Host Mark Thiele discusses business strategy and looking at IT processes for rapid growth with guest Bill Dougherty.
Bill Dougherty is an industry veteran with over 25 years of experience, designing, building and operating scaleable, highly available, and secure systems. Bill leads the IT, Security and Support teams for Omada Health, the leading digital health company focused on the prevention and treatment of chronic diseases. Prior to joining Omada in 2016, Bill was the CTO and CISO for RagingWire data centers, ran site operations for StubHub (EBAY), and lead the security team for Copart, Inc. In his spare time Bill is a father, photographer and notorious riverboat gambler.
Mark Thiele: Hi, Bill. Thanks again for joining me on this podcast.
Bill Dougherty: It’s great to be here.
You are CIO for Omada Health. Is that correct?
VP of IT and security and support and jack of all trades for Omada Health.
Wow, that is a jack of all trades; a couple of titles in there that struggle to find the same head in most organizations these days. What’s top of mind for you in your role at Omada Health?
We’re having a lot of fun right now because we are growing like a weed. We are expanding rapidly this year, bringing on lots of new customers and helping lots of new participants. We recently just passed a couple of big milestones for us, the first one being we have now enrolled over 250,000 people in our digital healthcare program.
Not too long after that, we recorded our 50 millionth weight value. One of the aspects of our program is we have connected scales that people will weigh in on. That’s one of the health values we track. We now have 50 million weights in our database, which is pretty cool. It gives us one of the largest behavioral health data sets in the world.
That’s really amazing. What you just talked about will probably align well with some of the information we shared before this as far as some of the questions are concerned, so thank you for that. That’s an exciting space, and lots of opportunity in it still. Thank you. Before we get into the rest of the geek stuff, what gets Bill excited when he’s not in the office?
That’s a great question. I happen to love my job, so I do a lot of work even when I’m not in the office. When I’m not working, I’m taking pictures. I’ve been a photographer since before I was in high school. On any given week I shoot a couple thousand photos, lots of my daughter and then lots of landscapes and other things.
Awesome. One of these days you’re going to have show me some of your work.
I’d be happy to show it to you any time. Anyone can see it at WLD1.net.
There you go. Excellent.
A plug for the side gig.
That’s right. Awesome. Thinking about what you’re doing with Omada right now and thinking about the term ‘digital transformation,’ ‘digital disruption’–I prefer to call it ‘business transformation,’ not that it’s critical one way or the other. I feel like we’ve all been doing digital transformation for 50 years of IT delivery. Now is really the time to try to match business function capability with the best that IT can offer into business transformation. Thinking about that from a healthcare standpoint, even as an interested observer, it seems like there’s lots of opportunity. What do you see as some of the biggest areas of opportunity for you guys as you do this?
It’s amazing. It’s why I joined Omada. Healthcare is ripe for innovation and is ripe for disruption. There are a few companies that are out there that are doing a great job of it. The foundation for our company was actually a study the CDC put together many years ago around how to prevent type 2 diabetes. Type 2 diabetes is a largely preventable disease.
We know what causes it; being overweight and inactive and eating poorly. These things elevate your risk factors exponentially. They did a study to compare the best drugs on the market with intensive behavioral counseling, and it turns out that intensive behavioral counseling is really what is most effective. You get people to eat better, exercise more, reduce stress, lose weight, get some sleep, and their risk factors come way down. The trouble with intensive behavioral counseling as it was done originally, is it was all in person.
There’s 100 million people in the country that are at risk for type 2 diabetes, roughly. There aren’t enough coaches to go around. Even if you could grab all 100 million people, there aren’t enough professionals to help them. This is a problem that technology can solve. Through the use of iPhone and Android apps, through the use of connected devices and digital coaches, you can meet people where they are in their life and give them the tools necessary to live better. If you do that, they have a much better life. They get healthier.
For their employers and for their health plans, it plummets their health-related risk. Everybody wins. That’s kind of what we do at Omada. Then we’ve gone beyond just the simple prevention of type 2 diabetes to other related diseases. We’re seeing this throughout the healthcare landscape that there are these old true stable business models that are just begging for digital disruption.
