Software Technology Executive
Global Head of Healthcare and Life Sciences
Joining Cindi today are two guests: Jon Osborn, Software Technology Executive and former SVP, Chief Technical and Data Officer at Ensemble Health Partners, and Todd Crosslin, the Global Head of Healthcare and Life Sciences at Snowflake.
On this episode, Jon, Todd, and Cindi discuss how people in the healthcare chain of command can collaborate with one another more freely than ever before thanks to cloud innovations that make sensitive data safely and instantly available to all parties, why the countless benefits of moving to the cloud far outweigh the seemingly hefty upfront price tag (provided you adapt your strategy to fit your budget), and how a data chief can build a case to make these points evident to the investors, accountants, and higher-ups responsible for paying the bills.
Experiment with internal and external stakeholders. The level of collaboration between people involved in the chain of providing healthcare is enhanced by the cloud in ways that have only recently become possible.
Choose based on value. The price tag for moving to the cloud is real, but the benefits are too substantial to pass up - and there are hacks to adapt your strategy to align with your budget.
Be brave. The ROI for innovation pays back with interest, but you need the courage to make a case for all your various stakeholders.
To be in healthcare, you are either insane or you care; it's not the flashy vertical.
With healthcare ... you're talking about people's lives, so it is by nature an industry that is cautious. Because of that, it can be very frustrating. ... Being in health IT and being in data, my career has been a lot of bringing new innovations to light and [pushing] us forward, and yet there's this caution, there's always this fear of the unknown. It's a struggle, so you have to be a little bit crazy to do this to yourself. I've been doing this to myself for 20 years in leadership in 30 years altogether, and it's just such a frustrating thing, but in the end, I can look back at early parts of my career and I can see that the world is a different place because of [these innovations]. And ... being in the software space or in the data space, we're typically one or two persons removed from a patient; you're not that far removed, so you are helping to take care of that patient. I think because of that, you have to care.
With his background in retail, what does Jon see as the difference when making the case for innovation in the healthcare industry?:
I think one of the things that holds back healthcare is that some of the data is relatively complex and the relationships are hard to build. ... Data still has issues and new tools are needed, like Snowflake and ThoughtSpot, to actually solve some of these problems. ... Traditional relational tools just hold these companies back, yet they're so grounded in them. It's difficult to migrate. It's difficult to get any sort of agility when you have not only tools that have been around for 25 years, but also people with mindsets that are 25 years old.
So I think healthcare moves at a much slower pace, and I'm not quite certain [if] it's because of the level of investments are less than in retail or if competitively, it's just an environment that doesn't demand as much out of your systems. Like retail -- particularly online retail -- if you're not fully agile in that space, you're kind of out of business, [but] hospitals aren't necessarily going to go out of business because they have slow reporting or that kind of thing, but it holds them back from maybe producing a higher level of care or getting predictive or having some of these newer healthcare models actually work for them.
What does Jon see as a potential solution to this disparity?
From my perspective, I think data and apps need to finish the convergence. They need to come together. And the reason for that is there's a whole lot of people with a whole lot of really good ideas, but because they are not labeled a data person or whatever, then they get dismissed by the people who are the data people. And frankly, some of the best app people I've worked with have come from the data side and some of the best data people are English majors who are super smart. So I think getting these cross-functional teams built is really the win.
Todd weighs in:
Collaboration is everything. I had this conversation with a coworker just last week and in our remote workspace, it's like people think that, when they go on Zoom, it's only a virtual meeting and that they can't actually do work while they're in a Zoom. So there's this lack of understanding that you can actually collaborate, you ... have doc sharing using Google and Microsoft products [and] you can actually sit there for an hour and you're not meeting to have an agenda or whatever -- you're working and collaborating together.
Some people don't know they're a data geek until they start doing it. You put them down in front of ThoughtSpot, they start asking questions, and ... they get lost in it and they don't come back up for four hours. Then suddenly [they share] these massive insights, because of their background -- maybe they are a physician or a nurse; you need to pull those people together and collaborate.
Todd presents a real-life example:
I was talking about genomic data and [a practitioner said], 'Hey, I get a couple of variants inside of Epic. Now I can see a patient has several variants for their genomic data.' And I said, 'So imagine that you actually had all of their genomic data and you could run a query against all of their data against all of their variants and have that present as testing moves forward.' And they're like, 'Oh!' We need to move the democratization of analytics out to the edges so that people are constantly interacting with an aggregate query and not a list. I think there are phenomenal practitioners out there that can read a list and they can do their own analytics in their head in real-time and they are phenomenal ... but I think if you now give them five aggregate things instead of a list of a hundred things, they are only going to get better. That's my hope.
The thing that really excites me about it is that these newer tools require so much less upfront engineering, and that's important to me because [to understand] your data, you need to get it in a place where you can start to play around with it -- actually put it into a place where you can drive some early insights out of it and then make some engineering choices that are actually informed instead of trying to guess up front, which is essentially what we're doing in some of the older tools. [These newer tools allow you] to just go so much faster and then [quickly] correct the things that maybe didn't work out so you're not going to paint yourself in a corner.
