Podcast Transcript

David Brühlmann [00:00:33]:
Picture this. You walk into a state-of-the-art biotech lab. You’ll see millions of dollars in equipment, some of the sharpest minds in science, and yet half the data is trapped in spreadsheets no one else can read. Data lives in silos. Spreadsheets bridge the gaps. Automation projects stall before they start. My guest today, David Hardy, market and innovation strategist at Thermo Fisher Scientific, has spent 25 years at the intersection of data, automation, and lab science. He’s here to unpack why connectivity is the quiet bottleneck few are solving.

David, welcome. It’s good to have you on today.

David Hardy [00:02:28]:
Thank you, David. Glad to be here.

David Brühlmann [00:02:29]:
To start us out, David, share something that you believe about data connectivity and data integration that most people disagree with.

David Hardy [00:02:39]:
That’s a great opening question. Where I see the biggest challenges, rather than disagreements, is that when you embark on a digital transformation project, it’s a long game. It’s not a “*let’s get this done in a couple of months, six months, a year, whatever.*” This is kind of a challenge almost for life, where new data is coming in, new ideas and new questions require new pieces of data or require you to go back to the original data. I think that’s one of the biggest challenges and misconceptions around, particularly, the automated digital lab and data connectivity: that it can ultimately be done very quickly. It can, in some regards, if you take a small project — just like any project. But really, this is a long game. It’s not something that can be done quickly overnight. You have to really plan for it and stick to it.

David Brühlmann [00:03:29]:
I love this topic, and I’m excited to be able to talk about it and peel the onion a bit more. But before we do that, I would love to talk about yourself. **Tell us what first sparked your interest in data and digital technology, and what were some of the pivotal moments along your career path?

David Hardy [00:03:46]:
I think it all started over 25 years ago. I was fortunate enough to be doing a postdoc at Astra, and I actually went through the merger with Zeneca. What I was doing there was NMR — nuclear magnetic resonance. The group I was in had a great way of storing and archiving the data. The company had a great strategy for when you created a new compound: you registered it, it received an ID, and that was then used to store it long term. Back in the day, it was on CDs, I believe. And then we got a kind of server, etc. With that, it got me thinking. A chemist would come up and say, “Hey, can I get this kind of spectrum?” So what I was doing was thinking about: how can we make that easier? How can we make it more accessible for the chemist in this particular case?

Bear in mind, this was 1999, I think. What I did was create a webpage using Perl back in the day to allow the chemists, from their computers, to access that kind of data, search the legacy data, and find it. And that really got me thinking: “Okay, this is just NMR data.” Luckily, we had pretty good metadata around it at the time. But it got me thinking about what the bigger context of this was. What happens if we start bringing in LC-MS data? We were doing some kind of hyphenated techniques. So it was really just the thought then of what’s possible when all this data is brought together and contextualized. And that really sparked my interest and career in data, shall we say?

David Brühlmann [00:05:12]:
And what were the steps leading you to your current function? Can you tell us what you have done since working on NMR?

David Hardy [00:05:21]:
Since then, obviously I was a postdoc, and then I moved into the world of largely analytical instruments. I’ve worked for a range of other vendors, primarily focusing on things like LC-MS. I did take a brief sabbatical out into the world of retail analytics, which was super, super interesting. They’re dealing with very large datasets. You think about how many items are sold in supermarkets every day. But I still had the draw of what I would class as the proper traditional scientific world. So I only spent a couple of years there before being drawn back into working for Thermo Fisher Scientific, where I am now. Along the way, I’ve done a range of jobs — from software developer to product manager. And actually, when I joined Thermo Fisher Scientific, I joined in a pre-sales role.

I’ve never really been at the very front end of the business, but it was great to get out there, visit lots of customers, and interact with a diverse range of customers as well. So not just pharma and biopharma customers, but all different types of laboratories, talking to them, seeing the labs, and realizing they all fundamentally had the same issues around data.

David Brühlmann [00:06:27]:
And what was the moment, as you were working with these customers, different people, and different stakeholders, that made you realize that most labs were operating with a fundamental connectivity problem?

David Hardy [00:06:40]:
I don’t think there was one particular “aha” moment — this is the challenge that they’ve got. I think it’s an accumulation, going back to my Astra and AstraZeneca days through to visiting customers in pre-sales situations and really not seeing a huge change in how they all handle their vast amounts of data. Over that time period, the amount of data being produced has grown exponentially. But fundamentally, the challenges are still the same. Particularly now with — I won’t say the advent of AI, but the adoption of AI — and the need for good-quality data, everything has really come together.

David Brühlmann [00:07:17]:
Data is very complex. There are a lot of changes happening today. We have new technologies, different mindsets. From my experience, also talking to different labs or different companies, you get different answers. What are some of the most common issues you see when people push back against new digital solutions or new ideas?

David Hardy [00:07:38]:
In terms of pushback in principle, I don’t think you get very much pushback around the idea that we need to digitally transform the lab, or that we need to make this data accessible for AI, etc. No one is going to disagree with that concept or that approach. Everyone is initially supportive of: “*Let’s do something about this. Let’s transform the lab. Think of all the things we can do with the data when we’ve got it all together, standardized, contextualized, etc.” The biggest pushbacks really are perhaps more around what I mentioned earlier — that it’s a long journey to do this successfully. As I mentioned, it’s not something that can necessarily be done in six months. It varies, and it’s something that you have to keep doing continuously as new instruments come in, new questions come up, etc. So the biggest pushback really is: “Oh, we have to keep doing this for a longer time, and there’s a commitment to it.” And that’s when you get into the culture of management and ensuring that it continues.

David Brühlmann [00:08:34]:
And what kind of culture do you need to ensure it continues? What are some of the mindsets teams would need to be successful?

