Frontier Forum: How VPPs earn grid-scale trust [partner content]

This is a Frontier Forum brought to you Adios. Tonight as the monster storm moves out, the deep freeze settles in. Frigid temperatures, blankets of ice, and feet of snow still gripping a huge portion of the country. It's been a very active winter across the southeast. Arctic air, ice storms, back to back cold snaps stretching across multiple regions at once, and for Duke Energy,

Those conditions drove the system to record-breaking winter peaks. 18,000 crew members from 27 states, including Canada, are coming to the Carolinas to help restore power in what could be a historic storm. We often don't have two of our regions at one time super cold two weekends in a row. And so first it was the Midwest and the Carolinas, and then it became the Carolinas and Florida.

Stacy Phillips is the managing director of customer load management at Duke, and she has been very busy this winter. As line crews mobilize across Duke's territories to restore power, Stacy is mobilizing something else entirely. Tens of thousands of connected devices designed to keep outages from happening in the first place.

those days leading up to the cold, we are really busy just making one last check to make sure everything's good. And during those two weekends, we called almost all of our winter programs. Those programs wrap together load control switches, smart thermostats, batteries, and water heaters into a distributed fleet Duke can call on when the system is tight. Energy Hub is one of the companies helping coordinate those assets.

You know, we've been running events like crazy. So a hundred and eighty megawatts. Of flexible capacity we we just deployed, about 120 megawatts from kind of a traditional thermostat program, layering in batteries. So we're sort of stacking up these resources to both firm the resource while growing it.

Seth Frader Thompson is the president of Energy Hub. He co-founded the company in 2007 to help utilities harness the growing wave of smart devices in people's homes. Nearly two decades later, that wave includes thermostats, batteries, EV chargers. Coordinated in real time and increasingly capable of behaving like a power plant inside a utility control room.

What's interesting about this is we're having this conversation in February and we're talking about VPPs and it used to be that VPPs were really kind of like a summer peaking resource and now they're a summer and winter peaking resource and that kind of points in the direction that we're all moving. What's really changed here is the adoption of customer-owned technology that can be grid or interactive, namely thermostats, batteries. We're seeing more grid-connected water heaters.

And so those are much have come with a lot less friction than the days of your of of switches. We've got all these pressures on the system, you know, rising capacity constraints, load growth, you know, political pressures around affordability. VPPs are kind of fitting into this nicely. Like does this moment feel pretty different? What's cool about VPP is isn't it?

You know, they are both a genuinely practical, effective resource that can replace a Picker power plant, but it's also an opportunity for for the sort of average customer to participate in a way. And it's like this is something that you can do. that relieves pressure on the system, which improves the reliability and affordability picture for the entire base. And that's just a big change, right? And that that has happened because you've had this wave of IoT adoption.

That sort of smart started in things like smart thermostats, but has now grown to batteries and EVs and other types of resources. And basically the more adoption of that that happens, the more opportunity is for customers to participate and the more opportunity it is for utilities to kind of partner with their customers to to activate these VPP.

It used to be that you would activate a program for three or four hours and you would see a lot of attrition and you would get customer complaints about it and no utility wants that. Uh we're learning how to operate even the even those switches in ways that don't cause customers to opt out of events or leave the program. Um, and I think that and batteries obviously are great for that. You know, people are waking up to the fact that customers can can participate.

So the more that we're able to talk to our planners and um and our operators about the min about minimizing the impact on our customers. the better they feel in the operations room using the programs. I'm Stephen Lacey. This week, a conversation with Seth Frader Thompson of Energy Hub and Stacey Phillips of Duke Energy about the evolving role of virtual power plants and what it takes to make them foundational to grid planning.

It was recorded live as part of our Frontier Forum series. Virtual power plants are not what they used to be. They've evolved from one-way load control switches to connected fleets with real-time telemetry, load shaping, and increasingly automated dispatch. Today we dig into how those resources are becoming core infrastructure and what still needs to change for them to be trusted like PowerPoint.

Steph, can you talk about the reliability of this resource now? Um, in particular, what has changed that makes virtual power plants more reliable than they have been in the past? Part of it is strength in numbers, right? You have you just have very, very large scales of of you know the customer participation. So Stacey's program I think has a hundred and sixty thousand devices that are participating, which means that

You know, not only can you afford to have some customers say, Okay, no, today is not a good day for me to participate, you you not only have the ability to absorb small percentages not participating on any given day, it's also so big that it's very easy to statistically forecast.

