Most companies optimize prices to either gain marketshare or maximize profit. When marketshare is the goal, costs can largely be ignored–assuming you are a company such as Uber or Amazon and can afford to lose a lot of money for a long time, you price primarily to stimulate demand. When profit is the goal, you have to know your costs to set prices.
What costs to include: Marginal costs
Marginal costs are defined as the costs that you incur for each incremental rental. The list is short, here in rough order of magnitude:
- Depreciation, per mile or day owned
- Commission, per reservation or rental (or both)
- Labor, per rental
- Maintenance, per mile
Let's break each down in detail, and while we do that we'll outline what not to include. We'll focus on car rental, but the concepts hold true for any business.
Depreciation costs in price optimization
Because cars are worth less over time, we depreciate them. This can be done per-mile, or per day owned (that is, per day in service). Typical depreciation ranges from $0.20/mile for popular models to $3/mile or more for exotics; this translates to anywhere from $5 to $50+ per day. What you as a company buy and sell the car for has a huge impact: if you buy in large volumes and get excellent terms, and resell at retail prices from your own lot instead of at wholesale auction, your depreciation costs may be much lower than a competitor's.
The formulae are simple, and both should be translated to a cost per day for the optimization calculations. Do this for each car class–it will vary widely!
For cost per mile:
(Acquisition Cost - Disposal Price) / miles driven
For cost per day in service:
(Acquisition Cost - Disposal Price) / days in service
They may be quite different and you should carefully consider which is more effective and profitable for you to use. A key assumption in using the per day calculation is that your utilization is relatively high and constant. For a Ferrari that is rented every couple of months for a day or two, mileage is going to be a much better measure. But for a fleet of compact cars at 75% or higher utilization, per day is easier to calculate and likely accurate enough.
The resulting cost per day could be anywhere from $0 (popular cars that hold their resale value) to $100 or more (looking at you, Range Rover!).
Commission costs in price optimization
Commissions such as fees paid to online travel agencies (OTAs) like Expedia, and global distribution systems (GDSs) like Worldspan, Travelport, or Sabre, can take a big bite out of profits. You might also have costs to a franchisor, or to other channels (e.g., a hotel concierge). Usually these are a fixed amount per reservation, and/or a percentage of completed rentals.
These are entirely incremental costs: if you did not get the reservation or rental, you would not have to pay. Therefore there is a strong reason to include them.
However, there is an even better argument for leaving them out. Direct business does not have this cost, but most OTAs stipulate in their agreements that you cannot advertise lower prices elsewhere. So some of your business incurs this cost, and some doesn't. If you include these costs, your direct prices will be much higher–probably costing you business. If you don't include them, you might set money-losing prices on the OTAs.
If your business is highly dependent on OTAs, or certain car classes are highly dependent on OTAs, then it makes sense to include commissions in your optimization calculations.
But if you have more direct business, or want to build direct business through investing in your website, you should probably leave commission costs out of your optimization. Customers aren't stupid: even loyal customers will look at an OTA before booking direct. If your direct price is 2x a competitor, they will not book with you–and they might never come back.
Commission costs vary from $0 (direct bookings) to $25+ (OTA bookings or franchisee fees). Because of this high variability, you have to think strategically about accounting for them in a price optimization.
Labor costs in price optimization
Labor related to each reservation should be included. Examples include time for a counter employee to complete paperwork and hand over keys, and for operations personnel to gas and clean the car. Delivery is a labor cost, too, if you provide it.
The equation is simple (note you can replace "reservation" with "rental" or do 2 separate calculations):
Cost of Labor per Reservation:
hours / reservation * $ / hour = $ / reservation
Hours should reflect the time actually spent engaged in the above activities for a rental. Hourly rate can and should include marginal overhead for that employee (that is, not your lease for the space they occupy, but costs above wages like health and benefits, workers comp, etc.)
You should never include overhead for merely having employees at a location who you would pay regardless of whether you had reservations. You could even make an argument that if you are required to staff a counter, there is no marginal cost for doing the paperwork at the counter (up to the point where you need a 2nd employee).
In general, labor costs costs might range anywhere from $0-15 per reservation–and possibly much more for delivery or operating a distributed fleet. While this may skew your profitability away from shorter or rentals, you may find that in some markets or for some cars these costs are negligible.
Maintenance costs in price optimization
There are two kinds of maintenance costs–routine maintenance, such as an oil changes and tire rotations, and non-routine, such as repairs like fixing chipped windshields or repairing body damage in an accident.
You may want to include the routine maintenance in price optimizations. This is usually a small number, like $0.01-0.02 per mile, but it can add up–if a car is driven 50 miles a day, that's $0.50-$1.00 in costs. It can be more significant for exotics or custom vehicles.
