 Hello good afternoon. This session is focused on using the data that you get from many different sources to drive energy efficiency. One of the things we see is how do you make decisions in terms of where to invest your money? It is always resources are going to be very very limited and we have a very good panel here of speakers. We have George Deniz from Kuchman Wakefield. He's the managing director there and he's supporting Adobe campus and Adobe facilities worldwide. We have Victor. Victor is from director of global operations on facilities in Procade and he's been looking into not only facilities but data centers and have used extensively the data to drive efficiency in his buildings. Ralf Rene is from NetApp. He's the director of site operations and he's been in the business for a very very long time and has some very good case studies and very good data. He'll show you how they have used the information to drive changes in their operations and also to do energy enhancements on the buildings to drive efficiencies. My name is Mukesh Kutter. I'm the energy director at Oracle and we also have a portfolio of very large number of buildings globally almost to 20 million square foot and we use this information to identify which are our hogs, which are our good ones and try to see what we can do. So what we want to address today is, I hope this works, it does. Okay so the question here is when you are looking for data, so what are the type of data that are easily available? You probably have a data that you get from your electricity bill. You get once a month. Some places I know in Europe they send you a data, you know, bill once in six months to small accounts. But you also have data nowadays available to you for every 15 minutes, every day of the month, every year. So that's amount of almost 36,000 data points and that's very invaluable data. It used to be that only large buildings could get that data but now every residential customer also has that piece of information. So you use the information to gleam through and say we have the data but what do you make sense of it? Is this information you're looking for or the data? So we look into how we interpret and analyze the data. They are going to be some discussions here after the presentations. I have a couple of slides that we are going to talk about and see how we identified what could be going wrong or what could be figured out. Are there any references and benchmark? If you are the only people doing the work, are you in the middle, are you in the high end, low end, how do you find that out? There are a couple of sources available for you. EPA Energy Star is one of the big ones which we can see where you stand in terms of the benchmark for the rest of the industry. This is a good one. Interval meter data that has been there for some time and smart meter data. How do you interpret that? And take some simple actions. And we have a couple of case studies that we can show you that you can do that. And then again, once you take an action you want to see that really did save energy or not. So with that I am going to pass on the button to George who is going to talk about his experiences and his work he has done at Adobe. Great. Thank you. He is going to take a second to change the slide deck. So I am George Denise. I am the Global Account Manager for Cushman and Wakefield at Adobe. I have actually been at Adobe for about 12 years. I oversee their global portfolio about 49 sites around the world, 3.2 million square feet. Nine of it is actually buildings, the rest of it. The other 40 sites are lease space and somebody else's typically multi-tenant building. So this is a little quick blurb about Adobe but I think everybody knows who Adobe is. So we will go to the next slide. A little quick blurb about Cushman and Wakefield. These are my sponsors Adobe and Cushman and Wakefield. So I have to put their slide up there but everybody knows pretty much who Cushman and Wakefield is anyway. If nothing else you know. They have those real estate signs on all those buildings. This is the Adobe headquarters building from a slightly different perspective. Again, nine buildings, about 49 sites around the world. 3.2 million square feet. Two million of that is the nine buildings and the other million is the 41 sites. So I'm mostly going to talk about what we're doing with the buildings. We do a lot with the lease space too but we don't have necessarily we submit whenever we can. We do work with it but we don't have as much control obviously in somebody else's building as we do with the own buildings. Key factors I guess we have 23 lead certifications, 17 at the Platinum level. All of the owned buildings and the one building that they lease entirely and manage are all certified through lease EB or e-bomb at the Platinum level. Over the last 10 to 12 years for the entire portfolio we've undertaken about 157 different projects. Most of those energy but energy and sustainability projects spent a little under $5 million have received back about a million dollars in incentives or rebates and have reduced total operating costs by about annual operating costs by about $3.2 million which is about a 1.2 year simple payback and about a 84% return on investment. Probably key here for this group is that we reduce our electricity use by about 50%. Also we have more recently installed fuel cells. We have on the San Jose campus about a dozen Bloom energy fuel cells. First generation they supply about 32% of our total electricity use there. In San Francisco we have two of the second generation. San Jose is about a million square feet, 12 first generation supplying about a third of the energy. San Francisco is about a third the size, about 325,000 square feet. Two of the second generation Bloom fuel cells supply about 50% of our electricity use there. Similar uses both the San Jose has four data centers and 68. I'll think of in the second labs. Development labs, testing labs. San Francisco has one data center and a smaller number I think eight to 10 testing labs. So really similar uses. This is just kind of a quick summary of what we're doing with the buildings. Lee Platinum, good energy star scores, good solid waste diversion numbers. San Jose site, similar numbers again to what you just saw. These are the kinds of things we're discovering by studying the data and some of the projects we've put in place. Our very first, well, basic monitoring I think on a building we've been doing ever since we've had buildings and that is to do our daily rounds and read the meters manually and you know that's that's what we did in the Stone Ages. We chipped that numbers into a stone tablet and preserved it and studied it. More recently and probably a very good tool, basic tool is energy star, Mikesh mentioned. Energy star is probably the most comprehensive out there, easiest one to work with. Looks at a number of factors including the climate zone you're in, the size of your building, the number of people in it, the number of PCs in it, other uses such as size of your cafeteria, your data centers, et cetera, et cetera, and then tries to compare you in terms of your energy efficiency