a brand new concept. Utilities had been doing something similar to demand response for decades, in the form of what they call interruptible rates. The largest industrial customers in a utility’s service territory would be signed up for these interruptible programs and the utility in return would give them cheaper power all year long in return for the ability to actually throw a breaker and shut them off during emergencies. And the way those programs were used, a lot of times it was more of a political chip—an opportunity for utilities to give favorable rates to the largest industrial customers to keep them from moving out of their service territory. And they would never call. They would go five years without actually interrupting. So that when they did interrupt, the industrial customer would say “What is this, there is no way I’m going to participate!”
X: They’d forgotten that they even signed this contract?
DB: Exactly. Maybe the person who signed it didn’t even work there anymore. Furthermore it was prior to the advent of the Internet, so it would literally be the utility picking up the phone and trying to track down Steve the facility manager to ask him to manually take action at his facility. Steve may or may not be at his desk. He may be at the golf course, or if he’s there he might say “Absolutely not, we are in the middle of a production run, there’s no way I’m shutting down.” So with interruptible rates, the demand response industry had a pock-mark on it or. It wasn’t deemed such a reliable resource because it didn’t show up when actual push came to shove and they needed the resource. That was one of the challenges we had to fight.
When we came about it was post the advent of the Internet so we have this ubiquitous communications solution that is there for us to tap into and do this on a very cost-effective basis. We are putting up on every one of our sites a data logger, a device to install that gives us real time visibility into that facility where we can tell on a real-time basis that they’ve taken action and furthermore we are in a lot of customer cases taking direct load control, direct control over their energy devices. We can actually push a button from our network operations center here in Boston and effect that reduction automatically. So what we’ve been able to do is really demonstrate that with that technology this can be a very reliable, quick-to-respond resource.
X: Did that involve innovating a new technology, or bringing together existing technologies? Specifically, I can see how the Internet and cell phones and the like would get you around the problem of whether Steve is at his desk or out golfing. You’re more likely to get through to him either way. But in terms of the control devices and being able to directly modulate how much energy a customer is using—did that require any new technology?
DB: We are much more of a systems integrator, especially at the customer premises. We don’t manufacture our own devices, we use commercially available technologies and we layer on some software and intelligence onto the device and then put it out in the field. The real intellectual property that we’ve developed is back here in our network operations center, and it’s the software and the applications we’ve created to manage this network. How do we deal with data quality issues when we’re talking to 2,000 different facilities out in the field and you can lose connectivity temporarily to some of them. When we get data spikes or data gaps, how do we do the data validation and quality work.
There is a whole tremendous amount of IP that we’ve developed around baseline methodologies, and what I mean by that is how do you measure demand reduction. The only way to measure demand reduction is to come up with a baseline which would be the best possible proxy for what customers would have been consuming had there not been a demand response event. And then you look at actual readings during the event and the delta between the baseline and the actual reading is what you reduced. There has been a tremendous amount of development work we’ve done in terms of creating that methodology.
X: The baseline methodologies are important because your contracts compensate your customers to some extent based on the amount of reduction during a demand response event, right? So you do have to know what the difference is between what they would have been consuming and what they are now consuming, once they’ve ratcheted down.
DB: It’s not just during an event. They get paid on a year-round basis for the capacity that’s they’re committing to reduce, when and if they’re called upon. So in addition to figuring out what their payment might be during an event, we need to make sure that for each customer site we are living up to our commitment and that in aggregate as a portfolio, we’re delivering 100 percent or 110 percent of what we promised to the utility in terms of the overall load reduction.
X: That must mean documenting pretty accurately what the potential demand reduction is at each site. What are the challenges there?
DB: It starts at the very beginning during the sales process. What we do is, we have a fleet of site technicians, facility engineers that will go and do a demand response audit. They will walk the facility floor at each one of the sites that we enable and they will meet with the operations team of the customer to understand, A, what are the business rules, the business dynamics that enable us to participate in the program, and B, what are the physical assets that we can actually affect to reduce load. We then get the customer to sign up to the program based on that potential, which we turn into