And so the ability to sort of know with a high degree of certainty what the outcome is going to be, what the output of a VPP will be, without kind of forcing any particular response on any particular customer is the is is the core of it. The other thing is just th the internet and the IoT p powers so much of our lives now that it has become fundamentally pretty reliable and people are used to all this stuff just generally working.

And so what this has enabled is sort of a shift. If you're a utility operator who historically you were comfortable operating stuff that you know you designed, purchased, installed, operate, and maintain. You know, the initial jump there to we're going to rely on things that are diverse and kind of chaotic at the edge, but saying actually those things are just as reliable and maybe more reliable statistically that than the traditional resource, that's a big change.

I think the piece there too, Seth, is repeatable. What we say is it's got to be repeatable and reliable at scale. So something that works great, but no one wants it um is not going to be something that we can really s bring into the portfolio as quickly as um something that a lot of customers want. And a good example of that is like batteries. We watched batteries for a long time before we felt like enough of our customers were going to be purchasing them that offering an incentive

was then going to um be of interest to them because surprise, they're not purchasing batteries so they can provide load flexibility to our utility. I mean, there's probably a niche there, but Our rates are pretty low and so uh compared to national average, we are under average. So the value proposition maybe isn't there like it is in other utility in other utilities or parts of the country for that savings. So we watched it. We watched the prices come down. We watched

solar um at the residential scale grow and grow and said, okay, now it's time. There's enough demand in this market, brought together the partnerships with aggregators like Energy Hub, as well as our local trade allies. And um and other stakeholders who were interested in that. And 18 months later, we've got 45 megawatts of it. And it's a great program.

Yeah, and I mean I think what that ends up meaning is that the resource goes from like demand response was there for decades, right? But it was it was sort of the resource you were glad would be it was there when you needed it, but you didn't really want to use it.

And now it's a resource because of that repeatability that you can use operationally. And you can be confident that if you need to use it, you know, today I think maybe dozens of times per year, but moving toward hundreds of times per year, that repeatability just changes the picture. I also want to talk a second though what about what you just said around the value. I'm not sure it gets enough airplay that having the resources in our back pocket in reserves.

what a value that is to our customers. I go back and forth on like the glamour of activating the resources all the time and doing cool stuff with software and load shapes and just the value to our all of our customers, not just our participants. Of their capability being in reserves, which allows us to not have to purchase as much or not.

be ready to start up some um peaking unit and what that saves all customers in terms of um affordability on their bills. Um so I think the demand response classic use case at peak. um is a great building block for what's to come in the future.

And so Stacy, you you said earlier that like the planning and operations teams are increasingly seeing this as a helpful, important resource. Can you talk about how that fits into broader plans when you think about serving load growth. Oh sure. Astonishing low growth. You're gonna invest, I think, a hundred million dollars over the next five years to support

um data centers and advanced manufacturing. So when you think about the overall mix and building generation, doing up grid upgrades and investing in virtual power plants, how do they all fit together? Okay, I love that question. Let me start from the beginning. We've been operating, in our opinion, virtual power plants for 40 years. We have a million customers who participate in our programs with us.

We've got two gigawatts across all the utilities in the winter and two and a half in the summer. This is already real capability for us. It lives in the integrated resource plan. And the second you've got something in your integrated resource plan, you are planning to use it at peak and beyond. But you know, the integrated resource plan is how you make sure you hit your peaks.

It's funny. I feel like the operators and the planners get a bad rap sometimes in load flexibility space for being skeptical. They are the most supportive of us growing these uh resources. They want more tools. And so if you're an operator, you've got a whole box of tools you can work with to serve the load. And the customer resources are one of those tools. And each one of those tools has best use cases and strength.

And it just it's our tools are kind of still growing in their uh in their capabilities to your point before, in the they're becoming more able to do more than just the peak. And so our operators are getting used to that and we're giving them more tools to do that. On the planning side, our planners actually they start first with what can customers help us do at P.