with all other buildings in the country. Probably the best part besides all of that is that it's free to use. So it's a useful tool, it's a simple tool, it's a basic tool. Beyond that I think, you know, we looked at going back to 2002-2003, PG&E had digital almost real-time meters but day late reporting. That wasn't good enough when we were doing a lot of programs. So the first thing we did was submit it and put in an ION software system so we could track it. That whole piece for the three buildings in San Jose cost us $39,000 and was very useful immediately as we suddenly had a graph of what our usage looked like. We could see startup spikes and pillows and other spikes for other reasons and these are some of the kinds of things we discovered with it. I'm looking at the right slide. But, you know, one of the things, for example, is we immediately saw a spike in the graph. It turned out to be an issue in the programming for the chiller plant that our engineers had always suspected was an issue. We had had engineering firms out to look at it a couple of times. They said, no, everything's pretty much the way it's supposed to be. But as soon as we had a graph of it, a picture of it, we could immediately see that there really was a problem. We brought them back out. It was a very quick fix and it saved $36,000 a year. That problem had been in place for, that was 2003, I believe. So that was the first building that was built in 1996. So that's $36,000 a year times that number of years. A lot of other things like that. So tremendous value just in being able to visualize, see a picture of what's going on. Fast forward a few years and we got more and more into it. And we've now spent almost a million dollars on our monitoring system. However, we've also found with all of the things we've discovered and repaired with this monitoring system. We call it IBIS, which stands for Integrated Building Interface System or Intelligent Building Interface System. We developed it with a company in San Ramon called Integrated Building Solutions. And it just gives us a lot of powerful information to work with. And the payback, as I say, with all of the things we've discovered and been able to address, the payback ends up being about three and a half years, which is a little long for some building owners, but for those of us who are in it for the long haul, I think most of us agree on most investments. The three and a half year return isn't bad. So if you look at this graph up here, these are a number of screenshots of different aspects of the of the graph of the system. This one is showing the energy usage in real time for the three buildings. And then this is a consolidation. The what happens is each night at midnight, our system, we've automated it essentially. So we're automated commissioning in a sense. At midnight, it looks at the weather local weather station, looks to see what the weather is going to be the next day. It then goes back over the past 12 months and looks by the five most similar days. And it creates that little green band. Those represent the the actual energy use within those five days. It then tracks forward if you see the darker, that's where that's the time right there that the shot was taken. And it tracks it. If this is running within that green band or below the green band, then it's doing what it's supposed to do. If it goes above the green band, I think right in here, you see a red spike right over here, you see a red spike. It turns red when it goes above that band, depending on the size and length of that spike, or whatever that overages that variance is, it goes into alarm. And depending on the parameters, it automatically writes a work order through our computerized maintenance management system to the engineers. One of the things we used to say is you can't manage what you don't measure. So then we developed all of these systems to give us the data we needed. Then we said we got all this data, you have to do something with it because we discovered building operating engineers aren't really analytical for the most part, some are. But generally speaking, they focus on work orders. So we decided, okay, let's have this system write in work orders. And now they're they're tracked and followed in, and they have to close those work orders out and have a satisfactory solution with what happened. Property in Seattle, this is about 160,000 square feet and is wholly leased by Adobe. Same kind of numbers. Electricity reduced 19% much smaller numbers because of the size. And while 19% of electricity may not sound that much better, when you consider that the cost of electricity in Seattle is less than half of what it is here, then that starts to make more sense. Same kinds of things. Again, some of the projects we did in their returns. I'm not going to spend a lot of time on this. But if you read quickly, you can see there's a bunch there. This is my favorite graph in the whole wide world. While we were doing all this work in San Jose, we added a third tower, we added two more data centers, we added 40 more labs. And we increased the population by 45%. So to look at an absolute energy use graph isn't that impressive doesn't tell that good a story. When you look at it on a per square foot, or a head count basis, it does. But the Seattle site over the past five years has had about the same population. We've probably built up the server room a little bit. We've added 40 people there. But generally, it's been pretty stable. And so you can see the on the graph, the first columns, the this is a 12 months of the year, looking at a five year period of time, the first column, the highest the light blue is the first year and each year successively it's come down. This is probably for somebody working with energy. For me, at least it's my favorite picture next to pictures of my kids, I think are two puppies I recently acquired. Latest is going in and tearing out of the walls in our space, opening it up, putting in motion act creating neighborhoods, putting in motion activated HVAC and motion activated overhead lighting, plug load based on since motion sensors. And in a building that had already had an energy star score of 100. We were able to go from 80 people per floor to 135 people per floor and still reduce electricity used by 65%. So really good impact. Quick pictures. This is looking at it from a three dimensional model. Some of the material entered the information pages we have. I'm going trying to go fast here because we've got other folks coming up and that's that's a quick overview of it. I wanted the others to have a chance and then if you have questions later. Okay. Thank you, George. Really appreciate it. Next, we have a presentation from brocade. Victor. Thank you, Chris. It's coming on. Okay. So jumping right into it. I'll talk, you know, give you guys a brief overview of a brocade. I think unlike Adobe, most people don't know what brocade does. Talk about three specific examples where we looked at interval data. So I want to share our findings and our results. And then also besides energy efficiency, you know, one of the things that we always contemplate as building