What Your Power Utility Thinks About You

Selling electricity causes one to become familiar with many delightful details about the rhythms of daily life. Electricity is a part of everything we do, so understanding how people use power requires us to understand how they live.

The operating schedule of very expensive power infrastructure is chained to when Americans brush their teeth or cook dinner. So when the power company adds you as a customer, they’re considering how the rhythms of your power use will harmonize with their existing customer base, and the impact that will have on their operations and investments.

First you have to understand the utility’s perspective. A regulated power utility¹ hooks up new customers, builds power generation capacity to serve them, and then charges customers for what they use.

As the economy in the utility’s service territory grows (and new residential subdivisions, logistics warehouses, strip malls, factories, etc. are built) the utility adds customers and has to invest in increasing its generating capacity. To recoup their investment in this new generation capacity, and to earn a return, the utility charges customers for the power they use.

The utility needs to have capacity to meet the peak demand on the system - the total demand on, probably, the hottest or coldest day of the year. If they can’t meet peak demand, they’d have to shut off people’s power at the worst time. This happened in Texas in February of 2021, and the cost has been estimated at $130-300 billion and hundreds of lives.

A profitable customer is one who increases usage of existing capacity, speeding amortization and increasing the return, more than they contribute to the need to increase peak capacity. Fortunately, this is true of almost all customers.

Imagine a grid with just one customer who has just one 10W light bulb. The utility would need to build 10W of generating capacity to serve the customer, but the customer would only use the lightbulb a small percent of the time - say, one or two hours per day, between waking up and sunrise. At that rate, the generating capacity might reach the end of its useful life before the investment to create it is paid off. This would not be a profitable customer.

Now imagine a grid with 1,000 customers who have, to keep it simple, just one 10W light bulb each. The maximum possible demand that could occur is 10,000W, but in reality the peak demand will be much lower. Half of the customers are early risers who only use the light before sunrise, and half the customers are night owls. Within each group, people have different schedules. And on any given day, some people will be out of town. So with 1,000 customers, the required peak capacity is not 10,000W but only, say, 900W. And that 900W capacity investment is spread out over 1,000 customers. This is a much more profitable situation.

The difference in these situations is the ratio of total system usage to maximum demand - the amount of usage that you can charge for, compared with the amount of usage you have to build for. The higher the ratio, the faster the utility can amortize its investments.

And critically, the reason that the ratio was higher in the second scenario is that there were more customers. The marginal customer increases the usage of capacity investments more than they increase the need for capacity investments. Every additional customer increases the profit margin, on average. It creates a virtuous cycle. As you add customers, you’re able to lower prices, which attracts more customers.

This is the real reason why utilities are monopolies. Wasteful duplication - the idea that it doesn’t make sense for competing utilities to build duplicate sets of power lines - is a part of it; high capital expenses are a fine barrier to entry for creating natural monopolies. But the more fundamental dynamic is the one I’ve just described.

Once this dynamic was understood in the late 1800s, it kicked off a wave of M&A activity. This was probably the most compelling synergy story in the history of private equity. People raised money to buy neighboring utilities and interconnect them to increase their customer base and therefore profit margins.

But not all customers are of equal value. The best customers have high usage but don’t contribute to system-wide peaks. The worst customers have a low usage rate but always contribute to system-wide peaks. A customer’s usage rate is called their load factor, the percent of their total capacity they actually use over some period of time.

Industrial power users are good customers. Many of them run at close to 100% of their capacity, close to 100% of the time - think of a logistics warehouse or a 24/7 factory. Residential power users aren’t so good. They’re out of the house much of the day, they sleep all night, and they all turn their air conditioning to maximum at the same time. Unsurprisingly, industrial power users are charged much lower retail rates than residential users.

Most factories take at least several minutes to shut down without the risk of damaging equipment, but some large industrial loads - like data centers that train AI or mine cryptocurrencies - are capable of turning off almost instantly. They’re called large flexible loads, and they’re the best customers for utilities because they have a high load factor but turn off during load peaks.

But every customer provides some value because they increase the diversity of demand schedules on the system, increasing the ratio of total usage to peak usage.

Commercial and residential buildings have a lower capacity factor than industrial, but commercial buildings tend to have high load when residential buildings have low load, and vice versa. So a customer base with both types of customer will have a higher capacity factor overall.

Parking garages have the lowest load factor of all, but if they offer EV charging, then their usage comes mostly at night. This is valuable because system load is typically low at night, meaning that overnight system peaks are rare and that generation investments are going under-utilized during the overnight hours. Here again, low factor loads are still valuable if they complement the system’s load schedule.

These are the kinds of things that your utility customer thinks about when considering you as a customer. Here are some others.