And then they backplan the rest of the centralized generation based on that. They actually one of them has a a term he should uh copyright, shrink the challenge. We use the customer resources to shrink the challenge. And there are always

happy to have more customer resources. Then it's really my team you have to get through. You've got to prove that it's repeatable and it's reliable. And Seth probably knows that our thermostat program with Energy Hub was not in reserves, I think, until year three. If you've got seasonal resources, it's gonna take you a couple years to figure out how to best use them and optimize them and forecast them.

And then yeah, you do even though the control room is um our biggest cheerleader, they all you know, they're accountable for reliability. That's a very serious job. And so we bring them the reports and we say, okay, this is ready now. We we think it's reliable. Here are our reports showing how we forecasted it and how we expect to be able to use it. We even bring in third parties to audit our work and really prove to them that this is real.

Seth, I I'm I'm kind of curious. Going to Stacy's comments about how a lot of these resources have been around in Duke's territory for a long time, like what distinguishes traditional demand response from a modern virtual power plant? I know it's a spectrum, but like how do you really distinguish those?

Yeah. I mean I think some of it is sort of foundational technology, right? Do you have like do you have two way data? That was the big kind of innovation when sort of modern IoT came around. Um the ability to The ability to see data about what's available right now, the ability to turn that data into forecasts, the ability to

You know, if you're if you're logged into the energy hub derms and you're you're you're activating your VPP, can you see what's happening in real time? Or are you depending on sort of the control center to look for the signal and the noise on whether that so that's a big piece of it.

I think scale is the other thing, right? I mean I think Stacy talks about there being a threshold like somewhere around fifty megawatts where you know, you you you are actually at the scale of of a of a an individual picker plant and that's kind of the um that's kind of the hurdle to clear to get the the the control room interested. I'll say though, uh too uh for utilities out there who are like, I don't have a lot of money to invest in fancy software. Um and

uh figuring out how to connect all these systems together. Um, even though Duke is big, I still feel like we are a scrappy utility sometimes trying to figure out how to make stuff work. And Um, people make fun of me because we have a million switches in the field and um and the vast majority of them are one way.

And what we do, because they are only one way, we're not getting that data back, is we use um our investment in metering infrastructure. We have advanced metering, we got some really smart data scientists. who um after events they go back and they look for the load drop and they then Tell us what they think we got. That goes back into the forecasts. And it's just that circle that you would expect like any other uh generator to have. So use what you've got.

And start from there to prove out those use cases. And one day you too will have a hundred and eighty megawatts in the winter. Aaron Powell The the thing I I wanna say about that though is that like Duke is Duke is a demanding client, a demanding user of VPPs, but Duke is also sort of unfussy.

Right. Like you are not waiting for some perfect system to be in place to to be doing this. Right. To your point, you've been doing it for decades. You were doing it when there was no data, um, but you were able to gather data to prove that it worked. Um you've since moved along. You're now integrating some of this stuff into the the the single pane of glass.

But the way you're even if you look at the single pane of glass, you're doing that in a way that's sort of like get the basics done first, give the operator sort of like one place to go to dispatch all this stuff. And that is in contrast to some other plans that I've seen where we're saying, like, let's build out twenty million dollars worth of of

software infrastructure before we even stand up a program. And that's backwards. So that stands out as like just do it first and then improve it over time.

I want to turn to this uh white paper that the team at Energy Hub wrote and it introduces this concept, Seth, called the Heels test, which asks Whether a grid operator could tell the difference between a VPP and a traditional power plant. Explain what the Hewells test is and why you developed it.

The Hules test is about solving I think the trust problem that we've been talking about, right? Like we know that the that VPPs are viable resources to offset peaker plants today. They are a core piece of the stack. At the same time, if you look at, depending on what the VPP is made up of, there are kind of different capabilities. And we felt like the industry needed a framework for this.

So we started with this basic test of like basically can you fool the operator? Could you swap out a a a VPP for a traditional power plant and they can't tell the difference. Viewing that as kind of a north star on how do we just constantly improve our game in terms of the quality of data, the quality and accuracy of the response, the ability to integrate into core pieces of utility kind of planning and operations infrastructure.

So that that was the concept of the Hules test. You achieve the Hules test uh when you have basically removed any excuse. for from uh an operator or a regulator to say, okay, well these things aren't quite equivalent. Yeah, practically it's offsetting it, but I don't like it because of X, Y, and Z. So the Hules test kind of takes that away. But what I think is cool about

Uh what as you think about it is you you start to think about can we go beyond the Huels test? There's some argument in the industry about should we be calling these virtual power plants or should we be calling them distributed power plants?