operators is, is that fine line between energy efficiency and reliability, because sometimes you can cross that line. So a little bit of a brocade started with four employees in 1995 for now 4500 employees strong operating in over 160 countries worldwide. We actually have a combination of virtual offices, least offices, multi tenant offices in known space, similar to to Adobe's portfolio. And really what we do is storage and networks, we send percent of the revenue in the market, we own that piece. And the way I like to describe brocade is, you know, we own 70% of that sand business where Cisco own 70% of the IP business. So we want to take over their space, they want to take over our space, just to give you a sense of again, who our competition is. So jumping into the first example, again, this is regarding interval data. And this is actually the actual readings of a lab. It's only about any racks. It's our sales and marketing lab in San Jose here. You can see that they're operating anywhere from 250 kw to 290 kw. And if I were to put it in terms of dollars, on a yearly basis, they actually spend around $300,000. So when you look at this graph, you can see that, you know, by month, we have it where we're training sort of the actual consumption. And then it's hard to see, but I'll try to point that out. You can see these flat lines here. That's if you average the monthly data. So when we did that, what we realized is that it's basically it almost looks like a step function. So then we started asking, Well, what's happening here? Well, who best to ask but the lab managers themselves. So what we learned is that they were actually turning off equipment when it was not in use. What a novel idea, right? I mean, something we all do at home, hopefully. So then what we thought of was, Well, let's take what they're doing and implement it across the company, right, where we have over 3000 racks, where my energy bill is close to $10 million. So just from this example, and by literally working with the lab users to change behavior, we've been working with them to leverage scheduling software that turns off equipment automatically when it's not in use. And you can see here in the second quarter of a tooth of May, April, you can see that they were growing pretty consistently. But when we implemented the challenge when they were turning off equipment when it was not in use, you can see that effectively, we cut their power consumption, you can argue by half or by 75%. Again, with no basically investment at all just really by changing their behavior. So that was one thing we learned. Another thing that, you know, we want to talk about is, again, metrics. So in data center, buildings or operations, a common metric used is the power utilization effective metric. It's really a ratio of the total building power used divided by your it power. So besides the it power, the second consumer of electricity is your chillers in your buildings. So here what we have is a scatterplot. You can see all the blue dots are actually millions of points prior to July of 2012. And then we put a polynomial on that scatterplot. And what we realize is here we're plotting basically our PUE. And you can see that we operate between 1.1 and 1.3, which is really best in class. And you could argue that basically 60% of the time we're operating, you know, below 1.2. So what we decided to do in this experience, we said, well, again, without investing a whole lot of money, what happens if we change the set point temperature for the chill water system? What we started seeing is this red trend line that we were actually by just simply raising the chill water set point and by leveraging our water side economizer, we were, you know, getting a better PUE effectively throughout the year. By benchmarking it and using it previous historical data, we could actually also see that at 69 degrees wet bowl, you know, that equation changed. So by again, looking at this, I call it interval data and somebody might argue and say, well, wow, this is over a year and a half of data. We were actually realized some savings, again, just by studying the data and making changes and also realizing that at some point there is a tipping point. So making changes there as well. So this is another good example. Again, a lot going on here, but basically what I want to call your attention to is these spikes sort of in the middle of the graph. And what effectively is happening there is a chiller was turning on and off every 15, 30 minutes. And again, going back to the principle that chillers consume the most amount of energy besides your IT load, we found that this chiller had been cycling on and off for about three to six months. So imagine like any pump or motor, when you start that piece of equipment up, it ramps up so high. It's really the exponential curve on a power curve. So again, by just identifying this, by looking at really, if you look at the x-axis here, you can see that we had a drill down to basically the hour or the minute, if you will, to find this anomaly where we had to basically again only change the sequence of operations to fix this issue to save a lot of money. So that was another example that we found again, looking at the interval data. Thank you. Thank you, Victor. You can see how this data has been used very effectively by Victor in trying to find opportunities for saving. Raft is going to talk about his experiences with NetApp, and he has equally good experiences in this area. Thank you. Thank you, Mikesh. And when Mikesh earlier introduced me, he said I've been around a very, very long time. So that must mean I'm old. But I met an old friend of mine in the back here, Kim Ryle, who was my PG&E rep like 20 something years ago. So I guess it's pretty accurate. Okay, a little bit about NetApp. Most people don't know who we are. We're not necessarily a household name, but essentially we're the enterprise storage provider to many of the things you use in your daily lives, like your Facebook page is stored on NetApp storage. Your Yahoo email is actually stored on NetApp storage. So we're behind the scenes, but we're everywhere. Well recognized for our great place to work initiative. This one is ranked number three in the world's best multinational workplace back in 2012. Similar to Bokeh, we got a presence in 46 countries with about 150 offices and over 12,000 employees worldwide. We're a $6.3 billion enterprises of our last fiscal year. At the Sunnyvale headquarters, we do have a total of about 1.7 million square feet. 1.3 million of that is actually energy star buildings. Most of our portfolio that's later than 2,000 are all energy star buildings. In addition to that, we began our lead initiative a couple years ago, but we've got about 600,000 square feet of lead buildings. Two of them are actually certified. The one building that doesn't have the energy star label sort of on the far right of this map is a brand new building that we haven't yet occupied, but we are with the application with USGBC for lead gold under new construction. On the site, we've got three fairly large data centers totaling about, well, of that 10.5 megawatts of peak demand. 