may take days or hours, and external data sources may only be available periodically.

In tackling the “speed of business” problem, one approach to attain access to current data might be executing a full ETL refresh every six hours, for example. This is a wholly undesirable proposition with high overhead on both source OLTP and target OLAP databases with the prospect of ETL processes taking more than six hours to complete.

Another approach might be to deploy an entirely real-time-based solution. By definition, the real-time approach is to issue queries at run time directly against the OLTP database prior to the next data warehouse refresh. This, again, is not desirable as OLTP databases are not designed for intense data aggregation.

A more elegant solution might be to apply a real-time solution in those areas where alerts or rapid decisions might be necessary, reporting against data only entered into the system since the last ETL refresh.

The following may be some considerations when applying real-time business intelligence.

Consideration #1: Use When Report Refresh Period is Less than ETL Refresh Period

Individual ETL processes can take several hours to days depending on the business entities being reported against, system performance, the complexities of aggregations and overall volume of data extracted. ETL processes are generally executed in batches in an overnight, weekly, or monthly fashion at times of optimal database execution or as data becomes available depending on the data source. In those cases, when data is required prior to the next data warehouse refresh, real-time business intelligence could be applied. With a potentially higher overhead on the CRM database, the life science organization should closely examine which enterprise analytics approach best meets the needs of the required business areas, documenting and prioritizing the business requirements.

Consideration #2: Use When Snapshot Data is Adequate or Preferred

Business intelligence data is not stored in the real-time application as it is in the OLAP system. Therefore, real-time business intelligence is limited to those applications where historical data, such as historical sales force alignments, is not required. It is an option, however, to combine queries across both platforms to leverage the high level of data store available in the dedicated OLAP and the most current data in the OLTP to generate a true real-time picture of business performance.

Consideration #3: Use When Reaction to Alerts is Critical

One of the main benefits of the business intelligence solution is the ability to turn enterprise data into business information upon which business decisions may be made. For those business applications where decision making relies on current data and may be quantified in a meaningful way, real-time analytics is an excellent means by which the decision maker may be alerted.

Table 1 summarizes the considerations of the real-time and OLAP business intelligence solutions.

Real-Time Business Intelligence

 OLAP-Based Business Intelligence
Useful when reaction to KPI-based alerts is critical.

In those cases when reports need to be available prior to the next data warehouse refresh.

Use when snapshot data is sufficient. 

Needed when historical tracking is required

Trend analysis
Ability to React
Ability for business to react to “major event” triggers.

Real-time value includes as its chief benefit the ability to react as quickly as possible to new business developments.

Real-time alerts via email and text messaging may be generated as defined by KPI thresholds.

Business decision reaction is limited to the latest data warehouse refresh.
Development Effort Eliminates the need to manage a true physical data layer.

Data structures and business rules are built as a metadata layer on top of the OLTP/CRM/SFA.

New data fields are more readily added by modifying the metadata and queries issued.

Some BI tools contain OOB transformations potentially reducing the data conversion effort.

In addition to query modification, new data fields must be added by the extension of tables, ETL revision, etc.
Performance Queries are executed directly against the OLTP database. Query overhead may potentially impact the OLTP application’s overall performance.

Direct querying of the OLTP is more suited to low volume data that has not yet been migrated and converted to the OLAP.

With tables optimized for aggregation, the OLAP is ideally suited for tracking calculation-intensive queries or aggregating vast volumes of enterprise data.

Frees up the CRM system’s bandwidth.

If developed well, equivalent queries issued to the intelligence-optimized database generally execute faster on the OLAP system than against the OLTP.
Data Store No data store. All queries issued are snapshots as data exists at run-time. 

BI data is written to fact and dimension tables that may be physically archived.

Table 1: Comparison and Considerations of Real-Time and OLAP Business Intelligence Solutions

Applying Real-Time BI in Life Sciences

Life sciences organizations are now looking to gather OLTP data in a more frequent manner to help manage their business activity and decisions more effectively. Throughout the life sciences organization, different user groups and business operations will have varying degrees of reliance on how immediate data is to be delivered or presented to the end user.

Although real-time business intelligence should not necessarily be the first option used in all cases, in specific applications, real-time BI may be applied in some truly beneficial ways. The following are a few business scenarios were real-time analytics can create greater business value.

