It’s Time to Talk About the Minimal Valuable Product, and Not Just an MVP

by Brigitte Ganter, Ph.D. | posted on April 12, 2024

I had the great opportunity to attend the latest Smartlab Exchange conference, which took place April 8-9 in Fort Lauderdale, Florida, USA. The theme of the conference focused on lab process digitalization, effective change management, and emerging technologies. As the name of the conference implies, there was a constant exchange between industry delegates and commercial solution providers, and wow, was there lots of exchange! One key takeaway message I took home from all the discussions and presentations: An MVP is not enough anymore; we need to think Minimal Valuable Product!

Discuss the Implementation Strategy Upfront! You get further!

It was interesting to see the enthusiasm and demand for process digitalization is higher than ever while also accompanied by the challenges that need to be navigated and addressed. The challenges, as I heard them, were around ownership clarification of change management, people being afraid of failing fast and picking the wrong solution, not enough resources with the right training for automation and digital transformation, implementation fatigue, and last, but not least, providing and selling a solution that is built with the end user in mind and not just selling a system.

Some great quotes I collected during the different talks and panel discussions:

• “People working in the lab take longer to generate the paperwork then actually working on an experiment – so we really have to think about it from an end user in the lab perspective.”
• “Unlearning is required to adopt change.”
• “Everything should be completely integrated, including LIMS, ELN, etc.”
• “All the data produced is going through submissions and needs to be good for AI applications.”
• “One of the most important things is understanding the data, what is required in terms of security and data privacy, and what can be achieved via partnerships or building the controls in-house.”
• “Data has to be data ready so it can be used for decision-making purposes, including single source of truth.”
• “The right people for the right technology, which includes skill set, and the right infrastructure in place, and we need the right dollar amount to support it.”
• “Collaboration is an important aspect: in the lab with our peers and across the company.”
• “KPI is important – do science more effectively which resolves around taking a science-centered approach.”
• “Lab of the future has to plan for me the whole lab, including inventory management, so I can focus on the science.”
• “Digital transformation means E2E.”
• “The architecture and how the structure of the data is going to be are so important – since it is a multi-year type investment and a lot of dollars, so everyone wants the benefits pretty quickly, and as such, the architectural foundation is very important.”
• “Patience is important.”
• “We have done a lot from paper to digital, but people need to understand what the benefits are to be able to get over the hump.”
• “To successfully achieve digital transformation means to assess the current pain points and build out from there.”
• “Compared to programming background 18 years ago, nobody had that particular training/background, but now this is a completely different story. Eventually, the same will happen with digital transformation, automation background etc. which we should expect in 5-10 years.”
• “Always look within the organization and see who is the biggest supporter within a department – for buy-in you need department specific support.”

Some interesting live polling that happened during the conference

The data was based on about 30-40 active delegates in the room during any of those polling sessions.

What are you seeking to use AI for? N=35

• Automated analysis – 77%
• Experiment design and optimization – 73%
• Predictive maintenance and optimization – 54%
• Real-time monitoring for safety/efficiency – 46%
• Robotic sample handling and analysis – 42%
• Smart inventory management – 33%

What is your strategy for dealing with disconnected lab systems? N=34

• Adopting standardized data formats and using data lakes/warehouses for centralized data storage – 76%
• Investing in integration platforms – 41%
• Implementing middleware for data integration – 32%
• Prioritizing system compatibility in procurement – 26%

What is the best strategy for overcoming disconnected lab systems? N=37

• A mixture of both – 89%
• Replacing systems – 8%
• Integrating systems – 3%

At what stage are you in your generative AI journey? N=31

• Thinking about implementation and establishing strategy – 52%
• Started implementing a few initiatives – 32%
• Actively implementing – 16%
• Not thinking about implementing – 0%

How frequently does IT share their vision for digital journey with scientists? N=24

• They never share anything with us – 58%
• Monthly – 21%
• Quarterly – 17%
• Annually – 4%

It Is Important to Not Just Sell a System but Provide an Actual Solution

The breakout panel discussion with Scott Lunk (Bayer Crop Sciences) and Robert Guenard (Pfizer) engaged with the entire audience of 70+ individuals in the room, focusing on Change Management with the big takeaway message: focus on the journey with shaping the path instead of just thinking about the outcome. Whereas the path requires having owners (and not just buy-in because buy-in means the decision is already done), tenacity, meeting people where they are (e.g., if you want to improve the lab, go to the lab to learn more), bringing in some training (which includes training of the leadership team), and implement hyper care post-implementation (provide additional system support for like three months plus and don’t let people and labs hang after implementation).

Which brings me to the Minimal Valuable Product

Considering all the information I gathered, the industry sector is clearly longing for a Minimal Valuable Product and not just an MVP. Besides providing a solution that digitalizes and manages all lab processes and data, for example, via a LIMS, ELN, or LES, such as a solution also must make data available in a ready-to-use format for decision-making purpose and prevent the data from going to the dark side – as was mentioned by someone at the conference.

A Minimal Valuable Product needs to:

• Be flexible, in that other laboratory management solutions can be fully integrated.
• Be built with the end user in mind, and not just IT, but also the lab personnel.
• Support collaboration across different teams and programs.
• Be highly secure with access controls and support data privacy.
• Provide data in a ready-to-use format for decision-making purposes.
• Function as a single source of truth.
• Support a science-centered approach.
• Have an architectural foundation that includes accessible and fully structured data.
• Address the current pain points in the lab.
• Be patient-centered.
• Support routine lab tasks, such as inventory management and scheduling.
• Be built by vendor organizations that:
  • Have a focused approach on delivering the right solution.
  • Understand the data science component just as much as they have domain knowledge.
  • Have alignment from top to bottom and have the forward-thinking culture to implement.
  • Provide hyper care post-implementation.

t was nice to see that L7 Informatics with its Unified Platform L7|ESP is nicely aligned with these Minimal Valuable Product requirements, as it is built with an architecture for digitalization in mind, which not only manages all lab and business processes via workflow orchestration, using a low code/no code authoring tool, it is also integrates numerous lab management applications such as L7 LIMS, L7 Notebooks, L7 MES, L7 Scheduling, Inventory Management, and many other applications. This the while collecting all the process data along the way as the data is produced. The data collected does not only include master process data, but also valuable meta data that contextualizes the data and adds meaning to any process data point.

The other alignment L7|ESP provides is that it can also integrate with existing lab management systems, including legacy systems, which, therefore, does not require any organization to replace existing systems but rather means it can function as a single source of truth. Furthermore, L7|ESP has the potential to provide lab and business insights via the L7|Intelligence layer, which includes data models, such as a physical database schema, ITDs (a set of normalized tables providing access to entity and process data), and Star schema data products. A data scientist can then use all the standardized and structured process and business data and extract insights to translate them into faster, optimized, and cheaper lab and business processes by applying data models (both internally within the platform or by exporting the data to analyze and visualize in your tool of choice outside of L7|ESP), including AI capabilities.