7 megawatts of it was represented in data center operations. It's just kind of a simple diagram of what goes into our energy management programs, but in terms of instrumentation or data acquisition, we do it through a variety of platforms predominantly our EMS platform, which is simply an energy management system. We do a parallel metering, a socket parallel metering, and we do a shadow building off the utility at our main switchboards, and then we do some sub metering downstream of the main. We actually use a square D ion currently, but it was a power measurements limited prior to the square D acquisition of that technology. Our building automation system is automated logic controls, and we use their energy reports function, which is quite comprehensive, and actually when you integrate EMS with BAS, you could really get a single dashboard similar to what George was doing with the Zibus. In addition to that, we did implement Accenture as a Accenture Smart Building Solutions platform, which is really a combination of automated fault detection and diagnostics along with regression analysis. And then in addition to that, we also use a, which is really in the DCIM space, but it's a product called MODIS that we use actually more for energy management than data center infrastructure management. So those are the inputs into the energy management programs. Some of the outcomes or measures we implement are continuous commissioning. We can do that through predominantly our building automation system, along with the interval meter data that we do and our parallel metering. In addition, we've implemented retro commissioning in about seven of the nine energy star buildings. Just every year, we'll budget for energy efficiency measures. In addition, we'll implement some capital projects to improve energy efficiency. One of the things I think the idea of the session was really to look at interval data. So what we're getting out of our ion system is really a daily report on energy consumption. So you could see the two sort of large spikes on this. Those are the buildings that house relatively large data centers. But this is really just doing a day over day, compared to the same day last week. The value of this is really if those peaks spike tremendously, at least we'll alert you to go look into and investigate. So everyone on my staff, including all of our technicians, gets these reports. They find it a little bit obnoxious because you get this every day. But once you get through kind of scanning it, you're, you know, we get it in numerical. And in your mind, you read it sort of graphically. Yeah. Alright, I guess you'll have to get me restarted. I'm going to. But the point of it is really looking at your energy consumption on a daily basis allows you to actively manage it. Oh, thank you. What's happening here? Yeah, all right, okay. Okay, so just a couple of case studies that I would be maybe interesting. One of our buildings pretty well performing. There's a variety of benchmarks. Certainly energy star is one of them. There's a database out there called CBEX and you could kind of characterize your energy performance against CBEX. This building does quite well on an annualized basis about 13 kilowatt hours per square foot per year. It's a lead gold under e-bomb energy star as I mentioned, but we do do chill water plants. So the picture on your right can kind of show you that we've got a chill water plan on there and just some simple energy efficiency measures like daylighting and CO2 monitoring, which is in many cases counter productive to energy efficiency since you tend to flood your building with fresh air when you don't need to. But one of the things you can get is some real good data out of a well instrumented building automation system. This was just looking at spikes in our supply air temperatures in a couple of air handlers that serve this building, but it leads you to implement some efficiencies or optimization as a result of really what is programming errors. And one of the difficulties when you're a controls technician is you program things, but you don't necessarily have the ability to really track performance and do some direct correlations. So you do have to study your data. But what we were able to do is smooth out some of the spiking and try to match the air handler operation so that one isn't essentially overproducing over another. The air handlers serve the exact same square footage. Another thing we also found out and the picture on your left is really you'll see a supply air temperature rising while the air handlers not operating. What this led to us is a discovery that we were actually turning on our reheat coils, at least flown our boiler loop pumps flown hot water without moving air. Terribly energy inefficient since you're not utilizing that hot water. But it was really just a simple thing that we were starting our pumps and turning on the boiler based on temperature calls. But we had the air handler locked out on an occupancy schedule. So what this was really resulting is very excessive supply air temperature set point that's just migrating up through the VVs all the way up to the center in the air handler on a five story building. But that was easily identified and we actually seen this first in the interval that at the building level where we actually seen the boiler and the boiler pumps come on at a particular time when no other energy is really being consumed. So we on the bottom left is just implemented and match the the reheat call with with the air handler. So some of these things that we identified were just a few items that we were able to do to incrementally improve energy efficient. We were actually in a very efficient building. So these are just incremental steps that we did to become a little more efficient. But low hanging fruit in the sense that the data drove you to these optimizations and it didn't require any capital investment. Another building. This one has a large lab in it. So it's got fairly large energy bill and a little bit more of a demand two megawatt lab in there. One of the things we do have though is an incredible amount of instrumentation and all of this data needs to be processed in order to optimize. But just kind of even in a relatively small building one hundred twenty two thousand square feet we've got you know thirty three thousand BMS points. It's a lot of data to process some of it you don't necessarily review. But it's you know as an example rooftop units will have 12 points on it. Each one of the cry units will have 10 points on it. The VFD's there's a lot of information you can pull down off a VFD. But even your VAV's in our case zone dampers there's some points we pull off of it. So there's a lot of information that you could obtain and analyze off your building automation system. Some of the things that we had discovered was an ability to optimize further and without going through the entire list it was just really looking at various set points in terms of supply air comparing that with the return air and trying to fine tune the building if you will. We kind of came up with these all five of these measures really