David Hardy [00:08:43]:
I think in terms of mindset, it’s having — again, I don’t think anyone out there would not have a digital mindset in terms of thinking about what we could do: the ease of creating reports, the ease of creating submissions, etc. Occasionally, there’s pushback around: “Well, this is the way we’ve been doing things.” So the cultural aspects needed are very much about having a long-term vision of what they’re going to do with the data, the types of questions they’re going to handle in the future.

And then that goes into the culture of leadership. Quite often, I’ve seen presentations — some great presentations — around things like knowledge graphs in particular, contextualizing data, relating everything together. But then you follow up with them a year later, and that kind of project has been shelved or postponed. It goes back to that longer-term vision of the data and what you can achieve with it.

David Brühlmann [00:09:34]:
Now I’m curious, David. I’ve seen various labs and different people talking about data, and you mentioned in the beginning of our conversation that datasets have increased exponentially. Now, a problem that I’ve seen is fragmentation. It’s stored in different systems. You have different legacy systems, different ways of doing things. When you walk into a biotech lab, how do you spot that digital fragmentation is creating problems — problems that these teams have not recognized yet?

David Hardy [00:10:06]:
It’s quite easy to spot. One thing you missed off there — and I represent a vendor, Thermo Fisher Scientific, and customers have a choice. They’re free to choose a range of vendors, which again can present some challenges when it comes to data transformation, the data journey, and getting all that data together. But really, when you go into a lab, you get a very early and clear picture of what state it’s in. For example, are they still using Excel to manage all that data? Hopefully they’re not writing it down and transcribing it onto paper — that still does happen. But I think as you get in there and start talking to them, understanding what they’re doing today and what their vision is around digital transformation, you get a very clear picture of how advanced they are on that journey.

David Brühlmann [00:10:53]:
And where do you think the fragmentation hurts them the most? Because I imagine they’re sitting on tons of data and a lot of insights. What are the main pain points, or what are they missing out on?

David Hardy [00:11:06]:
I think what you find is — and this also goes back to my previous comment about my time at AstraZeneca — that in some situations they have a very good process in place today for maybe archiving the data and generating reports from certain types of instruments. I think the challenges and complexities start to arise when they want to bring a lot of that data together and start asking questions across different modalities, different types of instruments, and different vendors. You see this a lot when reports, etc., need to be created. Thinking of CMC in particular, that pulls in a lot of diverse datasets. There are some great examples out there today where customers are using things like AI to help with that.

David Brühlmann [00:11:48]:
What things have you seen companies do well in order to make progress on connectivity?

David Hardy [00:11:55]:
There’s a combination of things that customers do. First, they understand that they have a problem and that they obviously need a solution. One option is to build something in-house to help with that. But then that can quickly snowball, shall we say, and become more complicated as things start to scale. The alternative is to bring in other vendors, third parties, to help consolidate all that data and provide guidance, tools, etc., to consolidate and standardize that data.

David Brühlmann [00:12:24]:
When people are on this journey of digital transformation, it can feel quite overwhelming because there are a lot of things you could work on. On the other hand, you often have tremendous time pressure. What are some of the things you’ve seen people skip over and then have to rewrite later?

David Hardy [00:12:42]:
You’re right. I’d say it’s the same as any large-scale project that is being undertaken. And I think what I’ve seen a lot of — particularly with the in-house approaches — is that it’s very easy to generate a proof of concept, to quickly create something that fits the need, collects all the data, transforms it, does whatever they need, and stores it so it’s suitable for their requirements. I think that part is relatively straightforward.

But then, as you start to scale that, it becomes a challenge. You start to move into things like authentication, for example, which is something that is often overlooked, as well as data governance. Making sure that the right people have the right access to the digitally transformed data is critical. These are things that are often overlooked in pilots or POCs, but they do come back to bite you at the end.

David Brühlmann [00:13:38]:
Let’s make this very actionable for someone who has this exact question, perhaps even in a smaller company where you don’t have that many resources. Where should they start? What is the most important decision they should make?

David Hardy [00:13:50]:
I think the first choice a customer has when embarking on a data transformation or digital transformation journey is really the question of: build or buy? There are lots of great companies out there, including Thermo Fisher Scientific, that provide tools to enable this digital transformation. Obviously, there’s a cost to that, and there’s also a cost to doing it internally. So then you’re back to looking into the crystal ball and thinking about: how does this scale? What might the future requirements be around this data?

But I think that’s probably the first choice. The key thing is to understand that they’re not going to know everything at the start — and that’s fine — but they need to realize that will be the case. So don’t build something that necessarily puts them into a particular corner. They need the ability to scale and adapt.

David Brühlmann [00:14:36]:
So does that mean it’s important to have a system that can also adapt to growing or changing needs?

David Hardy [00:14:43]:
Oh, absolutely. The one thing we’re finding with data, automation, acceleration, and the expectations around data — particularly post-COVID — is that things are moving so quickly, at warp speed, as we put it. As soon as something is done, it’s probably obsolete. There will be a new technology, there will be a new approach, and that’s fine. It’s inevitable, even as you’re putting something together, that there will be a new way of doing it. That’s just change management, and you have to be comfortable living with that. Change is inevitable.

David Brühlmann [00:15:14]:
Absolutely. And we are in a fast-changing world. This concludes part one of our conversation with David Hardy. We’ve started pulling back the curtain on lab digitalization, the signs of fragmentation hiding in plain sight, and where data silos quietly drain productivity. There’s much more to come in part two. If you’re finding value in these conversations, please leave a review on Apple Podcasts or your favorite platform. Thank you so much for tuning in, and I’ll see you next time.

For additional bioprocessing tips, visit us at smartbiotechscientist.com. Stay tuned for more inspiring biotech insights in our next episode. Until then, let’s continue to smarten up biotech.