Did you think car rental companies make money from renting cars?
Read on to learn how the business of renting cars really works, what’s changing in this dynamic, nearly $100 billion global industry and how artificial intelligence is going to revolutionize it.
The real car rental business
Rental car companies like Avis, Hertz, Enterprise, and smaller players like Fox, Payless, U-Save and others make money from buying cars cheap and selling them for more than they paid. Just like a vacation homeowner, renting the cars out simply covers their operating expenses while they wait to resell the cars. Crazy, right? But it’s been that way for decades. And yes, that’s why you can rent a car for a day for $6.50 in Las Vegas sometimes.
Things may soon change. Under threat from (1) car manufacturers like Ford and Tesla who may rent cars directly to customers (especially, autonomous cars), (2) taxi-like services such as Uber and Lyft and (3) peer to peer (P2P) carsharing services like Getaround / Drivy / Car2go which connect owners and renters directly, the future of car rental is as uncertain as ever.
Another threat comes from declining residual values, or the value of the car when the company goes to sell it a year or two after buying it. Hertz recently lost half (yes, 50%!) of its market capitalization due to writedowns on small and compact cars which had poor residual values. Residual values are likely to be under pressure in coming years as purchases of new cars ballooned after the Great Recession, and cars have become more reliable and able to stay on the road longer.
Car rental challenges
The complexity of the car rental industry goes beyond any other industry – even airline or manufacturing. Car rental companies deal with hundreds of millions of dollars worth of capital equipment (the cars!), hundreds of millions of dollars worth of real estate, thousands to hundreds of thousands of employees, and millions of consumers.
Enterprise, the largest, has over 1 million cars, over 200,000 employees, thousands of locations worldwide, and tens of millions of customers. Even a mid-size company might own 2,000 – 3,000 cars worth $50 – 100 million, operate a dozen locations, and have hundreds of employees serving tens of thousands of customers.
Take a look at just one Hertz location, at Sea-Tac airport, to get a sense for the size of the challenge:
This complexity makes pricing and forecasting unusually complex, because of the number of variables involved. Each location, car class (compact, mid-size, full-size, SUV, etc), time of day and length of rent can have its own distinct price. Unlike, for example, an airline, a car rental company can move the cars around – and so can its competitors. They can even subsidize customers moving their cars around, by offering cheap intercity rates. Also unlike an airline, while a large car rental company with an extensive network is difficult to compete against, nearly anyone can open a small rental franchise and compete at one location. Nobody ever cashed in their 401(k) for a Boeing 737-800.
Today, most sophisticated car rental companies manage prices only vis-a-vis their competition. They look at bookings coming in, make a best guess as to how many they will eventually end up with, and then look at their competition’s pricing and price relative to the market. This has proven dramatically better than just the “summer” and “winter” flat pricing strategies of decades past. However, because it is completely reactionary, it consistently results in a race to the bottom, frequent sell outs, inverse pricing curves (where last minute pricing is cheaper than early bird pricing) and lost profits.
A future-proof plan for the car rental industry
The future of car rental requires proactively matching supply to demand, using data every car rental company already has. It means avoiding the race-to-the-bottom of matching competitors whenever possible, and instead making your own, intelligent pricing decisions in a way that creates competitive advantage. It also means running the infrastructure to accurately forecast demand, enabling better fleet decisions now and in future years.
Companies that base their estimates on simply multiplying last year’s number by 1.1 and assuming 10% growth may get away with it for a year, maybe two – but not longer. Companies who outsource their pricing to their competitors will fail even faster. The future of car rental requires leveraging the data that already exists to make rapid, accurate decisions.
Poker, blackjack and car rental
Imagine sitting down to play blackjack or poker in Truckee, California, in 1849. The games haven’t changed, but if you want to win, the way you play certainly has. Back then, few people played the odds. Even fewer counted cards. Casinos didn’t need facial recognition technology. But today, things have changed. A team from MIT won millions from casinos playing blackjack. People practice for the World Series of Poker year-round online, and the top table stays consistent year to year.
This focus on data and probability has permeated business – and now it’s car rental’s turn. Amazon has predictive fulfillment, knowing what you’ll buy before you do. It’s time for car rental companies to open the AI and machine learning playbook, and do the same. Just like reading The Theory of Poker, you’ll never be able to play the game the old way again.
Current pricing companies, including Rate Gain and Rate Highway, have improved the fortunes of many car rental companies. With long overdue automation, their clients can now change hundreds of rates at once. With automatic price scraping, employees no longer have to go to Expedia and manually search hundreds of prices, which is a broken process, as one revenue manager pointed out, because “you’ll never find what you’re looking for anyway.” But where they fall down – and where PROS, the much higher end, more expensive software also falls down – is understanding demand and being proactive.
By leveraging demand data – something every rental car company has, somewhere – every company can make better decisions. Without thinking, you are counting cards. Without effort, you know the pot odds and place the perfect bet – every time. Inverse pricing curves can be fixed. CEOs can stop outsourcing their pricing to the competition. Managers can move cars around based on actual demand, not best-guesses. Instead of top-down forecasts, forecasts can automatically be generated for each car type, every day.
You can move from playing in a dirty, drafty saloon to playing at the Wynn.
There are periods in every company, every industry, where leaders recognize that valuable data is being wasted – and decide to stop wasting it. For Amazon, it happened in 2003. For rental car companies, we think it is happening now. Today. As you read this.
It is not a time for fear. It is a time for excitement. Every industry that has implemented demand-driven revenue management has seen profits across the industry rise – not fall. Profits only fall in a race to the bottom environment, accelerated by price-matching software. Profits rise when you match supply to demand, and actively manage your fleet based on highly accurate predictions.
Whether you partner with Perfect Price, or hire your own PhDs in hard math, physics and computer science to go do it on your own, now is the time to collect the right data, build the machine learning and artificial intelligence infrastructure, and take control of your industry.
We’ll leave you with the example of Sears, back in the early 2000’s. A small company pitched their executives on an algorithm to recommend other things people might want to buy (Amazon’s “more like this” feature). The executives lampooned the idea, citing their collective 100+ years in merchandising expertise. One executive insisted on an “A/B” test. The algorithm outperformed the gut-instinct of the century-plus experience by 40%. That was 40% more sales, immediately.
When you look at your team, faced with a potentially industry-changing technology, do you want to just carry on with business as usual, or do you want to enable them with the power to achieve a 40% increase in sales?