didn't result into any significant energy savings. But it also did validate that the building was operating relatively efficient and again incremental efficiency. This information however was gathered from the Accenture Smart Building System which was a capital investment in the form of software which really then came into a question as to the cost of implementation versus the return on investment. You know without disclosing the cost you could say that this had a two year ROI and these are annualized savings so that I'll give you an idea how much we spent to implement the program. What it also does is kind of prove as a testing ground do you kind of implement this and scale it out across your entire campus. That becomes a little bit cost prohibitive. You know George mentioned approaching a million dollars for his IBIS implementation. It is something we're seriously investigating but we're also very pragmatic in making sure that we've got a business case to justify that investment. Some things I just wanted to add this event actually took place Wednesday night and because I still have an operations role or not willing to give it up. These were alarms that came through about 1 30 in the morning and basically this is the interval data and as you can see there's a big dip there laser so you can see that energy dropped down significantly like 170 KW but then peak right back up to 720. Just looking at interval data you would then want to investigate that. So if you just had a PG&E meter you would still capture this. This is the PG&E meter. We parallel so it's PG&E's PTs and CTs that we're reading but this warrants investigation. Well what happened here? Why is this happening? So we investigate killed it. Yeah killed it by the remote. If you don't mind flipping. Yeah it's not even going forward with this one so I had to go back and whose PC is that? It's not even his. It's not the operator. I can assure you the clicker operator could go down once. Thank you. All right so apologize. These are actually screenshots out of our ALC and you can barely see the resolution on that but basically the what we're we're seeing here. OK so what you're seeing here is an economizer outside air damper opening up the chill water plant shutting down. So this has got an inverse relationship. This was at a specific moment in time that caused that dip was the chill water plant shut off and we went into outside air economizer. But what it also demonstrated is that the fan energy exceeded the energy of the chill water plant and that's why the spike back up was greater than the demand when we were running in mechanical mode. So it leads you to go investigate and figure out how to optimize this. And again this was derived off of interval data. Another just so you got to go one step further and go OK what are the environmental conditions that cause you to one go out of the economizer mode go into and the time stamp is somewhere I guess right about here. So we certainly had a temperature decrease which really opened up the economizer damper. But in addition very interesting microclimate going on here where we've got a essentially humidity spike in the rarely early morning hours. So then it hit this threshold which then brought the chill water plant back on because we have to go in and dehumidify. But in any event the point of this was really just to show that you can derive a lot of information just off of your interval data that will lead you to that sort of secondary investigation. All right. So that's it for me and Mikesh you have some slides. It's so interesting. With all this data you have to make good use of it. And that means requires more digging and digging of information. And one thing good is the more you dig into it the more easily you find what is a fortune what are the opportunities there. So I got some other slides to show you very briefly. Some of them are basically discussing one thing is good that I don't have to explain what Oracle is. Who doesn't know about Oracle. So that makes it very easy so you can skip on these slides very quickly and move on to the next one. But we do have an assortment of buildings. You know you talked about building management system. You think of any building management system in the market today we have it. We have over 20 billion square foot of real estate. Nearly half of it is owned. Other half is leased. Some of them we have a triple net lease we are fully responsible there are others that we have no responsibility we just lease it out and whatever information we get we get. It's not easy to get in gather information from all of the sites into a single place and be cost effective. So we have been utilizing the built-in capabilities of each of the building management system and try to use this information. One of the things which we do is a simple thing. Look at the yearly energy consumption data for all of our buildings which are in our own portfolio and you will see that they are not all the same. They're all varying up and down. So we try to figure it out. The ones which are using less are there any errors and we did find errors there because they missed some data there and the ones which are very high. And there are also other errors or not. And if there are trying to figure it out what can we go and do it. Learn the lessons from the lower one the one you use less and apply it to the one which are high. You will see a line there which is around 15.8 kilowatt hours per square foot per year. So I looked at the data from the energy star for all locations we have in the US and found out that about 15.8 kilowatt per square foot per year if you met that you would get into the energy star qualification. The what they call energy star building. And it was very interesting in some cases if they are very cold then your heating bills are high. The ones which are in the hot climate the cooling bills are high the heating is low but when you add them together the number was coming up pretty close to be about 15.8. So that's our benchmark kind of a thing we look at it if any time we can get energy consumption less than that we are very happy. These are some of the ways we are trying to see figure it out what took us down. So where we can apply it. This is one of the buildings we have in Virginia area. Basically this is you see in the schematic here and I wanted to ask you guys what do you think is wrong with this one. Can anybody see anything wrong going on here. What's happening in this building. You see here. What time is the building coming on. One third in the morning. And this is winter climate cold climate up there. And the facilities guys said I need to bring this building on that early in the morning otherwise people will be unhappy and comfortable. This is a Monday and he was very kind that says OK because you're coming back after a weekend and Monday I need to come on early. But other days I can be delayed but still coming at 4 30 in the morning. So this is something we want to go and one of the things I found out is no matter what I talk to my facility engineers and say change it they would change it when I'm there you walk away they'll change it back to where