Speaking as an industry insider with an external perspective: there's an awful lot of truth in these observations. I disagree with some of the smaller points and comparisons, and have questions about others... none of those topics are suitable for debate in Substack's comment system, so I'll defer those for another time and place and just note that I agree fully with the overall tone and high-level conclusions.
One thing I'll add, though. This statement:
"Biopharma operates as an N-of-1 economy: every organization builds bespoke data infrastructures, bespoke use cases, bespoke integrations — all for internal use."
One possible response to that is that this is solved within Big Pharma, i.e. that the large pharmas can effectively be viewed as large portfolios of projects (each such project being comparable to individual biotechs), and that Big Pharma solves this N-of-1 problem across those projects by creating shared assets within the company. Data assets, computing assets, knowledge assets, scientific assets, common methods - all contained within (largely) the same IP ownership space.
And yet, even within the walls of those pharmas, everything you're saying here is true. While there absolutely are shared assets, there are also walled gardens of data. There are endlessly repeated experiments. There is an enabled artisanal culture. There's deep competition within the companies (to the detriment of patients) often enabled by data obscurity. I've lost track of the number of times that I, as a digital leader within pharma, had to tell a scientist, "I'm sorry, but it's not your data. It's the company's data."
There are also places where collaboration and data reuse are working incredibly well, thanks to people trying hard to work that way, or where the costs make it obviously the right thing to do. But it's very difficult to do so even just within one large company for all the reasons listed in this essay. Systemic architecture, misaligned incentives, and deeply-rooted culture. That ... and, turns out, scientific data is hard. People have known this for decades: the pharmas that finally get this right are going to have a huge advantage.. even moreso in the age of AI.
Remy, as ever, your intellectual honesty and rigor is welcome and needed in this discussion. We need more like you, and we need a very public debate about all of this.
Excellent analysis of the situation. I would add that these structural and cultural problems flow from scientists trained - at least in the US and most of the EU - in an academic and funding model that prepares them to operate as agents in a discovery cottage industry, and trains them to behave as isolated practitioners in competition with essentially all other scientists in their field.
Research departments in universities are built around the careers of individual faculty members. These principle investigators are responsible for their own funding (and often to a large degree their income) and that of many or all of their trainees. This leads to a culture of control over every aspect of one’s lab: type of equipment, software, assays, chemical and biological materials, and most of all, the resulting data. Those data are the key to publications and those papers are in turn the means to the next grant.
In this ecosystem, siloing behavior may be entirely rational, and the founders of and incumbents in our biopharma industry are the product of this system. There are many encouraging examples of more collaborative science in academia. One of the early examples that worked in life sciences that comes to mind was a consortium that formed to study Huntington’s Disease, where group success was measured by overall progress, not individual achievement, and those successes were shared by all members of the group.
While there have been a growing number of examples of collaborative science, the basic structure and culture has not changed. Can industry help lead this change, or is a change in academia a necessary precursor to rewriting the operating model in industry?
This is such a profoundly important observation and insight and one worthy of one or more additional essays. I've tried to lay some of the groundwork for the cultural root causes, but not the more complete historical context, which is required. Thank you for weighing in!
Excellent post that captures my observations over 25 years. Would you also say the broken R&D IT model is a contributing factor to the diagnosis or is just a symptom of the broken scientific data strategy ? I have felt that IT leadership has typically lacked an understanding of the scientific process and scientific leadership has lacked an understanding of the concept of industrialization of the drug discovery and development process.
Speaking as an industry insider with an external perspective: there's an awful lot of truth in these observations. I disagree with some of the smaller points and comparisons, and have questions about others... none of those topics are suitable for debate in Substack's comment system, so I'll defer those for another time and place and just note that I agree fully with the overall tone and high-level conclusions.
One thing I'll add, though. This statement:
"Biopharma operates as an N-of-1 economy: every organization builds bespoke data infrastructures, bespoke use cases, bespoke integrations — all for internal use."
One possible response to that is that this is solved within Big Pharma, i.e. that the large pharmas can effectively be viewed as large portfolios of projects (each such project being comparable to individual biotechs), and that Big Pharma solves this N-of-1 problem across those projects by creating shared assets within the company. Data assets, computing assets, knowledge assets, scientific assets, common methods - all contained within (largely) the same IP ownership space.
And yet, even within the walls of those pharmas, everything you're saying here is true. While there absolutely are shared assets, there are also walled gardens of data. There are endlessly repeated experiments. There is an enabled artisanal culture. There's deep competition within the companies (to the detriment of patients) often enabled by data obscurity. I've lost track of the number of times that I, as a digital leader within pharma, had to tell a scientist, "I'm sorry, but it's not your data. It's the company's data."
There are also places where collaboration and data reuse are working incredibly well, thanks to people trying hard to work that way, or where the costs make it obviously the right thing to do. But it's very difficult to do so even just within one large company for all the reasons listed in this essay. Systemic architecture, misaligned incentives, and deeply-rooted culture. That ... and, turns out, scientific data is hard. People have known this for decades: the pharmas that finally get this right are going to have a huge advantage.. even moreso in the age of AI.
Remy, as ever, your intellectual honesty and rigor is welcome and needed in this discussion. We need more like you, and we need a very public debate about all of this.
Excellent analysis of the situation. I would add that these structural and cultural problems flow from scientists trained - at least in the US and most of the EU - in an academic and funding model that prepares them to operate as agents in a discovery cottage industry, and trains them to behave as isolated practitioners in competition with essentially all other scientists in their field.
Research departments in universities are built around the careers of individual faculty members. These principle investigators are responsible for their own funding (and often to a large degree their income) and that of many or all of their trainees. This leads to a culture of control over every aspect of one’s lab: type of equipment, software, assays, chemical and biological materials, and most of all, the resulting data. Those data are the key to publications and those papers are in turn the means to the next grant.
In this ecosystem, siloing behavior may be entirely rational, and the founders of and incumbents in our biopharma industry are the product of this system. There are many encouraging examples of more collaborative science in academia. One of the early examples that worked in life sciences that comes to mind was a consortium that formed to study Huntington’s Disease, where group success was measured by overall progress, not individual achievement, and those successes were shared by all members of the group.
While there have been a growing number of examples of collaborative science, the basic structure and culture has not changed. Can industry help lead this change, or is a change in academia a necessary precursor to rewriting the operating model in industry?
This is such a profoundly important observation and insight and one worthy of one or more additional essays. I've tried to lay some of the groundwork for the cultural root causes, but not the more complete historical context, which is required. Thank you for weighing in!
Excellent post that captures my observations over 25 years. Would you also say the broken R&D IT model is a contributing factor to the diagnosis or is just a symptom of the broken scientific data strategy ? I have felt that IT leadership has typically lacked an understanding of the scientific process and scientific leadership has lacked an understanding of the concept of industrialization of the drug discovery and development process.