PART 5: 3 Powerful Ways to Beat the IESO Peak Tracker

In this five-part series called "Tame the Beast," the goal is to help you understand:

  1. The different players in the Ontario electricity market - setting the field;
  2. What happened to Ontario's electricity prices - the beginning signs of trouble;
  3. The impact of electricity peaks and the Global Adjustment - the serious situation;
  4. The story of how many companies are trying to reduce their electricity costs - the struggles of demand response; and,
  5. How companies are successfully reducing their Global Adjustment costs - how to tame the beast.What is the Global Adjustment?

As outlined in the previous section, it is becoming increasingly difficult for Class A consumers in the Province of Ontario to accurately predict peak hours in advance. Given that accurately predicting the peak hours can lead to enormous savings for Class A consumers, many are looking for better solutions. Read on to find out how some Class A customers have discovered a way to calmly tame the beast, and save up to 70% on their electricity costs.

By maximizing their savings in the ICI Program with highly accurate predictions, some manufacturers have been able to get some of the cheapest electricity available in all of North America in the Province of Ontario. They thereby succeeded in eliminating some of the critical competitive disadvantages that have been inflicted upon them prior.

The publicly available peak tracker predictions provided by the IESO are becoming increasingly inaccurate as they are unable to account for the demand response efforts of other Class A consumers in the market. In addition, the simple linear regression spreadsheets employed by some energy consultants are also unable to deal with the enormously complex task of predicting the actions of thousands of different Class A consumers. Without accurate predictions, however, demand response efforts become far too expensive and inefficient.

With thousands of Class A consumers now eligible to participate in the ICI Program, the cumulative effect of their demand response efforts is making most predictive models obsolete. As most businesses use the same public IESO (Independent Electricity System Operator) peak tracker as the source for their energy demand data, they also respond in a predictable way – which causes concurrent and massive reductions in energy usage. One example of this occurred on September 26th, 2017. Many businesses were using the publicly available IESO peak tracker data, which predicted that the peak hour would occur between 6 and 7pm. As a result, all of these businesses reduced their usage at the same time, and caused a massive decrease in energy usage. This then resulted in 7pm actually not being a peak, with peaks instead occurring the hour before and after the target hour, creating so-called “shoulder peaks."

This means that the businesses using the public IESO data missed the real peak when they responded. This is a major loss to many businesses, both due to the lost productivity resulting from shutting down their production when they attempted to reduce their electricity usage, as well as the lost savings from missing the peak hour. The savings in the Program usually more than account for the lost production, but if a business responds at the wrong time, they still pay for the lost production without saving any money.

In fact, the businesses that missed the real peak hour actually incurred even more costs. For example, if ABC Corp. is normally 1% of the Province’s electricity usage, then they would be expected to pay 1% of the Province’s Global Adjustment costs each month as part of being a Class A customer in the ICI Program (this was explained in more detail in section 3). By opting in to the ICI Program, ABC Corp. was planning to reduce their usage during the peak hours, and therefore only represent 0.5% of the Province’s usage. This would have allowed them to save 50% of their Global Adjustment costs. However, because they missed the peak hour, while other Class A businesses in the Province did not miss the peak hour, ABC Corp.’s proportion of the actual peak usage may actually increase to 1.5%. Not only did they not save money during this peak hour, they have actually increased their costs as they now have to pay more in their annual Global Adjustment fees.i

Given this, although the Global Adjustment Class A Program may well look very attractive due to the significant potential savings, those savings are far from being a certain outcome. As an increasing number of businesses participate in the Global Adjustment Class A Program, the task of finding which peaks are the true ones, and differentiating them from false peaks, is a task that grows ever more difficult. Moreover, the more advance warning an energy consumer has about an upcoming peak, the more they can potentially save during their response. This is achieved by them being able to plan ahead effectively and thereby ensuring that they reduce their usage in a financially efficient manner.

In an effort to help customers manage the ICI Program and maximize their Global Adjustment savings, while minimizing the impact to their production, EnPowered has created a predictive model that is able to address the three primary issues that Class A customers face:

  • Advance Warning;
  • False Peaks; and
  • Peak Hour Precision.


When a business is given advance warning of a potential peak day, the business is able to more effectively organize its demand response efforts. With advance warning, a business can schedule a plant shutdown to maximize their savings in the Program. As an example of some of the more creative ideas that EnPowered has heard from its clients, a business could organize a maintenance day on a peak day, or even organize their annual company BBQ. When a Global Adjustment Class A business has 5-14 days of notice, the shutdown efforts can be much more cost-effective and efficient. The result is that businesses are able to increase the size of their demand response, while minimizing the impact to their production.