they were. So it's important for them to bring them on board so that once they apply it they will stick with it. That's the biggest challenge we have. And so one of the things we did was over a period of time explain them and said OK let's try it one we can see what happens. And this is what we did. We did try it. And we went on all days at 6 30 in the morning instead of far about 1 30 and then 4 30. And you could see here there was significant saving to be had. Now who has ever heard of the rule of a third. One third. Yeah. So you at any given time if you look at your buildings you have the potential to save about one third of the energy and we have proven it. I'll show you some data. So if you look at it but we did notice and I'm going to compare these two slides on top and others so that you can see and the alignment instead of coming back so early in the morning we are coming back late. But also we noticed that our energy consumption on the first day increase in the morning to 1200 or more kilowatt. And that is obviously because the building was shut down over the weekend and you had had a much higher energy use there. We learned something else in this building. How many of you know that most of the buildings in the U.S. even today do cooling and once it becomes too cold they start heating it up. It's very common. I mean 90% of the building today have managed in that way. They have a central system a complete VAV box that feeds about eight or 10 about 1500 square foot area. And you're supplying the cooling from the center plant to that particular unit. It cools it and in certain area becomes a little more cold. You bring on the heaters. So this is one thing we found in here too. And what we did was we look at the VAV boxes the variable air volume boxes and there from there we figured it out that in the morning when you're trying to do cool down it brings down the temperature so quickly in about 15-20 minutes that you didn't have to have it for four hours or two hours or one hour ahead of time. So the shortest time and this was it was this type of data that was able to convince the facilities engineers to see look this is what's showing up. You only need this much of time. So give it a try. And once they get into it then they love it. The same thing happened in terms of a airflow demand when you bring the building up early in the morning because there's a heat built up. You require a lot of airflow in the beginning. Once everything is set us down and you don't need as much of the cooling and this was another information when they saw it. Hey, there's no occupancy happening between 6.30 to about 8 o'clock or 9 o'clock and then the airflow begins to pick up. You really don't need to start the building so early in the morning. So which we'll look at it. But this is a very interesting slide. This is basically to prove the point that what you need to do you do this. This particular system is one of the system we have in the UK. You do the cooling of the air then goes and splits into the situation and the two different ducts. One of them starts heating it up again. And this is the kind of system I mentioned. Also what happens is that there are dampers and the dampers pinch up the airflow because you don't need as much airflow. You're trying to create a very high pressure here and then try to block the airflow so that you don't deliver as much air as you need. So those are our static pressure set point and every time you go and talk to an engineer that I want to reduce a set point or the person in the corner is not going to be happy because he's not going to get the air. And when you show them the data, look, your dampers are trying to restrict it. If this damper was completely open, for example, then you don't have to put high pressure on this fan over here. So if I have to summarize three or four big items that can save me energy today in any building towards reaching my goal of a one third, these will be the four items. Look at your start time, look at your stop time. Not only for the building HVAC systems, HVAC systems, look at the lighting system, look at the outdoor lighting systems. All of them you should really look at it and that will give you what you are looking for. Supply your temperature. You don't need to supply very cold air all the time and then start reheating it up. Reset the temperature up and you don't have to supply very high pressure and try to put a damper to choke the airflow. That doesn't work out too. So those are the kind of things simply can be done. And I was at a very recent conference where one of the person remarked very eloquently, we have done a lot of things besides this. We have done a lot of things on the periphery that would really be very, very advanced and give you savings like starting up early in the morning or when you know this is a hot day today, try to cool the building up and then let it warm up slowly. Those all help. But this person when he mentioned it reminded me of the very simple thing. He said somebody asked Warren Buffett during the energy, you know, during the financial crisis. Remember 2010 when everything was built down? He made a simple observation. He said, there is so much money to be made in the core operations that you don't have to run on the periphery. Same way, there is so much money to be made from the core. Simple things to be done in the building automation control systems that you really do not have to do necessarily do on the periphery. Good to do it. Good to go look into all of those things. But don't ignore the simple things. And this is what we did. This is what we are rule of one third. It took us 10 years to achieve it. But we did. We exceeded that. Over the years, this is for a two million square foot campus we have in the Redwood Shores. We started in the 2000 when those were the baseline. And over the years, we have been able to exceed that goal. And it continues to go down. This is for the electricity. We have for the natural gas, it keeps on jumping up and down. It was down in 2009 and then suddenly decided to go up. We figured to figure it out what's happening here. But this is the data that is very powerful to us. And also to our management on the upper end. So they can understand what's going on and also to the people who are really operating the buildings so that they feel comfortable. Sorry. I think that was the end of it. But I have some other slides and other can show you on the similar case studies. There are plenty of them. But we want to open it for questions to any of the panelists and see what we can provide you some insights. Yes. Good. I was wondering if any of you have explored technologies that report to desegregate whole premise, whole building energy use and to buy end use. And whether that further level of desegregation of interval meter data by major classes of end uses will provide any additional any additional value in the upside. Yes, I'll start. So similar to George and what Ralph described, you looked at systems that integrate a lot of that data. And then we do a lot of monthly reporting, not only internally, but also to our customers where we're showing them again, the