Because a VPP can do more than a traditional power plant and you have the ability to support the distribution network to to help kind of continue to hit on those reliability and affordability angles that are associated with electrification. So if we think about, let's say you have

Uh you have millions of EVs that are charging on your network. The ability to have that EV charging happen in a way that doesn't cause you to have to it doesn't trigger a bunch of necessary upgrades to the distribution network.

If you've got huge fleets of batteries scattered around, how do you coordinate those in a way that supports the distribution network? And that's where you go beyond the Hewell's test into sort of the, you know, the great beyond of level four on our uh our VPP maturity framework. Hey Seth, can we call it the customer power plant? Can we give them some some credit for it? We should. Yeah.

Um, well, you know what? I have never loved a virtual power plant because I want everybody to know that it's real. I think it should be called the customer power plant, but that's probably because I'm like old school customer programs lady also. This is this is what I do, customer programs.

I mean if you do an analogy to the cl to the cloud, by the way, like that supports the you know, w everything runs on virtual machines in the cloud now, right? And that is very capable and so Yeah, we could argue about it forever. Yeah. Um I also agree with you that about the operators not being able to see the difference. Here's the deal. They don't want to see the difference. We need to deliver to them streamlined decision making.

Uh that's what I like about that idea of like they shouldn't be able to tell the difference. Now, I also would argue that it's okay that they're not the same when we talk about getting to parity with the power plant. Um, kind of like you've got a lot of different people on a team that you work with and they all have strengths and things that they're good at. You know, the operators, they know things about.

Um, CTs and being different from coal and how long can things run? Or how long does something have to come offline? Or how much notice do you have to give it? Like they know those things and Um, and as long as it's repeatable and reliable forecasting of how long how much notice do I have to give it, because everything has notice, how long can I run it? And then what do I need to know about it afterwards and when can I run it again? That's what they need to know. Yeah. This is a great point.

Yeah. And I know that this framework that you developed was like intended to get industry reaction, utility reaction. Were there any other reactions, uh, Stacy? D do you like agree? Is this a threshold that's helpful to do? I think that anything that helps demand response, uh, load flexibility practitioners think about how to mature their portfolio and talk about it um ex within the organization is

really valuable. Again, going back to me being like old school customer programs lady, I don't look at it from the technology perspective. I look at it from the customer friction perspective. And so if I look at my level zero switches, and I don't mind calling them level zero, um a couple of years ago we would run them for three or four hours at a time and customers would get very hot. Now we run them one hour at a time.

And customers don't leave the program from using it. So I, in my mind, I have moved from level zero to level two because when the operator thinks they can run something and no one's gonna get mad about it, like a power plant. they're gonna run it more. Um, and that's that's what we want. And then on the value side, you know, I work in mainly integrated vertically integrated utilities where the value is very different. It's not linear.

Um, as I'm out there replacing those level zero switches with that are one way with 4G switches, I don't magically get. um more value out of that. But again, I'm cre I'm I'm building the case internally for the trust. that these are solid resources and that they're gonna be here for a long time. Going back to that planning thing, um, there may have been people a long time ago who thought that

These resources weren't here to stay or they might not grow. I think that argument is over. Um Duke Energy is investing a lot of money to um upgrade our switches so that they can be here because they have been such a good resource for us. Um I'm excited to see how other use cases um play out um at the distribution level or in ancillary services, but in our markets.

We are operating on the peak hour contribution. Our system value is set by the avoided cost of generation, transmission, and distribution in that peak hour. All those other hours don't have much incremental value. It just works differently than in the markets do. Um now we are still operating our programs outside of the peak hour.

um when we need them. And we are looking to work with our regulators where appropriate on where there is incremental value that we should bring back into those programs. Because by the way, when we have more value, we pass that along to customers. and customers participate because we pay them. So we also want to see that our system values increase so that we can bring more customers into the program. I think what what you're touching on here is this issue of um

Like every you you talked about how every resource exists on a spectrum of kind of what its capabilities are, you know, how c how fast acting is it, how long how durable is it? Can you use it every day, etcetera? And yeah, you you talked about how switches have one one use case, smart thermostats have another. You know, you can flex a battery every day basically. Um uh EVs you can smart charge every overnight every night.