In order to do that, you have to do it in a very methodical, careful way because now you’re dealing with people’s most sensitive data. There’s [very] little that you have about yourself that is more sensitive than your health data. You’re dealing with lots of federal and state rules around the processing and storage [of] and access to that data. It’s just a really exciting time to be in this space. I have never been in healthcare before I joined. I love what we’re doing.
It’s amazing stuff. Just to highlight a little bit of that, when I think about what you just talked about and I think about business transformation, so much of what I would assume to be the fundamental opportunities of business transformation are built into the service you guys are building and delivering. That’s really creating a better customer experience. A better customer experience can be measured a lot of different ways. My God, if a better customer experience is actually that they live a healthier life at the same time as they feel better about the relationship that they have with their providers, that seems like a major win-win all the way around.
Exactly. It’s so measurable that we then went a step further. We’re one of the first providers out there to develop outcomes-based billing. People join our program, and we want them to get healthier. We only bill for the people that are successful in the program. Our motivation is 100% aligned with the participants’ motivation and with the person paying the bill, either the employer or the health plan.
Everybody wants them to get healthier. Everybody is incentivized to it. We are able to transform from the typical doctor-patient [relationship where] every time you show up, you get a bill; to now I only get a bill if I got made healthier.
That’s pretty cool. First of all, congratulations. I think this is an exciting space to be. I’ve even made reference to healthcare in a couple of talks I’ve done recently, even simple things like just helping to remove delay confusion and stress from the life of a customer of the healthcare system is an enormous win. Hats off to what you guys are doing. I think that’s fantastic.
Putting [on] the CIO hat relative to what you’re doing, do you see the CIO or a role similar to the one you’re in right now as being in a great position to lead, at least from a technology standpoint, if not even from a process and opportunity standpoint, for projects like this?
Absolutely. I see it as a critical role in organizations for the CIO to look at key business processes, key challenges throughout the organization and figure out how to apply current and future technologies to those. We try to do that every day.
That’s fantastic. As an IT leader, what are some of the things that you might have done differently or that you could even argue you lucked into, if that makes sense, that made this process more successful for you at Omada?
I lucked into a lot of stuff because I inherited a pretty good organization and a pretty good company. One of the things that was different at Omada from previous organizations I’d been at is Omada is a cloud first and SaaS first company. For our core product, which my engineering team runs, that is all AWS. It was born in the cloud, lives in the cloud. Everything is cloud.
For the portfolio of our internal IT systems that I run, almost 100% of that is SaaS. I think I own two physical servers. What that does is it changes the role of the IT leader from being someone who manages lots of infrastructure to being someone who manages lots of vendors. It changes the trust relationships. We now have these very complicated four and five and six-way trust relationships that we can talk about endlessly.
It also gives me lots of flexibility in terms of how I bring new solutions to the company for complex business problems because I’m not just restricted by ‘What are the solutions I have in place? What’s the infrastructure I’ve already built?’ I try to keep abreast of what’s new and different and exciting out in the marketplace of all this wonderful innovation that’s in the SaaS community and can very quickly pivot to a new solution if it makes sense within the context of our business.
That’s fascinating and outstanding. Would you also argue that–and this is the part that might spin my beanie as much as what you’ve already talked about or more–do you feel that putting yourself in this position gives you a better opportunity to be a business person with the rest of the business and less of a wrencher that occasionally gets invited to business meetings?
I would like to believe I’ve always been a business person, but it certainly helps. It changes the nature of the conversations because in past iterations I would spend lots of time managing large capital budgets. We would be talking about nine month to two-year project cycles for large projects. It was different.
I would say I’ve witnessed many organizations where the rest of the leaders, if they have to call IT, they groan. They just know if I’m inviting IT into the meeting, that means it’s going to be difficult, costly, time consuming. The idea is to not make it difficult, not make it costly, not make it time consuming, and make it easy, fast, and cheap, but to do it within whatever the rules are of your organization.
In my organization we’re in healthcare. We have certain constraints on us. Some of those are bound by HIPAA, by federal law or by state laws. We do certain things because we are bound by those constraints. Once people understand those constraints within the boxes you create, you can act really fast.