The Ontario energy market, like most commodity markets, can be very difficult to predict without special insider knowledge. With that said, future electricity usage can be forecasted to a certain level of precision, which is a task that the IESO, independent stakeholders, and EnPowered all attempt to achieve. With the IESO, they have an admirable record of 96.7% accuracy with ± 470 MW, meaning that, on average, their predictions are within 500 MW of the actual usage. In contrast, EnPowered has a record of 98.3% accuracy with ± 250 MW.

The reason that EnPowered is able to be more precise is because EnPowered is not only able to predict what the Province’s electricity usage will be under normal circumstances, it also predicts how other businesses in the Province will respond during peak hours.

This additional accuracy is extremely valuable to Class A customers because it allows them to avoid False Peak days and thereby reduce their electricity and Global Adjustment charges. A False Peak occurs when the publicly available data provided by the IESO predicts that a peak will occur, but either due to changing usage patterns or a strong demand response from businesses in the Province, the electricity usage becomes lower than expected and no peak day actually occurs.

The difference in accuracy of 220 MW may not seem like much. However, that very same seemingly minute difference means shutting down just twelve times instead of twenty-eight in a given year, while still catching all five peaks. This translates to a significant economic savings, especially when one considers how much just one shutdown costs for a large business.

A False Peak day can lead to significant costs to a Class A customer. If the business responded to a False Peak, they have lost a significant amount of productivity with no benefit whatsoever. If a Class A customer is relying on inaccurate predictions, such as those provided by the IESO, they will be forced to shut down more frequently in an effort to catch all five peaks, leading to lower savings each year. In general, EnPowered would recommend that a Class A customer should only expect to respond to peak hours 10 times each year, and certainly no more than 20 times each year.


Rather than peak days, the true factor that matters is actually the top five peak hours each year. It is these peak hours which are used to calculate a Class A customer’s Global Adjustment costs each year. Most Class A customers, in an attempt to catch this peak hour during a peak day, will reduce their electricity usage for periods of 5 hours at a time. This is because they cannot be certain which hour will be the actual peak hour due to the inaccuracy of traditional predictive models. With increased accuracy, however, a Class A business can focus their demand response efforts on a single hour instead.

This allows them to increase their demand response efforts for the actual peak, and also reduce the impact to their production. In the 2017-2018 peak season, EnPowered was able to predict the actual peak hour with a high degree of accuracy. On average, EnPowered’s customers were able to avoid the peak hour while only reducing their electricity usage for 1.5 hours. With more accuracy in detecting the true peak hours, costly mistakes and perils like the “Double Peaks” described above, can be avoided.


Predicting the Ontario energy markets with such a high degree of accuracy is not an easy task, as previously outlined in Section 4. The way that EnPowered is able to deliver such accurate energy predictions is through the creation of a complex model that utilizes cutting-edge machine learning and artificial intelligence systems. Rather than using linear regressions and manually analyzing the many important variables that influence energy demand in the Province, EnPowered’s prediction algorithms are able to find correlations between data sets and different variables that interact with one another in an active fashion. This information is then processed and reviewed by looking at various variables and the relationships between them.

The forecasting itself is performed through a solution called a neural network, which allows for meta-pattern recognition. Where traditional forecasting falls short is when different variables themselves influence one another. However, this is not an issue for neural networks which excel at recognizing intra-variable relationships and do not run into the limitations that humans do when they analyze data. Specifically, for the EnPowered Ontario Demand Forecasting Model, the data that is used by EnPowered includes, but is not limited to, the following factors across all jurisdictions in Ontario updated and analyzed in real-time:

  • Wind Data;
  • Snowfall Data;
  • Precipitation Data;
  • Node Data;
  • Usage Data;
  • Temperature Data;
  • Response Efforts;
  • Real-Time Usage Information;
  • Special Events;
  • Among Others.

In total, the model processes over four hundred million pieces of data which are updated on five-minute increments in order to perfect the ability of the model to detect patterns and trends in the Ontario energy market, a level of computation which is simply impossible with normal predictive modelling techniques. Moreover, the model is constantly iterated and improved with new data and new algorithm adjustments for even greater efficiency and accuracy.