Every piece of the healthcare puzzle has its own needs. The provider wants the best patient care. The patient only wants to share their private data as far as it helps their outcome. The payer wants to control the costs as much as possible. So how do you get the conflicting priorities of these different pieces to play nicely together as a cross-functional team?
Everybody's got their question they want to ask. What's conflicted is if you try and put it all on the same data model; you're not thinking about the places that matter for each one of those constituents. If you can come up with a way to let people have the answers that they need and share the data that you need to share, then you can drive that conflict out of the system ... and you start to see some really happy customers.
Todd on the role of experimentation in this field:
I'm insanely inspired by Elon Musk and SpaceX and everything going on with that. I'm a sci-fi guy, so I just look at the technology it took to make those rockets go up and then come back down and land like they do, and you see the little thrusters go out the sides, and it straightens it out and all that -- how was that done? You know it was modeling. You know it's a massive amount of data that they use to do that. So the thing that gives me hope is that I know that that type of thinking is going to make its way -- and has made its way to a certain degree -- into healthcare. It's really [virtual] experimentation on a scale we've never had before.
Is the future of collaboration and data sharing now?
I think a tool like Snowflake is going to drive some new thinking about how to collaborate with partners on data. You may have ... petabytes [of data], which is impractical to share over traditional ETL. You can't use a flat file. You're always pulling a fraction of the data out that you're going to share with some third-party vendor who's doing a service, but with Snowflake, you can just put that data online and they can have all of it. There's no worry about truncating the stuff that you really wanted or not being able to collaborate because you don't have enough of it.
Without getting too technical, we are a public cloud -- all three public clouds. We are multitenant on those public clouds. We have this unique capability to have, within the database, live sharing so that you're not having to dump things out to FTP and send a petabyte across the world. The best [example is the company] Starschema. It aggregates 27 sources of data from around the world, and it's all sitting in AWS West and then it gets replicated to 20 cloud regions around the world in real-time at every hour if any data changes and gets replicated as a service so that anyone on any one of those 20 clouds with a Snowflake account basically is getting this pulse of live data.
The data sharing is this core feature underneath the data marketplace, which is a public app store for data and private data exchanges [that] I think will be very prevalent inside healthcare and life sciences because you can't put people's [protected health information] out there in a public marketplace.
Todd explains how this applies in the real world:
I'll go back to the genomics example. That's big. We're in these conversations now with companies that know genomics really well and companies that know patient data really well. And the beauty is that before, they were not able to have this conversation to say, 'What if I could just join these datasets? I don't want to move them and forklift them from point a to point B. I just want to join them.' And now they can do that."
While the benefits are clear from a data perspective, how do we justify the unknown or less predictable cost of cloud computing?
I always get asked: 'Is cloud less expensive, equal, or more expensive?' The answer's 'Yes.' [For] clients who are replacing something that's on-premise, generally speaking, [there's] massive cost savings. That's what happens if you're doing the same workloads that you were doing before. But what happens when I say, 'Oh, wow, I can now do 10 times what I was able to do ... a hundred times as fast!' People [can] get a little bit tipsy on it, and then they're like, 'Oh, wow' [when they get the bill]. But what I always say, being more of a business CIO myself, is: 'What is your ROI?' There are lots of tricks and lots of ways [to adapt your cloud strategy to your budget].
Jon weighs in:
<blockquote">Totally agree. And I think a lot of companies or CEOs, they miss out on some of the cost savings available. My experience with both ThoughtSpot and Snowflake is that you don't need the same number of skillsets that you used to. You do see the people costs can be less. And sometimes it's a hard conversation. But if you build that into your model and maybe retrain some of these people and build that into your cost model, all of a sudden this license cost isn't as much anymore, because now you don't need a traditional DBA person who's troubleshooting queries, for example. You're not creating indexes, you're not doing this older-mindset-type work anymore. You're doing all this newer work. And to your point, Todd, that's where you have to be clear about what the business case is, and where the ROI starts and where it starts to end.
Jon is an expert in navigating a sea of no’s. What does he recommend to someone trying to build a case for a greater short-term investment that will result in long-term benefits -- particularly, innovating a move into the cloud?:
This business case problem is at the core of being successful, in my opinion, and the reason for that is twofold. One is obviously business cases help us justify spend. But oftentimes the investment in these technologies, like ThoughtSpot and Snowflake, also come with organizational change demand with them. It's going to look different; your team is not going to look the same. And so getting through the business case where you're really not just handling 'Well, can I pay for the license costs?' but 'How am I going to be organized? How am I going to govern these tools? What level of skills do I need? Do I have to retrain people?' Getting through all of those discussions is really at the core of navigating the nos so that you can actually get to the yes. And then, of course, bravery matters, too.
Todd Crosslin is the Global Head of Healthcare and Life Sciences at Snowflake, with a history of success in start-up and high-growth/complex environments. He specializes in strategic business planning, agile lifecycle management, liaison and mediator activities, team building and leadership, key contract negotiations, and software development lifecycles.
Jon Osborn is a Software Technology Executive and the former SVP, Chief Technical and Data Officer at Ensemble Health Partners, with a proven ability at the executive level to build technical organizations, direct and manage technical projects and architecture, manage agile software development teams, and work with business owners to deliver value and modernize technology.