Design vs. Operation

Gauging Building Performance at Columbia’s Manhattanville Campus

In February 2023, we met with members of Columbia University’s operational team about the challenges of submetering its Manhattanville campus. Finally able to sit around a conference table, we talked about how the team is collecting and analyzing this data to meet sustainability goals, better understand their building systems, and turn stakeholders into efficiency champions at all levels of the university.
headshot of Molly McBeath
Molly McBeath, Content Writer, utiliVisor
Aug 07, 2023 (20 min read)

This hour-long interview has been edited and condensed for clarity. Click the video to also open the full transcript.

Tell us about what was going on at Columbia before utiliVisor came on board? What struggles were you having? 

Sean: Tim, I’m actually going to defer to you on this one. Since you originally received the RFP for the players that are not necessarily in this room, can you talk about how Phil approached you, and then kind of how we integrated in, then we can talk about all aspects of what we're doing. 

Tim: Absolutely. About 10 to 12 years was when the campus started getting built. And once it was built, we originally got called in about 6 years ago to look at the retail metering on campus. And then that kind of went away for a little bit. And then recently, we got called about a year and a half to two years ago because they had trouble getting metering data into an easy-to-read platform that was digestible, so they can make decisions about construction on the campus. So there was an RFP developed, there was a specification developed, we applied to it, we submitted pricing, we got approval. And then we started on the ground working to grab that information. The meters were in place because Manhattanville's a brand-new campus, so we tried to leverage everything that was on campus already that take that information. And there was a lot of work done that Mike and his team did prior to us arriving – for a year or two years – to kind of integrate that together. I feel like we were the last 10% to 15% to get that up to a platform that can be manipulated, understood, and distributed to people so they can take the energy information to make decisions about it in the future. 

Sean: Yeah, I'm the new member of the team with Manhattanville. I came on board this summer. Alex also came on board this summer as well. And we were looking at Manhattanville to be our pilot case for expanding submetering throughout the campuses, less about allocation of utility but more to understand how the campus assets, buildings, and plants consume utilities, and know more and for the future, since we are based in New York City, how that impacts our carbon emissions. And I'll echo what Tim said, years of work by Mike and the ops team at Manhattanville made this very easy for us to integrate the utiliVisor platform to the campus. Manhattanville had a bunch of forward-thinking with respect to submetering. I believe we have 193 thermal electric meters across 1.9 million square feet and additionally domestic and hot water meters that I'm not even adding to that number. They just needed a method for being integrated from BMS and SCADA back to something that was manageable and digestible for a lot of different aspects of the university.  

Mike: And me being the person who's been there the longest. So the initial design was to have all these utilities go to a centrally located place, but due to circumstances like COVID, there were multiple delays. And so in the past year, we've now worked with the utiliVisor to centrally locate all these different programs that Sean just indicated with two different BMS systems and a SCADA system and a centrally located hub center. So that's where we were before, until utiliVisor came on board. 

How many buildings are we talking about? 

Mike: There are five main buildings currently on the Manhattanville campus. There are two Columbia Business School buildings, Kravitz and Geffen. There's the Forum building and a research building called GLG. And then there's also a performing arts building, the Lenfest building.  

Sean: And obviously, there's a central energy plant where we take a vast majority of our metering inputs to those buildings to really understand how not only the buildings are performing but how the plant is performing as well.  

Mike: To follow up what Sean was saying about being based in New York City (and therefore needing to comply with Local Law 97): There are three boilers, two electric chillers, and one steam chiller. And part of this data is to see how we can trim down our carbon footprint as per Columbia's goal to go carbon neutral by 2050.  

How much of this project was prompted by Columbia's Sustainability Plan 2030?

Sean: I think this was initially spurred on, this relationship we have right now, for past reporting LEED requirements and NYSERDA grants from previous construction. So there was an appetite to understand the building's performance, looking backwards, and then understand how the campus interacted with the central plant for moving forward for the eventual expansion of the campus. The campus sits right now at about 1.9 million square feet, the final campus build-out, it's going to be roughly 6.9 million square feet. So really, from the development side, there was a large appetite to understand how the actual building loads were performing against design. The team that Alex and I and Mike as a result of running the plant are part of now are looking really from an operational perspective.  