The result to energy customers in Ontario is that they can receive highly accurate energy predictions that are updated every five minutes, 24/7. As such, the results of this model are highly valuable to Class A customers as it allows them to maximize their savings in the ICI Program while minimizing the impact to their production.


In summary, as a Class A customer, there are four main options available to save money in the ICI Program:

  • Do-Nothing;
  • Public IESO Data;
  • EnPowered Predictions; and
  • Hardware Solutions.


The Do-Nothing approach is the simplest option for a Class A customer. A company can opt-in to the Program without actually making any sort of adjustments to their energy usage. During the five peak periods that occur annually, an organization could merely choose to continue using electricity in the same manner as they always have. This could result in some minor savings, but it could also lead to increased costs if the business uses too much electricity during the peak hours. In making the decision as to whether this option is one that makes sense, a decision-maker should take a look at their energy usage and see how it compares to the total energy usage within the Province. By calculating their share of the total provincial energy demand, known as their Demand Factor, they would be able to understand their potential savings in the ICI Program, and also understand the risk and size of potential losses.


The next approach that a company could pursue is to simply use Public IESO Data as the basis for their energy demand responses. This can be done for free, and most Class A businesses are currently using this public data, as are most of the energy consultants in the Ontario market. Given that most consultants use the same information, it is better to analyze the public data directly, rather than hiring an energy consultant to analyze the same data. This does require a significant amount of time from internal staff, however, and may result in little or no savings when the public predictions are inaccurate for the reasons enumerated throughout this article and earlier in Section 4 of this series.

It is notable to mention that rather than searching for the IESO peak tracker daily, you could simply opt-in to our email notification service which sends the same IESO predictions to your email address whenever there is something that you need to be aware of.


A company could also use EnPowered Predictions to more accurately predict the exact peak hours, without causing significant disruptions to their production schedules. By using proprietary predictions that not only predict the energy usage in the Province, but also predict how the rest of the market will respond, EnPowered is able to provide predictions that are 50% more accurate than the IESO. With more accurate predictions, a Class A customer can catch all of the peak hours each year, and only respond 10 times per year. In addition, when they respond, they would only need to respond for 1 or 2 hours instead of 5 hours.

Combined, this means that a Class A customer can greatly increase their savings in the Program, while minimizing the impact to their production. All of this can be done with little investment of time or resources by the business itself, while still allowing the business to greatly reduce their electricity costs.


Finally, a company could opt to invest resources into Hardware Solutions, which implies purchasing hardware like batteries, cogeneration units, or other systems that allow a company to reduce their electrical usage during peak hours. This will require a fairly significant investment of approximately $1-3M per MW that the business is looking to offset.

Although it may make sense for a business to install a storage system that is able to move an entire facility off-grid during peak hours, this is usually the most expensive option available. Instead, it usually makes more sense to first reduce electricity usage during peak hours by curtailing production, and then installing a storage system that offsets the remaining usage.

This option may not be available to every business, but it can help to reduce the investment in storage systems by over 50%. For example, a customer may use 4 MW of electricity on average, but may be able to curtail as much as 3 MW of usage during peak hours. This then means that they would only need to install a 1 MW battery instead of a 4 MW battery, roughly $6M less in initial investment. This is why EnPowered always recommends that customers first attempt to curtail their usage during peak hours, before installing a storage system.

In addition, most battery providers will attempt to install a storage system that runs the entire facility off-grid for 3-4 hours, as they want to be sure that they catch the peak hours. However, by using EnPowered’s market predictions, a business can run a storage system for only 1.5 hours and still catch every peak hour due to the greater accuracy of EnPowered’s predictive model.

By using these market predictions, a business is able to install a storage system that is 50-60% smaller in capacity while still retaining the same amount of savings each year. Combined, by following these practices, a business will be able to greatly reduce their initial investments and greatly increase their ROI.


The ICI Program for Class A customers is a great opportunity for businesses to reduce their Global Adjustment costs, and save up to 70% on their electricity bills. Unfortunately, it is also a very complicated program to participate in. As more companies begin to participate in the Program, it is becoming increasingly difficult for businesses to maximize savings in the Program while minimizing production impacts.

If you would like to learn more about how these predictions work, or how you can better manage the ICI Program in general, do not hesitate to reach out through the following link:

Contact me directly.


i “Peak Tracker for Global Adjustment Class A.” IESO, Independent Electricity System Operator, 10 Feb. 2018,