TLDR: Summary of Recommendations for Digital Change Management
Build rapport with users incrementally by starting with a small focus group. After gathering feedback from this group, plan to scale to the larger organization
Find allies in the organization that are in leadership roles, and technical early adopters that are involved in the day-to-day of research & development
Make it easy to start by focusing on user experience, ease of access, training and documentation.
Focus on a few key deliverables before getting distracted by a scale-up plan.
Create community spaces that are physical and digital to support collaboration and self-directed learning.
Introduction
Digital transformation has become a buzzword across industries, and the biotechnology/biopharma field is no exception. It is all too common that scientific data are managed using a combination of spreadsheets, homegrown tools, and off the shelf systems. With biotechnology quickly becoming one of the most data intensive fields in the world, it is obvious that modern research organizations need better ways to capture, share and store their experimental data.
Introducing new software tools in a lab setting can be challenging, especially since research cultivates a healthy level of skepticism in scientists. An education in research requires skepticism, a critical attitude of questioning and challenging assumptions, generating hypotheses, and collecting evidence to arrive at reliable conclusions. Although a beneficial tool for research, these skills render scientists resistant to change. This is where digital change management comes in – it is the process of guiding individuals and teams through the adoption of new digital tools or technology, ensuring that the change is implemented successfully and smoothly.
The needs of organizations vary based on type of research, size of organization and regulatory requirements1. These drive how an organization approaches change and the benefits to successful digital change management are numerous - improved collaboration, better tracking of data provenance, and enabling the ability to make decisions using historical data. All of these benefits directly lead to fewer inefficiencies in reaching scientific insights and remove bottlenecks in drug discovery pipelines at Phase 1 and 22.
How do scientists currently manage their data?
Biotech companies devote significant time and money to improving scientific data management, but there is no silver bullet that solves everything. If you asked a scientist how they manage their data, it is guaranteed that spreadsheets are used in the data lifecycle. One of the primary advantages of using these GUI3-based applications is their flexibility. They let scientists customize their toolbox with an easy to use interface, which is immensely helpful when exploring a hypothesis or sketching ideas – because not all ideas can be conveyed in the tidy data format4 at the start. Spreadsheet-based desktop applications can be beneficial for both data analysis and data management. They provide a range of out-of-the-box statistical functions that can be used to analyze data, making them ideal tools for low-throughput tasks like assay development. Spreadsheets’ ease of use is another significant advantage, especially for researchers who may not have extensive training in statistics or script-based analysis using R or Python. With its familiar user interface and numerous online resources available, spreadsheets can be used by researchers at all levels of experience. Web versions of spreadsheet tools are also used for data management and collaboration with other scientists. This accessibility has made it an invaluable tool for students and researchers who are just starting in the field.
Despite their flexibility, these systems are incapable of providing a modern research lab a rigorous, scalable, distributed environment for data capture. One of the main drawbacks of using spreadsheet based tools is their limited ability to handle large datasets and lack of software elements5. While most spreadsheet tools can handle up to a million rows of data6, this is still insufficient for large research projects in the biotech and drug discovery space, particularly those that involve genomics or other large-scale data analysis. Another potential limitation is the risk of errors in data entry or analysis. Spreadsheets are not designed for scientific research, and researchers may make errors when inputting or analyzing data. These errors can have significant consequences, particularly if they are not identified and corrected early on. A well cited example is MS Excel’s propensity to auto-correct genetic annotations into dates.
Modern R&D software teams use purpose-made computational tools to automatically capture raw instrument data and metadata for scientists. These tools require swapping spreadsheets for a very different and complex set of software. Changing this experience for scientists is not easy – it requires Digital Change Management.
Path to digital change management
Digital Change Management is dealing with the human side of digital change. Change management should aim to make people feel welcome and help them accept, execute, and adopt new tools. These actions may not make sense in isolation, so let's talk about what you can do to enact effective change in your team.
Successful change management projects include leadership, IT7, and scientists (as active stakeholders). The software suite of choice must meet the needs of all these personas to address issues like making data FAIR8, protecting organizational IP, keeping data secure, and making the scientific process more efficient. For a deeper dive into incorporating FAIR data principles into R&D infrastructure, check out Bits in Bio’s previous article on this topic FAIR Data Infrastructure in Life Sciences R&D.
Show, don’t tell
Scientists can be highly discerning when selecting research tools and are accustomed to doing their own research on products before making a commitment. Their behavior and approach to adoption of new software can be challenging. Forcing a solution upon them without first building rapport will lead to failed adoption and engagement.