Um what we don't have in the industry right now is sort of regulatory recognition of those di different things. We tend to have kind of a one size fits all M and V approach. And so um part of why we felt this framework was necessary was, you know, this is a very data-driven industry. And it's also an industry that like, you know, lives and dies by like real results, like the power was on or not.

Um but in the context of this, there was a lot of hand wringing, or sorry, not hand wringing, a lot of hand waving about like what the differences are, we felt like let's just write something down and say, here is a clear set of milestones. We had to make some trade-offs in kind of how we defined those.

And then put some numbers out there on what we think the value is. The next step for the industry is to see do do regulators support that? Can Stacey get more credit for being able to reliably and repeatably flex a battery? three hundred and sixty five days a year and do it for both system wide peaks and local.

So let's go back to the VPP maturity model you and just dig into the different levels a little bit deeper. Can you lay out the different stages from level zero to level four and where do most VPPs kind of fit into that? Yeah. So so Stacy touched on level zero. Um you know, push a button and hope for the best. Um but with decades of experience, right? So what it what it lacks in uh in real-time feedback, it benefits from twenty years of studying it.

Level level one is sort of what Energy Hub started doing, right? Taking sort of connected devices with two way data, but still kind of push the button and hope for the best. And so it was It was open it was closed loop on uh did it work, but open loop on what's the load shape.

Level two is where we started building in consistent load shaping and typically going from sort of single season to multi-season. And so this is something we've been doing with Duke for, I don't know, uh two, three years here, where you can produce a really flat uh output from your VPP, you can do it with a single class of DER or you can blend multiple classes of DER. Um but it's basically still enhanced DR.

Level three is where you are you are you have a fully automated VPP, but this is not just about the capability of the VPP. To to pass level three, you also need to be integrated into utility systems.

Um, we feel pretty strongly that at level three this can't just be sort of like a bilateral contract where you promise you're gonna participate. It's really gotta be it's gotta plug into the tools that the operators are using daily. It's gotta plug into unit commitment, it's gotta plug into forecasting software. It's gotta plug into the models they use for for developing those IRPs.

And then level four is where you go beyond the capability of a of a power plant and you now have the the ability to support the distribution network. You're typically d solving for multiple objectives at the same time. So an example of that would be let's say you've got Um you've got local congestion. You're trying to prevent having to reconductor uh a feeder, you're trying to avoid

uh replacing a a whole bunch of expensive transformers with long lead line lead times. So you're doing that, you're looking at bulk grid conditions, whether it's congestion or prices. And then you're also looking to solve value for the customer. So you maybe the customer's on a time of use rate and you want to make sure that

uh that that their DERs are behaving in a way that saves the customer money while simultaneously providing those grid services to the utility. So that's sort of zero through four.

Does mass deployment of le level four VPPs feel like it's on the horizon for you, Stacey? Like what is that what's gonna take for us to see those level four resources operating consistently? Well, I only know my utilities very well. Um, and I'd say we are like a strong level two. Um and as you heard before, obviously we've got a lot of devices in zero and one. So

Um, but we are operating year round and thinking about this all the time. I think there's a lot of work to do in terms of connecting data systems. um internally and externally, a lot of cyber work to be done. I mean, I'm excited for customer resources to be the Swiss Army knife of you the utility toolbox. Um I I think it's further off, um, but aspirational and um exciting. Yeah, Seth, what is aspirational? What feels near?

I agree with Stacy. Like the smart thermostat programs are are are at a strong level two in general. And at level two, like I said before, they are they are clearly already providing the value that a peaker plant would would provide. Um so th again it's this is a trust issue, not really a capabilities issue.

Um I think the the the VPPs that are that are made primarily of batteries, you could make a pretty good case that they are at level three or very close to level three, depending on how well integrated they are into other systems. And then uh the work we do with EV smart charging, I think you could make a good case that kind of in isolation, you're just about at level four. But the whole point of level four is that it can't be in isolation. It has to be integrated.

And so for example, if we're doing you know managed G V charging, we need to have an accurate network model from the utility. We need to have asset data about kind of what the thermal load limits are of different pieces of equipment or what the sort of congestion looks like. And most utilities are realistically not at the point where there's a system we can just plug into and do full level four. So I I think level four is

I don't think it's a ten year goal. I think it's a five year goal. And I think level three is a is like that's what we're working on right now. Consistent level three across D E R classes is I think what we're working on right now. And it's not just the technology. There's some regulatory um pieces that some utilities might have to work with. In some of our um uh uh utilities, we have hours limits.