Our view is how are the buildings performing now looking at both CU Plan 2030, which is our internal sustainability goals for carbon reduction, and then also understanding how we are performing against our regulatory commitments with Local Law 97. So we jumped on the opportunity, right off the bat, to partner with utiliVisor to look at both past performance for development and future performance for operational needs. That's why I think this is a really interesting project, because there's a bunch of teamwork across several disciplines to get to where we are right now.  

Mike: And to add on, this data has not just been for our own benefit. It's also going down to the development team, called the Manhattanville Development Group (or MDG), for the next version. As Sean said, What is the next piece of equipment that's coming into the plant? What designs need to be changed to correspond to our goals for 2030? Or net zero 2050? What equipment will be the future for the central energy plant?  

Sean: The MDG group is responsible for the design and construction of new buildings at the Manhattanville campus. And those buildings get turned over to the Manhattanville operations team. Mike is part of that operations team, and I oversee the central plant. But then from an energy perspective, I have the second part of my team, which Alex is a part of from Manhattanville, which is deep energy analytics and analysis to drive toward our goals. Alex, maybe we can talk about how this benefits you? 

Alex: Yeah, having any sort of historical information is really necessary for us to identify where to actually do projects down the line. Not so long ago, there wasn't much in the way of good data even on a building level to understand what was happening across all Columbia, not just at Manhattanville, It's just monthly data and there was acknowledgement of some issues with collecting the data. And there were also many snapshots at a certain period of time. So what we really need to develop is just going deeper and deeper in terms of historical energy consumption at more and more a micro scale right now. We're getting down to the heating loops’ and cooling loops' level, and in the future, hopefully, we’ll be getting to the space level. 

Sean: The beauty of having this platform now and understanding the data at a 15-minute interval is that Alex and his team are able to then perform regression analysis, which feeds back into our utility budgeting for future financial cash flows for the campus. So again, it's a collaborative group, and there's a lot of different facets in the organization to take advantage of the product we were able to put together. 

Mike: That's where utiliVisor comes in – the validation on the metering itself. Going back to look at issues, is it internal with our system, our data collection, or is it the actual meter? So there's this collaboration with human beings, not with other machines, saying, Hey, there's an error!  

There's been a big collaboration between myself, Alex, and Sean along with the utiliVisor crew. Ed has been one of the lead people with us from utiliVisor, so it's a human interface. Instead of just dealing with emails and computers, Ed will pick up the phone and say, How do you guys think things are looking? There's a back and forth on the database that we're looking at. So if Sean wants to see something particular drilled down, Ed is right on it and working on it. Or if Alex notices a discrepancy, Ed looks at whether it's on his side or if it's on our side, on the meter or in the manipulation of the data going through the two systems.  

Sean: That's what's been really impressive to me is the help, how we've just continuously followed an iterative path with this, with cooperation between Columbia and utiliVisor the whole way. You know, utiliVisor notices a discrepancy: they communicate to our team. We notice a discrepancy: we communicate to the utiliVisor team. We have a completely customized product, and the response from utiliVisor has been excellent the entire time. There's never been a “No, we can't do that.” It's a “Well, let's figure out how we can.” 

So what was the moment that you decided you needed to call Tim and Ed at utiliVisor? 

Mike: I'll take that. It was a different division, the Manhattanville Development Group, that started the process up. And once they got on board, we kind of added onto it. We saw the opportunity, as Sean said, that once we saw that utiliVisor was working on a product for the development group, we thought this was a perfect time for integrating the carbon footprint side of that into our portfolio.  

Sean: We just hijacked it from MDG. Honestly, just like Mike said, we saw an opportunity. And Tim and his team were ecstatic about helping us create exactly what we wanted from an operational perspective. And while the development group got utiliVisor in the door, I think this group here is really taking over to build out the product that everybody needs and that everybody can take advantage of. And that has been the fun part of this. Everybody’s got a piece of the puzzle. And we're still working on trying to get even more groups involved. Because we see the power of being able to see deeply into our buildings and deeply into our plant in real time and have interfaces to allow us to manipulate the data in any way we need.  

Were there any obstacles to the project?  

