Accounting for Customer Attrition

Electricity customers buy on price and customer services, switching suppliers as a direct function of competiveness of price.

Despite the fact that the Retail Energy Supply Association claims that only 50 percent of residential consumers know they have a choice in where they buy their power, customer loyalty has been a relatively negligible factor.

Knowing that, customer attrition is an almost inevitable blotch in a power retailer’s profit/loss statement but planning for that attrition, with proper hedging and risk assessment, can determine the color of that blotch.  Not enough customers, combined with long contracts on power, turns that blotch dark red.  Setting the right rate and buying forward contracts for the right load quantity can keep the books in the black: the kind of outcome Ascend Analytics software users have come to expect.

Planning against customer attrition isn’t easy.  Guessing on the wrong assumptions with an unsophisticated deterministic model sets utility retailers on a high-risk trajectory.  With Ascend’s PowerSimm state-space modeling simulations, retailers gain a confident outlook of not only the expected impact of customer attrition on revenues, but also a clear idea of the uncertainty surrounding this revenue impact.  Using this information, a retailer can keep risks to a minimum.  “You know how much to contract for,” explains Ascend analyst Scott Nelson.  “You know how to hedge your loads properly.”

Ascend Analytics products operate at the state-of-the-art edge of statistical modeling.  With PowerSimm, retailers input their own predicted customer attrition curves.  These curves are then combined with the load and price simulations rigorously benchmarked to capture the key structural weather-load-price relationships and aligned to the current market outlook.  Within each simulation path, the incumbent rate is compared to the market price, and based on that differential, a percentage of the contracted load will likely leave.  The value of PowerSimm is in the precise quantification of the uncertainty surrounding this revenue impact.  Software that generates simulations using deterministic modeling can’t quantify the uncertainty of this revenue risk, since deterministic models fail to capture the uncertainty in volumetric risk, price risk and the correlations between the two.

Retailers often find themselves in a situation where a competitor can make a strategic decision to lower rates.  The shoppers might walk out of the store.  “All that revenue is basically at risk due to attrition,” Nelson notes.  Retailers cannot predict every competitor’s move but a robust simulation of market price can help anticipate the likelihood of such rate changes, allowing retail managers to plan accordingly.  “You might want to look into basically setting up some kind of hedge to mitigate that risk,” Nelson adds.  “You can’t actually model very effectively what rate a competitor is going to put together but you can say ‘this is where the market is and this is where it’s going.’”  If a retailer has an idea of where wholesale power prices are headed, that manager may be able to set a rate that slows or eliminates customer attrition.

So while customer attrition may be inevitable, the uncertainty around its impact can be calculated.  PowerSimm does that accurately, and automatically.  The customers might still walk out the door.  But when their departure is accounted for in Ascend’s PowerSimm, they won’t take the profits with them.

Plug-In Hybrids and Plugged-in Headaches

With estimates of new electric vehicles and plug-in hybrids climbing in every research report, you can already hear politicians boasting about “electric highways” with charging stations sprouting out of the ground outside every Starbucks.

But the truth is most EV owners are going to charge up at home and how long and when they charge will ripple across the power industry, affecting everybody from early adopters accessing “smart house” applications with their iPhones to coal miners shaking the dust out of their boots.  Ascend Analytics’s PowerSimm and Curve Developer programs can help portfolio managers keep that ripple from splashing over their profits.  Modeling for load, regulations and fuel prices, Ascend products factor risk and uncertainty into decision analysis.  Electric vehicles owners promise plenty of both.

A report released in July by the Electric Power Research Institute placed their mid-range prediction at 35 million electric vehicles on the road by 2030 and pegged the annual power consumption for those cars at 80 terawatt hours.

EPRI’s more optimistic estimate brought the number of EVs on the road up to 65 million, more than doubling the electricity demand.

At either number, the impact could be dramatic when residential customers start plugging their cars into high-amp chargers all timed to click on at 9 p.m. spiking the load on a system tuned to off-peak evening hours (a study released in September noted that 89 percent of customers intend to charge their cars at home).  Businesses running fleets of EVs and plug-in hybrids could be switching their chargers to “on” at entirely different times.

Utility managers know that off-peak and on-peak load can be served by different mixes of generation assets and they know that mix will shift constantly over time with fuel prices, regulation costs, transmission upgrades and new renewables coming on line.  Ascend Analytics products can handle all of that with precision, tallying uncertainty and optimizing financial performance.

With simulations for load, spot market fuel costs, renewables, forward option volatility, weather and regulation modeled in short-term and long-range runs, utility executives can optimize their portfolio to meet the potential growth in off-peak demand.  PowerSimm models load by the hour and into the years, with portfolio managers able to look at dispatch scenarios constrained to meet a user-selected set of inputs.  PowerSimm also models renewables by site so a manager might know that the night wind picks up at a particular wind farm just in time to charge a factory fleet of plug-in hybrids.

That kind of data granularity will be indispensable to utilities and retailers customizing rate plans for homes equipped with smart meters and vehicle charging stations or business fleets drawing power under demand resource contracts.

Nobody is predicting electric vehicles will replace internal combustion autos tomorrow, but with Ascend Analytics software, portfolio managers can be ready for those changes today.