data to help us understand what's what's going on. And by the way, how can you help us change it? Right. Because they're really the ones ultimately leaving the lights on using their computers or are using that it equipment. So far, we haven't got it to pencil out. I mean, even with our with our lab project, where we're close to a million dollars, it's just really costly to implement. It's also a long time. And what we're finding is that we can do a lot of it without getting to that level of reporting. We can do it with just whatever data we have. We can produce a lot of savings. And I would say the same thing that, you know, we've done an awful lot of sub-metering of subsystems and so forth. But I really have to say that in spite of that, the bank of the buck has been just from looking at the overall graphs and capturing it. So to answer the question about this desegregation of data, I think it's one of the most powerful tools we can have. I'll give you. We have a lighting circuit in our building, which is feeding about 30 different lighting breakers. Now, if you talk to your people, they say, Oh, lights are not supposed to come out in the night time. They will stay off. But there's no way to verify that. If I had a means of desegregating data from all these 30 circuits from measuring only one data point, I can identify which particular circuits were left on or were not turned off when they were supposed to go on. So I've been I talked to some companies which have been working on it. They are more focused on the residential side. So they want to know when the refrigerator comes on, when your dishwasher comes on or dryer comes on. But they have not really focused on the commercial side. And I would be very happy if somebody comes up and says, here's the meter data. Let me take a look at it and I can desegregate and tell you which particular circuit was left on or turned off will be very helpful. Jack and Victor, Victor, you talked, I believe, about having a software application that monitored plug loads with motion sensors. And if I remember your comments correctly, could you talk a little bit more about that? How that operates? Sure. So I guess one correction. So it's not tied to we have motion sensors in our labs, but it's just for automated lighting, right? We have separate zones of cold out of the hot aisles. The software that I was referring to is there's two components. There's at a plug level now, down with equipment, even A and B bus power supplies, we can monitor how much power we're consuming. We can also lock it out, turn it on, turn it off. So we're now trying to understand power consumption at a utilization level, right? So what's the capacity? Really, how much are they using? And instead of buying more equipment, if there's 50% compute power there at lab, challenge them on buying more. So that's sort of a one side of the equation. The other side is we are looking at scheduling software. So the concept there is instead of, again, having 10,000 pieces of equipment, only 50% of it's being used, instead of buying more equipment, how do we create a software tool that reserves the equipment so that we leverage the existing capital instead of buying more equipment, building more lab space, consuming more energy? So that's kind of two separate software components that we're looking at. So to expand on that, one you mentioned about the isolase, the person who developed the isolase, who built and sold it, sitting in right here in our audience, can you raise your hand, Jim? So he's the one who built these systems and they were very helpful. What happens, it comes as your plug, which has got six points. Two of them can be run continuously, but four of them are based on emotion sensors. That is like your monitor or printer or something else that you don't need to be running all the time. If it doesn't detect emotion, it will go offline. So if you have a radio sitting in your office or if you have some fan or electric heater or something, those can all be cycled off very easily. It's a very, very powerful device. I bought a bunch of it for all my campus and it helped me a lot. Thank you. A lot of the comments and examples related to monitoring base commissioning and ongoing basis, as far as the capital projects go, internally, what would you say your hurdle rate would be? And if you have projects beyond your hurdle rate, would you be open to go to third-party financing of these projects? Want to address that? So we've implemented monitoring base commissioning and took advantage of the rebates. The hurdle rate in terms of IRR is north of 10%. The simple payback periods we generally need to target about two-year. The question about whether or not we'd go third-party finance, we're quite cash strong, but we've just recently announced the dividends so that position's probably changing, which also means we've got some debt that we're gonna do financing capital on. So I think the payback scenarios are gonna be challenged by our finance group, but essentially third-party financing or going into the debt market is essentially the same equation. It just becomes a question of how much premium we've gotta pay, but I think the deployment of capital, regardless of whose capital, is still gonna have to meet a strong business case. So depending on the project, I'd say two years, usually can get through quite well. Three years, you've gotta probably sell the management a little more. Anything north of three and above five is almost not considered. For the question in the back first, yeah. Two questions, should look pretty quick. First of all, I'm curious if the insights that you're gaining from this data suggest to you different from the amount you might expect to train or hire building operators? And if you think that's something that needs to, how long might propagate the industry? The other question is whether these insights have led to changes in how you think about longer-term investments, maybe downsizing children or other things that are gonna happen infrequently, but you're better prepared for them, perhaps. Particularly to your first question, yes, it's changed how we hire operating engineers, for example, significantly. As I mentioned, we found that generally speaking, they're not necessarily analytical. Very interesting story, I think. When we first certified their first building back in 2006, one of our engineers went off to tell his dad about it, who was a chief engineer for another building and thought, he learned a lot in the process. He was a skeptic on the front end, learned a lot in the process, became a huge enthusiast. What to tell his dad, he needed to do the same thing with this classic building in downtown San Francisco, and his father looked at him and said, why would I wanna do that? It's more work for me. If anything goes wrong, it's my fault, and if it's a success, then the manager takes all the credit. Probably represents the thinking of a large percentage of building operating engineers, but there's also a whole new wave now that are really much more analytical, much looking to the future, excited about being involved with it and so forth, and when I'm looking at