Ed: One of the major obstacles at the Columbia Manhattanville campus is that while the campus is relatively new, in a typical large campus situation, as new buildings come on board, the construction team will typically take the lowest bid options for the controls – let's put it that way. Specific to the Manhattanville campus, they have a BMS for the buildings, they have a BMS for the central energy plant, and there's actually an additional BMS for one of the other buildings on-site. So there's three separate BMS's, and then there's the IT department.  

The challenge, in particular to the site, was security. We run into it at multiple locations, but Manhattanville had additional levels of security that we needed to overcome to get the data out of the building. The first layer of complication was integrating different data from different automation systems, from the SCADA to the simple BMS systems. And then the additional layer was getting that information off campus to our data collection of the site. An added layer on top of that was some of the metering was never networked together. The meter was installed, but there was no cable. So that meter information is there: you can read it at the cabinet and write the number down, but should you want to, extend out that information over an interval period, they couldn’t do that. They have some very advanced equipment on-site at Manhattanville, but they didn’t have that capacity. Getting all of what you paid for is what utiliVisor likes to do. Moving that capability forward was a big part of the project. 

Mike: And it was a step-by-step process. At the end of the day, it was a compromise between security and dealing with limitations of the communication to get the data through. 

Ed: I have no problem coming on site. utiliVisor’s a relationship, it's not just a platform. 

Sean: Yeah, the hardware was there. I think the networking was the biggest challenge. And just like Ed said, the various different technologies and systems we have are not consistent from building to building, or building to plant. And that provided some challenges. But we overcame every single one of those challenges in pretty rapid fashion to be able to get a pretty comprehensive view of the campus energy data. 

When you say rapid, what kind of time frame are we talking about? 

Ed: We went in in August, and I think it was metered by September.  

Tim: I think we were fully operational in our system and collecting data within a month and a half.  

Ed: A lot of that was the team on site.  

Tim: This was a unique situation, because everybody was so involved and so motivated to get this thing up and running. We would say, “These are the steps we need to take; please do this.” And things got solved pretty quickly, which was great. 

Sean: Having the motivation from a lot of personnel at Columbia made this a very, very fast process, and then getting utiliVisor directing us on how to achieve that technical success was why this went so quickly. And that goes back to the relationship, the open dialogue, and the very custom package that we created, which is very important to us. It's an iterative process. Every week, we come up with something new to meter and a new way of metering, a new way of collecting data. And that communication occurs constantly. It does help that there's a very technical, motivated team at Manhattanville. But without utiliVisor's guidance, we would have never gotten to where we are. 

Tim: I think you pushing the boundaries on us makes us do our job better and opens up our eyes to different opportunities also. So we take that information, we go back to the drawing board with our development team, and then integrate that request into our system as well as our internal processes. 

So is it accurate to say that everyone approached the project with a beginner’s mindset? 

Sean: I would say that's 100% correct. Once we finally made the leap of faith as an organization to get into collecting data and submetering, I think all that past effort really motivated the team to efficiently get through the networking problems and the calibrations required from the meters. A lot of things converged at the right time – getting the right people on board and expanding the importance of why submetering and why data collection are really our only option moving forward.  

Mike: My motivation with the submetering data is to show how we're running the plant and the facility, so we can prove to other facilities and to upper management that we're running the system as efficiently as possible. Having proof of that, actually being able to put it on data and pivot tables so everybody can see what’s going on, helps out a lot to say we're doing a great job. And you can show your employees how we compare to other facilities. 

Alex: Also, like what Sean was saying, we had all this existing metering equipment that maybe wasn't being read anywhere. Or if it was going somewhere, nobody was paying attention to it. Ultimately, we need to roll this up to the rest of the campus. So part of the motivation is to show the value of doing this in the most well-equipped space that we currently have as well as to take whatever data we can from it and use it in the most valuable way for as many people as possible.  

What made you pitch this internally and to Columbia as a whole? 

Sean: I'll speak from past practice. I happen to come from another campus site that had over 2000 submeters across the site that all fed back to a centralized system, it was primarily for billing, as it was a commercial property. However, it allowed my team of energy engineers to do great things with it. And I saw the power not only from the billing and revenue recovery side, but from the energy efficiency and the M&V side of the business to really understand that small changes in a building can actually have a great impact. And being able to see things granularly and be able to track a trend can have a really big impact on a building's performance. So I brought that mentality with me to Columbia, and that was part of my introduction to Columbia. And what I found was that there were many people around me who had the same mentality, and we just happened to kind of get together and team up right at the same time.  

