I recently attended the SmartLab Exchange conference in San Diego, California, a small but high powered conference focused on digital transformation in scientific laboratories. As I sat and listened to panels, presentations, and roundtable discussions and talked to many of the delegates, I heard digital transformation being put forward as a solution... but to what problem? The purpose of digital transformation is turning data into decisions, increasing efficiency, and reducing human error. Many described this transformation as a journey, a multi-year strategy dependent on consistent execution. But a journey to where? To execute what? As “transformation” is a nebulous outcome, how do we get more specific to focus on the scientific, operational, and compliance outcomes? And where do we start? Scientific business analysis is the start of that journey.
Scientific Business Analysis is about surfacing problems
There are seemingly obvious starting points when completely transforming a business reliant on the usage of scientific data to drive decision-making. Often these starting points are major overhauls of legacy systems and practices, for example, electronic lab notebooks (ELNs), lab information management systems (LIMS), and data lakes to capture all scientific data and metadata to make it useable for artificial intelligence (AI) and machine learning (ML). The ultimate goal is to improve the quality and throughput of drug discovery, development, and manufacturing. However, once one or more of these systems are in place, biopharma companies are still struggling with how to translate these implementations into
- Faster time to value for improving scientific processes
- Necessary change management to get all teams involved on board
- Connecting people and teams who desperately depend on each others’ data
- Tangible AI-driven scientific and business outcomes
Scientific business analysis is the key to translating the benefits of new technology and tools into real-world value. Crucially, it helps organizations determine which problems are the most impactful, thereby providing a baseline on which to improve first and where it will be beneficial to leverage Scientific AI. Over time, solving the problems with the biggest impact provides ever-improving baselines towards hitting long-term digital transformation goals. Business Analysis is not new to the IT community, where for many years digital transformations allowed ERP systems to be connected to CRM systems and so on. But within the lab, this type of analysis has been sparse and AI has rarely been used.
Scientific Business Analysis is an art and a science
In order to surface problems, you need to ask lots of questions. Seems easy enough, right? In fact, it is, but with the myriad choices of scientific instrumentation available today, researchers from different laboratories can choose from a wide variety of systems and vendors to achieve the same outcome. Take for instance the workflow of high throughput screening (HTS) in drug discovery. The goal is simple…find quality lead candidates for optimization to enter the pre-clinical and clinical funnel. However, biopharma companies approach this goal in vastly different ways. Some have fully automated suites that can screen millions of compounds a year, while others rely on traditional bench-scale processes that are mostly manual. Understanding the magnitude of the pain in each of those workflows and their associated impact is the science. Knowing what questions to ask, to whom, and when is the art. This requires not only domain knowledge in the scientific area being explored, but in the data realm, understanding user inputs and outputs, file formats, how data gets from one place to another, and how Scientific AI can potentially help to improve scientists’ everyday work. Scientific Business Analysts have uncovered many bigger problems and potential improvements than the initial ones identified by users just by knowing which questions to ask.
Other critical parts of scientific business analysis include:
- Documenting existing workflows and related pain points consistently (“as-is” state)
- Shadowing scientists in the lab and/or whiteboarding scientific workflows to uncover expressed and unexpressed needs
- Translating workflow requirements into technical user stories
- Proposing solutions with tech-savvy scientific data architects and data scientists
- Testing that solution to save scientists’ time, allowing them to focus on more important tasks
- Documenting the improvements the solution provides, and comparing to the baseline (“to-be” state)
- Recommending next steps or other workflows that can benefit from the same or similar solutions
- Exploratory data analysis to see what questions can now be asked of the data that can be answered by advanced analytics and Scientific AI
In addition to this, scientific business analysts with deep experience can then start to share industry best practices with customers. This will further speed up the implementation of digital transformation strategies.
Time savings, higher quality data, and new AI-driven scientific insights
The best way to initiate scientific business analysis is to brainstorm all the potential areas where digital tools (such as replatforming all scientific data to the cloud) could have high impact on current problems. Then, dig into 5-10 of those problem areas with scientific business analysts talking to subject matter experts in the customer’s lab. Home in on the most painful scientific workflows…how many steps are there and which of them are manual, how is data generated at each step, how long does it take to complete each step, how does data get from one place to another and by whom, and what errors can be introduced that could have downstream effects. Scientific Business Analysts always work to identify the impact during these discovery sessions. They have found some extraordinary metrics that establish a baseline. These include:
- 10-15 clicks to get scientific data from an instrument to an ELN
- 50% improvement of throughput without increasing the lab footprint
- 1,000 hours per year spent manually transcribing assay results
- 7X higher probability of having to repeat experiments simply because scientists cannot find the data
- $30,000 in extra consumable expense per HPLC system due to lack of visibility to instrument/method performance
These metrics can be put in front of management to support data-driven decisions on what problems to tackle first, and how to measure the improvement.
Some say that to surface problems with no solution is a waste of time. But isn’t implementing a “digital transformation” strategy even more of a waste if you don’t know what problems you are solving for, neglecting the potential of Scientific AI, or how to employ such strategy on various scientific workflows across the organization?
Contact us today to find out which problems you need to solve and which solution can help you to make your digital transformation strategy successful.