If you own an electric vehicle, you probably charge it mostly at night, so your load factor is probably higher. If you have solar panels on your roof, your load factor is likely lower overall, because you draw less power during the day while your solar is generating, but you’re less likely to contribute to load peaks on hot sunny days. These are both attractive qualities, especially in combination, and competitive retailers will make you special offers that help you both realize their value.

So far I’ve talked about the grid from the perspective of an old-school vertically-integrated regulated utility, which owns the power plants, transmission and distribution lines, and the meter and customer relationship. But in much of the country, these monopolies have been successfully broken up. The details differ, but generally the utilities continue to own and operate the distribution lines as a public utility. They’ve sold their power plants to independent companies that make money selling electricity into a competitive wholesale market. And now competitive electricity retailers buy power on the wholesale market and charge retail prices to industrial, commercial, and residential power consumers.

The principles I’ve explained so far continue to apply in competitive markets. Competitive retailers also benefit from larger pools of customers, and for similar reasons to the vertical utilities.

Although competitive retailers don’t necessarily own the generating capacity themselves², the costs of that capacity are passed on to them. The mechanisms by which those costs are passed on depend on the market structures created by regulators in that market, and they differ between markets. In most competitive regions, the independent body responsible for the stability of the grid forecasts capacity needs, then requires retailers to pay for part of the required investment according to their share of historic peak demand³. Those payments go to the owners of the power plants. So these retailers still prefer customers with a high load factor and low contribution to system peak demand.

But there are some considerations that are unique to competitive retailers.

For one, it’s easier to forecast the demand of a larger customer base. Retailers want to be able to forecast their customers’s demand so that they can buy wholesale power in advance and be market-neutral in real-time. Errors in their demand forecast will cause them to have a long or short position when they don’t intend to.

Their portfolio of customers will differ in the predictability of their loads. Residential customers probably have no idea how much power they use, whereas industrial customers may have contractually-defined load schedules⁴. (This is true for traditional utilities also.)

Another challenge unique to competitive retailers is that their customers opt-in to being customers, rather than the situation faced by vertical utilities, which is a captive customer base in a certain territory. The fact that customers opt-in creates selection effects, which can interact unexpectedly with some of the topics we’ve discussed already.

Selection effects could result in a retailer’s customers having too much correlation in their load schedules. For example if a retailer gets all of its customers by advertising during sports broadcasts, their customers will have a disproportionate number of sports fans. Their power usage will have an unusually high correlation with the schedule of sports games, and if the retailer doesn’t account for this, they could under-estimate demand on days with big sporting events on. The result would be that they don’t buy enough power in advance, and they’re forced to buy the difference in real-time at the volatile spot price.

Another customer behavior that the competitive retailer needs to forecast is account switching - signups and churn. The retailers will have many longer-term hedges in place⁵, and these hedges need to account for future growth or churn in the customer base over that time.

Customer switching is the key challenge faced by retailers. Because it’s hard to predict how their customer base will change, retailers are more hesitant to lock in long-term hedges - certainly moreso than power producers, who face relatively little quantity risk. This mismatch is why power futures typically see liquidity fall off 1-2 years in the future.

In theory, customers will switch to a different retailer that offers a lower rate. If a retailer locks in a hedge, and then market fundamentals cause lower power prices, then the retailer will have to offer higher rates than competitors. This could trigger unexpectedly high switching. In practice, customers don’t seem to care much, but retailers can’t count on this.

Selection effects could complicate this forecast too. Suppose a retailer markets itself as having renewable power, so its customer base has a high proportion of people passionate about environmental causes. If it comes out that the retailers’ CEO hunts baby elephants, then the retailer could experience unexpected high churn as their customers quit in protest. With their long-term hedges, they would have bought more power than they now need. The result would be an unexpectedly long position in the power market.

Or maybe their customer base over-represents certain occupations or industries, creating correlated economic risks. A downturn disproportionately impacting their customers could decrease load factor, increase switching, or increase the rate of bill non-payment.

Interestingly this is one risk that competitive retailers are better able to hedge than vertical utilities. Vertical utilities are tied to the growth within their territory. When the plant closes in a factory town, the local utility will see declining usage of their power plant. Competitive retailers are able to serve any customer on their grid, and they don’t necessarily have any large capital investments in power plants.

There are two advantage of having a ‘captive’ customer territory. First, traditional utilities don’t have switching risk. And second, they can make energy efficiency investments (eg. installing insulation) in the housing stock of their customer base. That kind of investment, that takes years to pay off, makes less sense for competitive retailers because of the risk that customers will switch to someone else.

The takeaway from all of this is that, to the power company, you’re not just a customer - you’re also a load factor, a switching risk, and a non-payment risk. In optimizing their marketing to maximize ROI, they’re taking into account what the channel that they’re reaching you by says about you, and about how your profile complements their existing portfolio of customers.