Can you tell a little bit more about the mindset and the conventional wisdom around submetering and energy plant monitoring? 

Sean: We're different than a commercial operation. We are less about traditional utility consumption tracking and being able to isolate electricity, steam, and chilled water to an individual school or asset. We want to be able to look at it from a different angle and understand more on the building performance side, which is pushing the envelope in a lot of ways. We really want to understand asset performance against design. And the next step forward was understanding. And once we understand the performance, how do we change building behavior or equipment response and have the ability to measure and verify that against our future targets and goals? And the future targets and goals are right around the corner with Local Law 97 and our internal goals with CU Plan 2030.  

Mike: We want to make sure that the data gets to utiliVisor, and the data comes back to us without endangering the university's network system and security. So that, you know, it's a fine line of wanting data transfer to be as fast as possible, but at the same time as safe as possible. You need to make sure protocols are set correctly and that you have agreement with IT, internal networks, the BMS system, the SCADA system, and your (utiliVisor's) network system. Everybody has a different way of doing security and a different way of processing. It wasn’t really an issue of what didn't work; it's what works for each individual system and communicating the correct way. 

What was it like to get IT on board? Was that difficult? How did you work with them on that? 

Ed: We actually had to change our methodology and data transfer from what we usually use for situations such as Manhattanville, with multiple BMS systems, and we worked with the IT department to solve it.  

Tim: We adjusted quickly, because we're used to these kinds of situations where we're dealing with a high security system.  

Are there any tools or components we should highlight as essential to the solution? 

Mike: I don’t know if you want to discuss the Niagara system. I leave it to the technical person, Ed, if you want to discuss that.

Ed: At Manhattanville, they actually have a pretty unique situation where they have a dedicated person for their controls. With all those separate systems that they have in place, actually, there's one point person who is in charge of those various systems and maintains them for the site. There is an additional layer that that controls person uses on-site for a Niagara- or Tritium-based system, which is basically an open protocol system. With this open protocol, we were able to quickly add on to the system that's installed, and they were able to get us a dedicated device that we're able to load our software onto that can connect to their open protocols system to gather the information. 

Sean: From an end user perspective, there's a lot of interesting tools right now. Mike appreciates interval data in a graphical interface on the utiliVisor dashboard. The energy group loves interval data in in, you know, Excel format, or something like that, so we can manipulate the data. For me personally, getting the ability to develop KPIs and metrics for reporting is the tool I use the most. So I think everybody's using different aspects of not only the system provided but also the data output, either from visualization or from just raw data.  

One of the interesting things that Tim and I have been talking about is how to integrate this data into another tool to allow even more visualization by other groups outside of just the dashboard. And I appreciate the fact that UtiliVisor's going through the effort to develop tools and visualizations to help us not only manipulate data but report on data in a smart fashion. So again, it's really dependent on the end user, but there's lots of different ways now to take the output of these various meters across the campus and really try to get the full picture. 

Tim: And kind of piggybacking on what you're saying – we're taking the meter data, we're scrubbing the data, and if there's a system that the university is going to use, like an ESG platform or something that they may be using in the future, we're happy to take that information that's scrubbed already and then get an API to them so they can use the information that we're harvesting from our system right now. These are all options that you have on the table. 

Sean: There are many future uses, too. Just like Tim alluded to, there's the potential for development of a building system from this, and there's a future potential for carbon accounting of this data. There's, there's lots of different ways to slice this. And I think we kind of brainstorm together as a group, and come up with various ideas, and nothing's been off the table so far. We're using the real-time and historical data right now. The future is pretty much open. 

Sean, what KPIs you are using right now? And which are the most interesting to your stakeholders? 

Sean: My stakeholders right now are interested in rolling EUI (energy use intensity) on the buildings. We have a very energy intensive building on the campus, the research building, and that is the building that primarily could drive early Local Law 97 penalties. I use the EUI data to report upward, and then the actual interval data is going to allow us to evaluate projects moving forward specifically for that asset. That will really be our tool for monitoring performance and seeing if we hit our goals with the building. We've been able to break down each of the buildings on their apportioned share of carbon emissions in real time. And we can see what that means for the future and use some regression analysis to understand where we're going to be if we implement various ECM- or RCM-type activities in the building. It’s incredibly valuable as a tool to allow us to really stay on top of the various things we're going to do. And instead of guessing, now we can see. 