And um and it takes some time to kind of work through that and talk about what's the in again, going back to the system value is at peak. Um, what is the incremental value to customers? um and to the um the general system um for operating more hours of the year. And so we're we're working through those things, but they're not always as fast as you'd like. But I know I've also thinking about level four.

I've talked to um our energy control center director around like what would it take for us to be integrated into the energy management system? Where, you know, uh is that what this looks like when you actually have parity with um power plants? And the answer is yes. As long as our resources are sitting outside of that, they are kind of different. So let's get to four.

And Stacy, I would be remiss if I didn't ask you about this concept you're calling the the single pane of glass, but unifor unifying, you know, core utility platforms. Like how what does that mean and how are you integrating VPPs into that? In the Carolinas in particular, um, control software um for switches, there was one and for thermostats, there was another.

very niche software for our team. For a long time, if the control center wanted it outside of if they ever had an emergency, they would call And we would talk about it first. And that's just not a way to scale. I think we all know that. And as we're talking about here, we're looking for um the expansion of thinking about our products. So that can't sit on our team.

Meanwhile, in Florida, we had long had full integration of only you of all of our devices, which are entirely utility-owned. So we had some sense for If the operators have something that they can use that's lightweight. um and shows them what load drop they will get, we bet they'll use it. And so we had had a team here at Duke who had just finished building a tool for our energy control center. They already knew the way they liked to see things.

They were going back to the beginning, we were talking about like you don't have to spend a lot of money. Um, they figured out like what how do you make the dis how do you streamline the decision making for the operator to make it easy when the rest of us know it can actually get pretty complex. And so they built them something and

Um, we operated side by side for a little while and got the operators comfortable with hitting it. We showed them the data that we'd been running. We figured out how to run the programs without lots of customers leaving. Please start using this. And I actually last Saturday night, they did it. They didn't even I mean they've been doing it for a little while, but in an actual

Well, I gotta be careful with utility people here. I shouldn't call it an actual emergency. In an unanticipated need, they decided to hit it on a Saturday night. And um and uh they have hit it other times too, but we had like maybe talked about or we saw it coming, but it was really like that scenario that you build these things for. So that in the moment, an operator can trust it that it's gonna perform. And it did. It was great.

So if a couple years from now, if we're thinking about Whether VPPs have crossed this threshold to from, you know, sub more supplemental to foundational. What does that market look like, Seth? So I think we touched on some of this stuff. So I think if if VPPs are a core part of the IRP, meaning a core part of the planning process. If they're always on one of those screens in the control room, you know, they're being used daily by operations teams as sort of normal management of the grid.

If they're solving I it would be fine to just solve the peak problem, but ideally we're solving that more than just the peak problem. Ideally we're providing Um grid services twenty four seven, three sixty five. We're helping to affirm renewables. We're helping to to offset distribution needs. We're hopping we're helping to um keep energy prices low. And then I think if these portfolios are consistently performing at level three and four, sort of beyond the Hules test.

I think that combination of things will mean that VPPs will have played an important role in addressing affordability, improving reliability, helping utilities. uh meet their strategic goals and and and helping with the the decarbonisation mission. Stacey Phillips is the Managing Director of Customer Load Management at Duke. Thank you so much, Stacey. My pleasure. And Seth Frader Thompson is the co-founder and president of Energy Hub. Thanks, Seth. Thanks, Steven. Thanks, Stacey.

This conversation was recorded live as part of Latitude Media's Frontier Forum with Energy Hub. This is an edited version of the discussion. There is so much more. And if you'd like to watch the full video, including audience questions about customer participation, How data centers are leveraging VPPs and broader grid planning, head to latitudemedia.com/slash events and click watch recording.

Energy Hub works with more than 160 utilities across North America to build and scale virtual power plants using its EdgeDerms platform. Read Energy Hub's white paper outlining their VPP maturity model, and discover what VPPs can do for your grid at energyhub.com. And for ongoing coverage of virtual power plants, load growth, and grid innovation, subscribe to our newsletter at latitudemedia.com slash newsletter. Thanks so much for listening.