Adverse Events

Drug safety and its close monitoring have increasingly become government regulated and have become a major concern for both the public and the life sciences organization. As new drugs and medical devices are developed, tested and go to market, the ability to track adverse events (AE) is a major task.

If a call center supports adverse event management, real-time reporting and analytics on the activity data within the call center may be used to issue alerts triggered by a KPI such as the # Open AEs, # Category 1 AEs, or Average Time-to-Close. Real-time analytics is helpful using any call center metric that should trigger a rapid decision or behavior modification.

Once a predetermined threshold is crossed, the system may issue an alert to the AE manager to address what may be a bad product batch and begin the process of a potential recall. The manager would also be placed in a position to quickly react and reallocate resources on the call center floor to address specific open AEs.

Medical Information Requests

As in the case for adverse events, a marked increase or influx in the number of medical information requests of calls might also trigger the need for a product recall or investigation. This critical decision might be triggered in reaction to a real-time issued alert report that measures the current week’s calls against the previous week as captured in the OLAP system. In this case, only the incremental and most recent OLTP call data would be leveraged for optimal report performance.

Product Launch

Business intelligence applications that track call data may be used as excellent tools for coaching opportunities to ensure reps are meeting reach and frequency goals, etc. Generally, call activity data reports are steady and repeatable on a day-to-day basis and therefore do not require that report data be refreshed in less than a day. In this scenario, data may be refreshed once or twice a week with coaching taking place once or twice per quarter as the need requires.

During a product launch, however, it is critical to generate a certain number of calls in short timeframes. The life sciences company cannot afford to wait a week to ensure that the proper number of calls have been made. In this case, real-time analytics may be applied to track calls and assign resources accordingly, and to maintain call goals and frequency.


In conclusion, business intelligence leaders should take into consideration the current and future business requirements when deciding how and when to implement a real time business analytics solution. The BI leader should also look at a solution that is adaptable to change and is well aligned to the current business process of the commercial life sciences organization. This will allow the BI solution to adapt as the business processes change. Furthermore, the proper organizational change management should be included in any real time BI deployment. For example, the proper user training and organization communication will help accelerate the adoption of the real time solution and drive the optimal amount of organizational value.

  • Anthony Frank
    Anthony is a currently a Senior Consultant for IMS Health's Commercial Implementation Services practice. Anthony has over ten year of experience in implementing CRM and analytics solutions across the life sciences industry. Some of Anthony’s many life sciences clients include Baxter, Johnson & Johnson, Hologic Inc, Otsuka and Amylin. The functional implementations at these firms include sales force automation, call center support and field sales analytics.

    Anthony has implemented several leading edge technologies such as Siebel (Oracle) Analytics and Business Objects. He has done application development work on these platforms as well metadata modeling of the underlying database systems. Anthony is a graduate from the University of Toronto with a degree in Environmental Geosciences.
  • Michael ZubeyMichael Zubey

    In his role as a principal in Information Management Consulting at IMS, Michael offers pharmaceutical clients support in critical business processes involving outsourcing practices, sales reporting and analytics. His expertise lies in the business and technical challenges of IT, sales, marketing, market research and sales operations.       

    Over the past 13 years, he has been a sales, marketing and finance leader for several technology and service organizations such as Verizon and DecisionOne. Prior to joining IMS, Zubey was the director of Solutions Management for Unisys Global Outsourcing and Infrastructure Services where he developed IT service offerings to meet the future needs of Unisys’s current and prospective clients. For his first four years at Unisys, he was the senior manager for the Unisys Business Intelligence program. In that capacity, Zubey created business intelligence solutions for customers such as Pfizer, Premera Blue Cross, and Fortis Health.

    Zubey has coauthored a textbook, Customer Relationship Management: A People, Process, and Technology Approach with his associate, William P. Wagner Ph.D. He is also a speaker at industry venues including The Data Warehousing Institute.

    Editor's note: More pharmaceutical articles, resources, news and events are available in the BeyeNETWORK's Pharmaceutical Channel. Be sure to visit today!

Recent articles by Anthony Frank, Michael Zubey



Want to post a comment? Login or become a member today!

Be the first to comment!