Alex and Mike, how you are using the utiliVisor data and what metrics matter to you? 

Mike: One of the things that I'm looking at is whether the plant is being used in the best economical way. There are different ways of running the facility and seeing how much it costs per ton to produce chilled water. And how much does it cost per pound of steam and seeing if we're doing it the most efficient way.  

Alex: As Sean said, I'm mostly playing around with the 15-minute interval data, looking at any sort of consistent trends, trying to develop models for how buildings are using energy on average days, comparing to see if there are certain days during periods that we're looking at that are outside of the expected demand, figuring out what we can do to reduce that demand going forward, and then planning for projects so we can actually identify savings right now. There's a certain number of projects that have already been implemented at Columbia where the savings are sort of invisible. If you're looking at monthly data for multiple buildings, you're not going to be able to see a small or even moderate-size change in energy consumption, because there might have been an increase in occupancy during that period, or maybe there's construction going on on some of those floors. So a much more zoomed in view of energy consumption at a small timescale is incredibly helpful. 

Sean: And we are not alone, we being Columbia, in looking at this data and analytics. Ed had great observation with respect to our system performance on free cooling and made a really actionable recommendation to setpoints associated with our free cooling system that's going to save us probably about $20,000 a year in utility costs, and we plan on taking those actions in recognition. We're watching, we're manipulating the data, we're analyzing the data, but we have partners in the background who are also looking at it as well. 

Are there any other results we should be talking about? 

Mike: This data is to help Columbia University make determinations on future projects at Manhattanville. So this information is going to set up for the development group for what equipment is going to be needed in the future for carbon neutrality. That's the big thing. At the end of the day, it's this data that's going to help us get to the goals of the university. 

Are there any other ways that you are measuring success that we haven't covered? 

Alex: I'm going to rehash something that was already said. But we want to bring this data back to every different group we can find. And again, within Facilities and Operations, that's every group, whether or not they're aware that this data could help them going forward. We’re using this data and creating dashboards and metrics to bring back to anyone who could use this – faculty too – to say, “This can be valuable for you, too. This can save you time and save Columbia money.”  

Sean: When you set goals, you have to have the ability to measure success, right? Columbia has set very overarching, aggressive goals for the entire portfolio. And now we have the ability to measure our success and our incremental path to reaching those goals. 

Now, we're not really looking for flattery, but we don't mind either. What would you say is the most valuable thing that utiliVisor has brought to the table? 

Mike: For me, it's cooperation and the ability of the team to get us over a difficult position: it wasn't all on us to get there. There was this team goal to get there, have suggestions and ideas and walk around the facility to say, “Hey, these are opportunities that we can improve and move on. It wasn't “This is what we're doing and that's it.” It was more about cooperative action.  


Alex: Yeah, that comes to mind right off the bat as well. Columbia's huge, and there are many things to focus on. Obviously, we know we need this and we're really excited to do it, but also having all of them physically there has been great to remind us to stay on track. And then also making changes as our needs change.  

Sean: For me, it stops at the relationship. This project has been a lot of back and forth, a lot of recommendations, and continuously bringing to the table new ways of doing business and new ways of looking at things. Oftentimes, when we close out a scope of work, a partner leaves or just collects their quarterly payment. I don't feel like utiliVisor is ever going to leave us in our journey here. And that's been the most important part – relationship. 

For your fellow engineers and operational folk, are there any counterintuitive experiences that you've had with this whole process or any mindset barriers that you'd like to make some recommendations about? 

Alex: The first thing that comes to my mind, just because you said "counterintuitive," is the sheer number of people who aren't interested in getting more data. I think, coming from my background, and the same with Mike and Sean, getting this data, you're immediately excited about it. And you're like I can do so much with this, there are so many different ways to make it valuable. I think I was surprised at how much we needed to sell this internally or needed to continue to show that this can help so many different people on so many different processes. You just need to actually look through it and figure out what are your goals? What are your KPIs? And then you can start making actionable changes to reach those goals.  

