No Drug Can Save The Ailing Data Warehouse, But There Is A Prescription For Life After Its Death
In this fast paced and cost-conscious industry, few Life Sciences companies have the appetite to spend millions of dollars or wait for months to access something that is all around us–data. Information is the basis of every decision we make and yet for an industry so dependent upon innovation for its success, we often find ourselves relying on legacy practices to gain access to data. There is no changing the fact that data needs to be collected, integrated, cleansed, and optimized before it can be made available. But it’s a bold new world out there and it’s moving fast. For many, the approach to accomplish those tasks has transformed and the traditional data warehouse simply cannot keep up. In fact, the concept of an enterprise data warehouse is fading fast.
Everything from how and where data is created and stored to how information is consumed has changed dramatically in the past several years. Mergers and acquisitions have left us with multiple diverse systems and repositories of data. Many of our information assets, such as physician notes, clinical records, vendor agreements and invoices have been digitized, leading to various formats and structures of information. Advancements in sensors and connectivity (LTE, etc.) now allow us to connect directly to internal or external devices to capture massive amounts of data. With the growth of the cloud and collaboration with strategic partners, data is now coming from many places that extend far beyond our company’s four walls. The reality is, regardless of the technology we choose; there is no single repository, no data warehouse, no data lake, or data garage that can hold all of the data we need to be competitive. And even if they could, it would become only a piece of the puzzle as soon as the next corporate merger or new data source came along. Even the way we work and interact with data has changed. It is no longer just data scientists and statisticians that need data to do their jobs; it is all people including executives, sales reps, clinical operations, supply chain analysts, quality engineers, scientists, external partners, and sometimes even machines. Consumers, who traditionally would have been happy with standard reports, are now more advanced. They look for self-service data exploration and visualization capabilities in order to consume and interpret the variety of data that supports their individual interests in the timeframe they need it.
“With the growth of the cloud and collaboration with strategic partners, data is now coming from many places that extend far beyond our company’s four walls”
Here are 10 tips Life Sciences companies can follow to succeed in the world of data analytics even as the data warehouse fades away:
1) Get out of your comfort zone and be innovative! Given the current pace of change, everything we do with data should be looked at critically. Create benchmarks to trigger change. Data driven projects that will take more than 6 months to deliver business value or that come at a significant cost should trigger a thought to consider alternatives.
2) Take data directly from the source systems when possible. Modern technologies will insulate the source systems from overuse and facilitate high performance results. This allows you to quickly create applications that meet rapidly evolving business needs while significantly reducing data management costs.
3) Focus on centralizing master data rather than all data. If you are able to keep your master data current and accurate, it makes it much easier to bring together the various sources of data available to you. This will also force you to think more carefully about what data you should, and what data you should not, store and maintain.
With the growth of the cloud and collaboration with strategic partners, data is now coming from many places that extend far beyond our company’s four walls
4) Move from technical to business centric application design. Rather than trying to store all data to address every possible business need, start with a specific business use case and find the data that supports that need. Look beyond your own organization and identify information that will add unique value for your business users.
5) Bring data consolidation closer to the business user. Involve someone from the business in the entire process from data sourcing through to consumption. The market is filled with technologies that empower the end users to collect, integrate, cleanse, and consume data for purposes of analytics.
6) Adopt a “move fast” mentality. In order to take advantage of opportunities, your business has to make decisions quickly. Giving them something good now is far more valuable than giving them something perfect six months from now. Decide which data will add the most value and start with that. Expand incrementally in rapid fashion; plan to deliver in weeks not months.
7) Treat data quality as an on-going effort that provides continuous improvement but does not act as a barrier to providing information to the business users. Communicate and be clear about the state of the information and let the business decide how useful it is. Fix data at the source and be sure to involve business users in the data quality effort.
8) Create consistency through process and behavioral change. Don’t let the idea of multiple applications or replication of data paralyze you. A “single version of the truth” is still very much alive, despite the fact that data is coming from multiple places and then is used in different ways by different groups. Create consistency without losing your focus on agility by creating lite-touch process with defined standards and approved, common data sources, reports, connectors, and visualization objects.
9) Empower individuals to be self-sufficient and create a culture of sharing. Business users have unique needs but can gain efficiencies and be more productive by sharing information. There are various levels and forms of self-service. It is not helpful to insulate users; get started and move them along the maturity curve.
10) Consider full or hybrid cloud-based solutions for data infrastructure or analytics in order to reduce your dependency on internal skills and improve your ability to shift from one technology to another.