Faculty of Science

Big Data: A Game of Gas

After this winter, it’s safe to say most of us are looking forward to the first un-official long weekend of summer. Gassing up the car or boat might be low on the agenda, but Matt Davison thinks we should all start thinking about it now.

Dr. Matt Davison

Dr. Matt Davison

Davison, Professor of Statistical and Actuarial Sciences at Western University, says consumers can take the pain out of the pumps with a little strategy: buy your gas on Thursday before the long weekend price spike. In fact, he thinks we should always look for the pump with the cheaper price. It’s move, counter move, in the game of gas.

Davison and Daero Kim, Department of Applied Math, are using big data and game theory to help retailers determine the “sweet spot” for gas prices. One key research outcome is a mathematical model to determine exactly what gas stations should charge to maintain demand, make a profit – and smooth out the annoying highs and lows consumers face at the pumps. The team is working closely with Fredrik Odegaard, Assistant Professor in Management Science at the Ivey School of Business. Odegaard specializes in using analytics to inform decision-making and is a consultant to the gas industry.

“Gas is a particularly complex arena and it’s really interesting to think about how different retailers compete,” says Davison. Retailers have access to a lot of data: crude oil prices, wholesale gas prices and operating costs per station are some. Competitors’ prices are easy to obtain with a quick scan around the neighbourhood.

Undercharging? Overcharging? Davison's work helps retailers find the best price.

Undercharging? Overcharging? Davison's work helps retailers find the best price.

What retailers lack is a strategy that pulls together quantitative data, factors in consumer behavior, and generates a specific gas price for the day. Without a strategy, the team’s initial research indicates that retailers are potentially missing out on millions.

Enter game theory. Davison explains that the industry is like a multiplayer game: the constantly fluctuating price of crude oil is one player, individual gas stations are another, and the consumer is a third. Everyone is making decisions to optimize their position.

“There are many theories of how things should happen,” says Kim, Davison’s PhD student who is leading the project. “But how do they really happen? That is what we are looking at.”

Kim’s research has uncovered that gas retailers are a few cents off their optimal price per litre: either too high to encourage demand, or too low to maximize profit.

Daero Kim

PhD Candidate Daero Kim

What seems like just pennies can add up to significant revenue industry-wide: “If a gas station sells about 40,000 litres a day, 365 days per year, that’s over 10 million litres per year,” says Kim. “10 million cents is a lot of money.”

Kim is monitoring industry data such as price, costs and sales volume daily from 100 different areas each with seven to eight retailers. These variables are factored against the constantly changing price of crude oil and a good understanding of human behavior. Price is the main motivator for gas purchases unlike other commodities where brand preferences play a stronger role.

When gas prices rise sharply, more consumers panic and shop around. Emotion comes into play, and many people will drive a little further to get cheaper gas. This forces gas stations to lower their prices to maintain, and hopefully gain, market share. In this scenario the retailer is chasing consumers with lower prices, and there’s less margin to make.

When gas prices rise sharply and stay there, consumers start to change their habits by taking more public transit, riding bicycles and walking. Davison recalls the general panic around rapid price hikes from 70 cents per litre to $1.35 following Hurricane Katrina, and the renewed interest in transportation alternatives. This scenario results in even less profit for the gas industry.

When price skyrocket finding a place to fill up may become more of an emotional reaction

When prices skyrocket finding a place to fill up may become more of an emotional reaction.

Conversely, when prices fluctuate by only a few cents, consumers don’t change their behaviour. They are content to visit the same neighbourhood gas station as they always do. Kim says that in this scenario retailers can gain significant profit by adjusting prices upward in smaller increments.

This is all good news for the gas industry, but what about consumers? Like most people, Davision says he’s not going to change his plans for the long weekend just because of a hike in gas prices. But he might start shopping around.

“Consumers can force change by “voting” for the cheaper price,” says Kim. His theory suggests that if more consumers continually make the effort to purchase gas at the cheaper station in their neighbourhood, even on a long weekend, other stations will follow suit and lower their prices too. It’s a rare game where all the players can win.