Mike: The platform is not an “I got you!” tool. It's more of a tool to say, “Hey, we're doing great” or We can improve,” and showing where the improvements have to come from. So you're not going to be initially pleased with all the data that you get because before you haven't been collecting it. And now you have collected it. And you see that areas you thought were efficient were not. And maybe you don't have to have the mindset of change. Change is okay. Some people might think it is an “I got you!” tool, but it's not. It's a tool to improve and get to Colombia's goals.  

Sean: I would say to any organization that has trepidation over submetering and trying to get as granular as possible – every major real estate owner or organization has ESG goals, whether they're in New York City or not – this is a very minor investment to be able to see that the actions you're taking to reach your goals are successful. If you have any trepidation about installing a system, about collecting data at a granular level in real time, it needs to be wiped away, because we're all trying to achieve very difficult things with very non-traditional systems. And the only way we would see that is to dive very deep. 


What are the next steps?  

Tim: I think we're going to continue to grow this and make this as useful as possible. We know the end goal, I think. But it's a winding path to get there. We're hoping to assist on numerous aspects of this on the submetering side, exposing that information to different parties. I liked what Sean said about it. We got all this data and we got excited. Not many other people got excited. I feel you, because I talk about submetering all the time and nobody gets excited. So I know what you're talking about. We're continuously going back to our team saying, How do we make this digestible for people to realize exactly what's going on? How do we make it simple enough so we don't bore them to death? And how do we make it actionable? The fact that we know these meters in the field are correct and we can rely on them is a key to the success.  

Sean: For me, I envision the utiliVisor tool as being something that can track carbon emissions in real time, at a very granular level, and also understand our utility data that's incoming, as we have marked changes moving forward. Time of dispatch is going to be an interesting concept moving forward, dispatching for carbon emissions. And, you know, trying to develop those types of programs and be very forward thinking is going to be where we're going over the next several months to several years. But where we stand today versus August of last year is phenomenal. And it gives us a good springboard for the next steps. 

Sometime in the next year, we're looking to probably break ground on the next 500,000-square-foot building on the Columbia Manhattanville campus, with another 300,000-square-foot building to follow suit shortly after that. There are future plans to expand that campus well beyond those two additional buildings on property that we own in the area. So there is a lot of potential growth. Do you know the beauty of having a system in place right now? We can do it – right – in the new buildings from Day 1, which is exciting. 


Columbia University
Sean Scollins, assistant vice president of energy & engineering
Alex Dulega, energy engineer
Mike Haeser, director of the district energy system for the Manhattanville campus

Tim Angerame, chief operations officer
Ed Harvey, lead engineer

Key Takeaways

Before Challenges 

  • No visibility into real-time energy usage 

  • State of the art system was not integrated/networked 

  • Multiple BMS systems + SCADA made networking challenging 


  • utiliVisor Operations Center 

  • utiliVisor Data Analysis Platform 

Project Goals 

  • Connect entire system under one observational platform 

  • Show stakeholders how the system serves everyone– from the board to the classroom

  • Be able to evaluate past performance to inform future design decisions related to campus expansion

  • Have data in a usable format to evaluate future performance for regulatory compliance and sustainability goals 

Key Requirements 

  • Host the consumption of each building’s usage of CHW, steam, and electricity 

  • Identify areas of opportunity for improved operations 

Why utiliVisor 

  • A family-owned, service-first company

  • Depth of experience; flexibility to work with on-site systems rather than force ugly integrations 

  • Renowned reputation for service & collaboration 

  • Manufacturer-agnostic & comfortable with all brands of equipment 

Results to Date 

  • Full visibility into and data collection of CHW, steam, and electricity consumption 

  • Identification of $20,000 in savings via new cooling strategies within the project’s first few months 

  • Easy visualization of rolling EUI throughout the entire campus 

Ready to discuss how we can help your organization?

Contact one of our energy experts to discuss your needs

About utiliVisor

Your tenant submetering and energy plant optimization services are an essential part of your operation. You deserve personalized energy insights from a team that knows buildings from the inside out, applies IoT technology and is energized by providing you with accurate data and energy optimization insights. When you need experience, expertise, and service, you need utiliVisor on your side, delivering consistent energy and cost-saving strategies to you. What more can our 40 years of experience and historical data do for you? Call utiliVisor at 212-260-4800 or visit