Jim Ellis has led diverse, high impact teams across the world and held senior roles at industry giants like DuPont, Sabic and Microsoft, where he most recently served as Corporate Vice President of Global Construction. Now he’s helping organizations deliver safe, sustainable and innovative capital projects all around the world.
We recently sat down with Jim to hear his valuable insights on scaling data center projects and the role of modern construction strategies. Here is an excerpt from our interview:
Very large projects are initiated in the data center market. What key drivers did you look at while at Microsoft and how did you help evolve their process in these larger projects?
Jim: The simple answer is in order to scale, you have to have a systems approach. The systems approach becomes extremely important when you’re scaling from what data centers were, nominally 40 Megawatts to 100 Megawatts, to today where they’re sitting somewhere upwards of half a gig to a gig. That might be one of the biggest challenges in terms of how you drive a systems approach to getting these data centers to the point where they can deliver to both the company’s expectations as well as the client/customer expectations.
Today, one of the bigger challenges is because the scale is such a huge scale, and time to market is the key priority for data centers, we have to go fast.
But if we’re going fast we also have to go safe. We have to go sustainable. We can’t ignore cost, because cost still has to be efficient and we have to be able to drive that schedule and we’ve got to be able to drive quality and the assurance of quality across the end-to-end cycle in a systems way. So that at the end of the day, when you turn over that asset to operations, they have something that’s safe. They have something that’s reliable. They have something that’s efficient.
About half of data centers right now are using modularization in their project strategy and 93% of data center owners and contractors say they plan to use modularization in the very near future. How do you see that as aiding in delivering across key drivers?
Jim: Modularization is something I have been championing since the early 1990s and I became a believer very early on. Modularization has the ability in a safe environment and controlled manufacturing approach to deliver expectations with more certainty than a stick build construction job. When you’re in a closed environment for manufacturing-driven prefab or modularization or integration, you’re going to buy better safety. You’re going to be able to improve your quality assurance. You’re going to be able to get efficiencies on both cost and schedule delivery. The technology and approach to this manufacturing has improved so much in the last 30 years to the point where, yeah, a lot of the data center providers today are exploring that. The other piece is also about eliminating the constraints.
One, power. Modularizing the power availability to the site of construction, maybe even early on into delivery to operations, through these bridge power, modular strategies.
Second, permitting. You could have the module sitting in the prefab yard while waiting on permits. The building itself can be built in a modular fashion on the site and save time. I’m hearing all kinds of stories from different providers that can come in and cut the time of a building install in half, just by putting in modular buildings.
And finally think about the shortage of labor. Modularization helps encourage folks to be in the facility environment versus being out in the weather, mud on boots and all that kind of stuff that goes into construction.
So, is it pivoting over to be a primary strategy with many of the hyperscalers? Yes. With all? No, but I think they’re all exploring at this point in time. Will it evolve with time to be the prominent strategy to go to construct? Yes. You’re seeing it already and I think it will continue to grow with time.
How do you balance safety, quality and cost to manage that process as a system so that you deliver long-term value out of these data centers?
Jim: Safety is a core value, that’s not something that anybody is going to compromise and I really feel good about our industry both in the industrial side and commercial side that across the supply chain, that’s not just a priority but a value because at the end of the day it’s about people.
In the early part of my career I was told about the three-legged stool; you can have cost, you can have schedule or you could have quality – but you could only pick two out of three. So we made trade-offs based on that early in the project cycle. What I find today in the data center world is that those trade-offs don’t happen as they used to because the stakeholders say, ‘We have to have it all.’
It starts very early on in the decisions being made around site selection, decisions being made on how to develop the project, who you’re going to develop the project with and what degrees of freedom you have in order to deliver across a spectrum of KPIs that have to be met. We still have priorities in the market, but it doesn’t forgo safety, quality and cost.
When we talk about safety, it’s usually about the individual worker and not the lifecycle of the facility. We maybe inadvertently sacrifice quality because we’re chasing schedule. What do you think about the way we incentivize contractors on a project to deliver faster?
Jim: All teams involved in a project feel the pressure, so we make these decisions. Folks will say it’s prudent risk, I say it’s risk that’s not always prudent. That whole balance of making sure QAQC happens across that lifecycle and it doesn’t get pushed out. Like safety, I think about it more as a core value. When you do that, it’s going to reduce your rework.
You don’t have to do the work twice, but it’s also going to create a lot of other opportunities around risk reduction. Mitigating things like schedule delays. Mitigating cost overruns, mitigating warranty issues, which can be very costly. If you want faster time to market – proactive quality controls, streamline things like the install process, the commissioning process. Reduce the retesting. There are tools right now that can help, like Cumulus, that add tremendous value to streamlining these processes.
Quality is sometimes viewed as a given, but we’ve been told you get what you inspect, not what you expect. What have you seen in terms of rework rates in data centers vs. oil and gas?
Jim: The rigor in oil and gas because of the process safety management concerns, seems to be more robust and a lot more focused. Let’s start with the build itself. The rework that goes on in a construction environment is probably in the order of 1% to no more than 2%, which is fairly low. That’s not just having a good QAQC program. It’s also associated with good planning, making sure you understand the scope. The scope is frozen. You can deliver on that with certainty because you know exactly what you’re doing.
The data center world is a little bit different, and a lot of that is driven by speed to market, right? Speed trumps everything else in that process. I didn’t see the same level of rework in oil and gas as I did in data centers. In the broader industry across data centers, I would give you a range on average of 7% or 8% and some of those required stoppages too because of safety incidents. And so there’s an opportunity there as I see it.
Our full interview with Jim has released on our Work Done Right Podcast, as part of a new series focused on the data center industry. Be sure to subscribe here to be notified when new episodes air.