That’s awesome. Not that you need my opinion, but I would argue that you’re doing this the right way, Bill. Congratulations.
Considering that you’re freely admitting that you kind of lucked into a good organization and a company with the motivation to do this the right way from the beginning, you must be able to think about what you have versus what others struggle with and have some advice for trying to get to the model organization that you’re talking about, one that is more about doing things from a business perspective and less about ownership and silos, etc. What kind of advice would you offer to a CIO who maybe has an organization that’s a little bit more legacy than yours?
I’d say two things. One, be willing to let go of some control. The other is be willing to write off some costs. The second one is pretty easy to understand. You’ve got investments you’ve made. Those investments now are an anchor that at some point you need to move away from them, even if it’s costly because it’s going to cost you more in the future.
The first one is the bigger point, which is be willing to let go of some control. We have a very democratized IT organization at Omada, by which I mean the individual business owners and business groups for the most part select their solutions and oftentimes manage those solutions. They do it within a framework that we have created.
We actually have what I would [call] four categories of IT. We have generalized IT, and generalized IT applies to the whole company like our email system that is owned, operated, and managed 100% by my organization. We then have departmental IT, so a good example of that would be a collaboration tool that just designers use. There’s five or six power users, and they’re the only ones who need that tool. They go through me to get access to that tool. We integrate it with our single sign on. We manage the contract and the ‘spend,’ but they administrate it on a day-to-day basis and track the new features. If they want to change to a different one, they’re the ones that control that.
The third category is similar to that departmental IT, but it’s stuff that’s connected into production. We have departmental apps that have no access to any production quality data of our core product. Then we have others that share data. A good example of that is... I’m fairly public with the fact that I use Zendesk. I’ve done a case study with them. We use Zendesk for operational support. We use it for all kinds of fun stuff.
We also integrate it directly into our core product. It is sending and receiving protected health information. Those applications, I may run it or the department leader may run it, but there’s a much higher requirement of governance because now we are connecting it into the keys of the kingdom, which is production.
The last thing is production itself. Our head of engineering runs that. That’s the core product. That’s how we deliver our system. If you think about that in the context of any general IT organization, all four of those are important. If any one of them isn’t working right, you’re having a bad day at your organization.
There’s only one category that I control 100%, and that is the generalized IT that everybody uses. Everything else I’ve democratized out, which means I have given up some control. I’ve maintained control over the things that we need to maintain control over: budgeting, terms and conditions, price negotiations, integration to single sign on. Beyond that, I don’t need to care about if they’ve added a design or removed a design or turned a feature on and off in that departmental app. It doesn’t impact my life. They can use the app any way they want and can make better decisions for them than I will ever make for them.
That makes sense. Go ahead, sorry.
I just gave you a lot to unpack there.
You did. I’m trying to swallow it all. It’s fascinating stuff. I’m going to have to just come in and spend some time with you because we don’t catch up enough.
Part of the problem we’re trying to solve is we don’t want rogue IT. People who get afraid of SaaS, it’s because they don’t want rogue IT. They don’t want stuff hanging out there that they don’t have any visibility [in]to. Partly because I’m in a regulated space and therefore I have the big stick that comes with that, and partly because of coordination between my finance team and my legal team and my compliance team and myself, we had some rogue IT when I first got here.
We’ve wrangled it all in and said “If you go rogue, bad things will happen. If you don’t go rogue, we will offload some of the burden off of you. We will handle your renewals. We will make sure there is money in the budget. We will make sure it’s integrated with our other tools.” Therefore, the business leaders are incentivized to work with us rather than go around us.
While some could assume when you say ‘rogue IT’ you’re talking about all shadow IT, you’re not actually. What you’re talking about really is that if someone is going to buy an application or use an application that wasn’t originally funded out of the central IT budget, that’s one thing. It’s another thing to be doing that without being able to apply any of the controls that protect the business or ensure the investment is successful long term. Is that correct?
Yeah. What we’re trying to do is we don’t want shadow IT. We don’t want anything in the dark. Do it out in the open. Tell us what it is you’re trying to do. Let us help you.