Start projects with focus groups where teams present their current workflows. This could be a session where you engage with scientists, ask questions about current pain points, or have a walk through describing ‘a regular day in the lab’. This will help you become educated on user needs and pre-validate solutions you think might help.
Find your early adopters and influence incrementally
Change agents only have so much time in the day and walking through every lab with every scientist just doesn't scale. To broaden your message for change, find allies in your organization that understand the science and can help you build a group of early-adopters. These early adopters will bring technical subject matter expertise to your project and help bring your message to bench scientists across the organization. Working with aAllies that bring leadership support and eEarly aAdopters that bring scientist’s support, you will have the most successful digital change.
Early adopters can do groundwork and support pilot studies with smaller groups prior to institution-wide release. Running these pilots is a high leverage activity but has high cost per user acquired. To increase the impact of the pilot, early adopters should be encouraged to share successes with their peers. Change agents can be seen as outsiders and having scientists convince scientists is going to be a more effective use of your time and emotional capital. A recent case study captures the successful implementation of an ELN for 800+ users that was achieved via an initial pilot study with a smaller group of 67 early adopters.
Make it elegant and easy to start
Change is hard. Change is near-impossible when there are barriers. Remove these barriers to drive users to your solutions because hard to access platforms, bad documentation, bad UX9 are all things that will kill interest in your solution before it gets off the ground. UX matters. A lot.
Collect user metrics as often as you can to get early feedback on adoption, and take temperature checks with your early adopters to see what you can improve. Usage metrics include page views, average time on page, and average session duration. User metrics are good proxies for user engagement and ‘stickiness’ of a new software tool but also require careful review to ensure you are making the right conclusions on the health of your digital change.
Always be thinking about a roadmap of enhancements and how you can challenge and build on your assumption. Don’t gold-plate solutions and instead focus on rapid releases of incremental value that can slowly overtake the legacy solution you are seeking to replace.
"We haven't moved off of Windows 95 because it is just too much work to validate and migrate 10,000 computers across our network" - IT Manager, F500 BioPharma
Keep the vision focused
As a leader in your organization, your visions are of a perfectly harmonized lab – but you won't get there on day one. Although it is useful to have a 5 or 10-year plan on scaling data management, you can’t lose sight of the practical issues that scientists deal with in their day-to-day lives.
Keep an eye on the big picture, but focus on the here and now with your users. Make sure you engage with their thoughts of the future even if you have a vision that conflicts. Focusing on immediate and practical pain-points to show usability for researchers will earn trust early and help ensure successful adoption.
Create community spaces
Between remote work, distributed teams, and demanding lab schedules, teams can't be together all the time. Make space for (either digital or in-person) knowledge sharing and ideation. Teams may teach you something new about your work or solve a problem you didn’t know you had. Try starting with a weekly ‘drop in’ session where people can come to ask questions.
Technology-led companies need to incentivize lab scientists to be software stakeholders and create a company culture that encourages data literacy. Culture and empathy matter in place like this so make space for improvement and encourage the inevitable growth you will see in teams.
Parting thoughts
The science of studying disease and the living world is an age old art, and the practitioners of this art need better tools. It is hard to break old habits and traditions within research, so organizations that take the path of digital change management must treat it as a long game. You are instilling the right policies, reward systems, and ultimately better science across your organization – don’t give up!
If you’re interested in learning more about digital change management or discussing these topics, make sure to check us out on Slack!
HIPAA, GDPR, GxP, SOx, etc
U.S. Food and Drug Administration. Impact Story: Modeling Tools Could Modernize Generic Drug Development https://www.fda.gov/drugs/regulatory-science-action/impact-story-modeling-tools-could-modernize-generic-drug-development (June 2020). Accessed March 27, 2023.
Graphical User Interface
Wickham H (2014). “Tidy Data.” Journal of Statistical Software, 59(10).
Think of elements like data schema, package import/export, and logging.
https://support.microsoft.com/en-us/office/excel-specifications-and-limits-1672b34d-7043-467e-8e27-269d656771c3
“Information technology”. Used here as a catch-all for the organization responsible for digital system selection, creation, maintenance, deployment, and support.
Wilkinson, M., Dumontier, M., Aalbersberg, I. et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data 3, 160018 (2016). https://doi.org/10.1038/sdata.2016.18
User Experience – the interactions a user has with a product, service, or good.