resumes, I'm looking for a lot of discussion about energy management and sustainability and so forth, and I'm looking for the ones that are excited about the things we're doing and wanna be part of that. The other thing is I spend a lot of money on training them, besides sending them off to take the classes with the control system companies for whatever building it is, and they enhance classes as well. There's a lot of other classes I'm sending them to now, and I'm also giving them bonuses if they get their lead AP, and this is a whole variety of things because I just found there's a tremendous return in it when everybody's aligned and marching in the same direction. There was a question here, yeah, thank you. Seems like everyone on the panel has achieved very strong energy reductions. Is energy efficiency a done job for you? Is it finished, essentially, or if not, what's next? Well, number one, it never stops because the technology is constantly improving and the processes and the way we think about it and look at it is constantly changing. That's why I like that one graph of Seattle is because each year they just keep coming down and down as we find more things and implement more things. And the rule of one third remains in place. We looked at the building, which was doing very good, one of the lowest energy consumption, and we will still continue to drive it down. So there are, only thing is it becomes a little bit more involved, but there are opportunities are always there. But I do want to, so we are subscribed in our ISO 14,001 target objectives to reduce energy on a kilowatt hour per square foot basis. And we're hitting that area where we're finding few projects that we can implement, optimize, and now we've got to resort to capital investments to get incremental energy efficiency. So that's maybe that threshold that you say, okay, well, what are your goals? And we're looking at a continuous improvement of roughly 6% square foot per year, which is getting to be almost unachievable. So we're maybe resetting that goal to, okay, how do we beat energy star as events were? But I think there is that problem was, okay, when do you got to make capital investments for incremental energy efficiency that then still has to meet a business case? Which I think was going to answer part of your question in a older building, you can probably afford to upgrade chill water plants, for example. In newer buildings, it's much more difficult. In fact, you've got a tremendous amount of book value remaining on that chill water plant, irrespective of whether or not it's operating optimally and use cases changes. In fact, we have that exact problem in that building that has a pretty good energy star score. We actually sized that chill water plant to have flexibility to build a lab in that building. We've subsequently decided we weren't gonna put a lab in that building. Now we've got oversized chillers, oversized pumps, but we implemented a basically a water side economizer project that allowed us to optimize that deficiency of just use case changes. But we can't afford to change it because there's the almost, on a six year old building, the majority of that capital investment's still on the books. And the industry is changing very fast in certain areas like IT equipment. How many of you have the old CRT type of monitors anymore? And the laptops that go off, go into sleep mode all the time. So a lot of savings came from there. We are taking the credit. Actually, the credit is your technology. That's because you're not working. I'm just gonna add, so our campus are buildings. They're only three years old. In terms of the Title 24 calcs that came, we exceeded by 24%. Our original PUE was 1.3. So even in new buildings, with all that commissioning we did, and as energy efficient, we've been able to improve on that and continuously find projects or ideas or even the environment's changing, right? Or our expectations are changing, expanding the temperature ranges, whether it's in the data center or in the office space, especially with condoms, right? It continues to say, often we operate more efficiently, whether it's on a chemical basis or a headcount basis, I think that's also driving us to continue finding ideas. Then it is a question of, okay, is it low-hanging fruit? Is there lost in fall? And sometimes, some years it's more of those less low-hanging fruits. We find that, yes, it continues to tell us people can fix it in a few years, let me guess what, it's back to where it was, right? So if you wanna change something, and if you're not looking at that then I'll hold it back for you, so. So that's the whole idea behind continuous commissioning and not any continuous commissioning, is that you are constantly watching and monitoring everything and checking it there, because it does be great over time. And I also think about, you know, back in 2006, when we started, actually 2005, when we started to go out for leave certification that first building, we'd done about 30 projects. And we'd also had a couple of engineering firms look over our shoulder and tell us, yeah, we'd pretty much captured all the low-hanging fruit. In the process of going through certification in one year, we found 34 more projects that were, we spent as much again as we'd already spent, but had an even better return on investment, 148% return on investment for those 34 projects that, according to the engineering firms, didn't exist. And, you know, now it's about 140 projects or something like that. And overall, in the beginning, our payback was less than a year when we bundled them all together. Now it's, I forgot what I said, 1.2 years or something. And we're looking at projects that we maybe have a three or four year payback on a regular basis. So yeah, we've moved up the tree higher, but there's still things happening all the time. I mean, I can't tell you how many times we've gone through and changed out all our lamps to the next generation. And even with that cost of changing it out, have the whole thing pay for itself in less than a year. We're on, you know, in some cases, our fourth generation of lamps in 10 or 12 years. Last question. To what extent has the insights that you gained from this data enabled you to participate more broadly or nimbly in demand response programs? We don't participate them no many more. We started out, I think we were one of the ones in the very early stages. Now we have our own program. We basically are operating in-house doing demand response all the time because we're saving a lot of money doing that. And we're now starting to work with a company to go back and see if we can figure out a way to maintain and capture the peaks so that we have somebody to compare it against so that we can continue doing demand response, you know, multiple days a week during the summer, for example, but still recognize the peaks to measure against so we can also get the return from the demand response program. So I'm sure there are still more questions. I know there is another session starting shortly. You are welcome to come to the speakers. They will still be here. So please feel free to come back. Thank you.