We created a lightweight process that lets people onboard vendors quickly. It makes sure my legal team is happy with the terms and conditions. We rarely just accept click-through conditions. We make sure we know how it’s going to bill so it doesn’t get turned off because it was on some employee’s credit card and they left. We make sure it matches all of our document retention requirements and our treatment of protected health information parts. All of that gets handled in this framework that still lets the business owners choose the tools they want, have a lot of control over their day-to-day usage. It’s the best situation I’ve seen for dealing with shadow or rogue IT.
I like it. I think it’s the best way to go. I’ve argued off and on, not specifically about rogue IT, but rather about the risks of creating an environment where people aren’t talking to each other. If the risk was bad when people were installing small databases and creating their own web pages back in the 90s on PCs in the marketing group, it’s theoretically way worse now.
The amount of data, the amount of company that could be given away by accident or costs that could be given away by accident with today’s availability of SaaS tools means that a lack of visibility is a real risk for the company. I think what you’re doing is outstanding. Hats off to you.
We talked for a little while about transformation and how it relates to the healthcare industry and how IT in what might be argued in a model way, could be operated.
I want to take us to a different discussion now about edge computing. You and I shared a couple of comments via LinkedIn recently where I’m not positive we both disagreed with each other, although that’s fine if we did. Maybe it was context. From your perspective, talking about edge, I am well known as being a vocal stumper for the potential opportunity of the edge marketplace. We started talking a little bit about edge as it relates to data centers specifically.
What’s your perspective on the opportunity versus the hype of edge? When you talk about the data center market, I’ll give you a couple of feeders, and you can pick and choose from those or come up with something else. When you think about companies that are out there building technologies that are meant to populate in edge–that doesn’t actually have buyers standing in line to buy yet, at least not visibly, what do you think about that? What do you think is likely to happen there?
It’s funny because somehow I’ve become the ‘edge contrarian’ on LinkedIn. I did a webinar on the edge with a data center guy who was building edge data centers. I was saying I thought his business model was way over hyped. I think the potential for edge depends on how it is you define what the edge is. What I’m referring to is people building co-location data centers in tertiary markets.
You and I have both been data center people. I have a nice big campus in Sacramento as well as Ashburn and Dallas. I’m pretty familiar with the industry. I have long said that I think there’s more hype than reality in terms of the demand for edge data centers, which is very different than the demand for edge computing. I think there are absolute workloads where edge compute makes sense. For those workloads, I don’t think it’s going into a colo. I think it’s going into the device.
The theory behind doing edge data center is latency. Roundtrip time really matters. To me most applications that isn’t actually true for if you’ve properly coded your application. For the applications where it is true, you’re going to want the calc done on the device anyway.
Everyone likes to talk about self-driving cars. Self-driving cars need 5G. They need edge data centers. What they actually need is for the car to make decisions without the packet ever leaving the car. The same is true for lots of applications.
I’ve also been saying that I think the enterprise data center is dying and that everything is going to move over time to large centralized data centers, either wholesale colo or public cloud, things like that. Again, where that breaks down is where you need large amounts of edge processing power. Let’s say you’ve got a very large factory and your robot is making real time decisions. There I think you need edge compute, but I think you’re going to put it on premises. If you’re latency sensitive, either you need a sub millisecond or you can withstand the roundtrip time to one of the major markets.
I don’t see the in between where I don’t need sub millisecond, but I can’t tolerate 20 milliseconds. I’ve got to be in two. Therefore, if my factory is in Nashville, I need a data center in Nashville. I could be wrong, but I think that market is way over hyped and certainly not where I’d be placing my bets. I’d be placing my bets and did place my bets on major markets.
We all know what they are: in Ashburn and Northern California andDulles, Virginia, Chicago, places like that. The only reason I was in Sacramento rather than the rest of the Bay Area really was cost in earthquake risk. I actually thought it was worth the extra two milliseconds of latency for people to drive the earthquake risk down to zero compared to the Bay Area.
Right. A lot of what you said I would outright agree with. Some of it I might debate a little, but more on perception of what the market could be versus something that I write in stone and know for a fact that it would happen. Those are very different things. One of your last points about the factory with the data center versus a factory with on prem, I tend to agree with that. When I think of edge computing, I generally don’t lump on prem computing with edge computing.
If that were the case, then every company that has its own data center has edge computing for that campus that that data center is on. I don’t lump those two together. I think you were doing the same thing in not lumping them together.
But I do agree that if I have a factory or I have three or four locations for a factory, it doesn’t make sense necessarily to worry about trying to figure out an application that is geo distributed and that I could install on a lamp post or I could install in a CVS back room that’s been purposed as a data center or in somebody’s one megawatt colo somewhere. [If] I’ve got three or four locations, it probably makes sense for me to put that rack of gear or half a rack of gear, whatever it is that I need, right there next to the equipment. You could call that edge if you want, but realistically it’s on prem IT for that function in that discrete environment.
I actually was essentially calling that ‘edge.’ My greater point there is if you think about how that company operated ten years ago, ten years ago next to that factory they had a big data center footprint because in addition to the real time control systems and the monitoring systems and the SCADA systems, they also had their ERP and their email and their file server. They had everything. My real point is what’s going to happen as you take all the stuff that isn’t latency sensitive and get it out of there because your cost of real estate is too expensive–you don’t want to manage it. You put it in a colo, some place central.
Anything that is not latency sensitive, get it off your books. Then that valuable factory real estate is used just for the applications that have to be there. You make that footprint as small as necessary to run the factory and no more. To me that is no different than putting ten servers on a rack in a factory and the same as putting a computer into a smart car. You’ve put your computational power next to the use that requires super low latency. You move everything else out. In that case I’m calling it ‘edge.’
That’s fine. The truth is that there are still many definitions to be made and many perspectives to be captured relative to what the “edge marketplace” might become or the different aspects of edge might develop into. As much as I spend a lot of time in it, I’m not going to sit here and say I am the oracle who knows all the answers. I think this is a good perspective, and I appreciate that. Thank you.
We’ve come pretty close to 30 minutes already, but if you’ve got a minute or so, tell me a little bit about what you see as some of the companies you’ve looked at or heard about that you think might win or fail in this space?
In the edge space specifically?
Yeah. You could be specific to data center or data center products if you want or some other app delivery model or whatever it is that you feel most comfortable with.
It’s funny, I think that the CDNs keep winning. I think Amazon and Azure and to a lesser extent GCP win. I think the large wholesale colos, so the digital realties: Equinix; my alma mater, RagingWire, Switch. Those companies win. I think the tertiary data centers will lose or will just get bought up. I’ve been wrong before and will probably be wrong again, but that’s how I see it.
I see the industry continues to consolidate.The data centers and colos, most of them are REITS. It’s a straight up commercial real estate play. Things like cap rate matters. It lends itself to consolidation, especially when interest rates start rising and cost of capital starts rising. I think for most people, Amazon has got enough distribution to deal with latency needs.
That’s just where I see the market going. That’s also from the perspective I’m in now. I think three years ago if you had asked me, I might have had a different answer because I was at RagingWire and had a different perspective.
I think that’s a healthy perspective. While I might put up some suggestions for some alternatives here and there, I’m asking for your opinion on this webcast, and I got it. I appreciate that. I am not going to sit here and say I believe anyone else is wrong so much as we have slightly different perspective on how the future might play out. I appreciate both your experience and your well thought out words in this space.
I just want to say, Bill, it’s great catching up with you again. Thank you very much for setting aside some time to join me. I hope we can find some time in the not too distant future to get together over a coffee or a beer or something.
It’s my pleasure. I love talking about this stuff. I love sharing ideas. I love having people in the comments and things tell me where I’m wrong and prove my thought processes. I think we’re all finding our way in this transformative moment together. The only way we’re going to do that is by sharing the best ideas and weeding out the bad ones.
Being open to the fact that a different opinion might actually be as good or better than the one you formed yourself–I face that every day. Bill, thank you very much. Enjoy the rest of your evening. We’ll talk soon.