For non-routine maintenance, you can either ignore it if you do not have statistical significance, or you can average it out by car class or across your fleet. Remember, you probably have insurance and collision damage waivers (CDW or LDW). These non-insurance products, combined with insurance, may actually make this a profit center. If it is, you can ignore this "cost" and might even include it as ancillary profit per rental, lowering your per-day price.
Expect this cost to range anywhere from $0 to $0.02 per mile. It is worth converting into a cost per day, like depreciation, for use in an optimization.
Turning costs into pricing action
Just having an understanding of cost does not mean you can use it. For example, knowing that your depreciation is $0.21 per mile on a Camry does you nothing unless you have some estimate of how many miles a Camry will be driven.
Per rental costs are by far the simplest, such as the labor required to gas and clean a car, or fill out paperwork and hand the keys over the counter. While some may be considered fixed costs (see below), if they are truly variable costs, you can simply include them in pricing. These costs are also least likely to vary by the rental itself–it doesn't usually take much longer to fill out the paperwork on a compact car than it does on a Ferrari.
Per mile costs can be complex, but are easy to use when simplified into a cost per day. You can always ask us what simplifcations are productive and what might lead to problems when dealing with maintenance, depreciation, and other costs.
Per reservation costs are just like per rental costs–unless you have no-shows. No shows mean you have to make an estimate of no shows, and what those will cost you, and then decide whether to factor that in or not on a per-rental basis. One way to simplify this is to either move channels with high no show rates to prepaid, or to ignore commissions (see below for why this might be the best solution). Not that you ignore this cost in other business decisions–but for setting prices, you might want to. Like an estimate of miles per day, the best you can hope do is an estimate.
Once you understand costs, you can then translate them into actionable numbers, with per mile and per rental estimates, that your revenue management system can use. Setting prices is different than making strategic business decisions, so simplifying and estimating is a best practice, not a short cut.
The costs never to include in a price optimization algorithm
We've discussed the many costs to include, and addressed specifically where not to include certain costs. Here we'll summarize what never to include all in one place.
Never include costs that you incur regardless of whether a specific car is rented. These are called fixed costs. Some examples of costs to exclude from a price optimization are:
- Vehicle financing costs
- Business financing costs
- Concession costs
- Flat franchisee fees, membership costs, or connectivity fees
- Your website development or maintenance costs
- Building costs: lease, maintenance, insurance, other financing
- Cost of sales, marketing or corporate administrative personnel
- Location overhead
- Insurance costs you pay regardless of whether a vehicle is rented (but always include any insurance paid for per-mile)
- Employee salaries or hourly wages not tied directly to a rental
Do not ignore these costs. The price optimization algorithm is simply the wrong place to include fixed costs. Instead, make strategic decisions about your business with fixed costs in mind–to close an unprofitable location, or move fleet from an unprofitable to a profitable location.
If you put fixed costs into your prices, you will do tremendous harm to your business. Adding fixed costs to your price optimization leads to a death spiral. Here is how.
The fixed cost death spiral
To calculate fixed costs per rental, you divide fixed costs by the number of reservations. Because the number of rentals is a function of price, you get fewer rentals at higher prices. Meanwhile, fixed costs by definition do not vary with the number of rentals. As rentals continue to shrink, your cost allocation increases (fixed costs, by definition, do not change), making prices go ever higher–until you are out of business.
The fixed cost death spiral can result from any fixed cost. Let's say you start allocating $1,000 in fixed costs across 1,000 rentals. That's $1 per reservation. You add that to your price optimization algorithm, and the resulting price increase reduced demand for your cars by 20%. That's 20% fewer rentals, at a slightly higher price.
Now you have 800 reservations, and still $1,000 in fixed costs (remember, they do not vary). Each reservation is now burdened with $1.25 in costs. Demand usually declines rapidly the further you move from the optimal price, even in non-competitive markets, because people are price sensitive. The extra $0.25 led to prices that resulted in only 600 reservations.
Now your cost is $1.66–leading to only 400 reservations, a new cost of $2.50, and so on. To zero. Meanwhile your fixed costs do not change, and you have fewer and fewer reservations contributing to them. This always ends badly.
Still don't believe me? Here's the Oxford published textbook our advisor wrote explaining it complete detail, with all of the formulas and proofs. It'll be the best $50 you ever spend.
Accounting for costs in P2P enviornments
Peer to peer companies have an additional challenge: do they include the costs of the peers to supply the car, hotel, or car and driver when suggesting or managing prices?
This is a strategic decision of the P2P platform, and generally it seems most do not include these marginal costs for their suppliers. In this way they can stimulate the most possible demand on their platform, and manage supply separately (with bonues and other incentives). In theory, if suppliers view prices as uneconomic (ie, they lose money on each incremental rental) they will withdraw from the marketplace. This does not seem to have happened anywhere; meanwhile, low prices have driven dramatic business growth.
Uber has received significant press for underpaying drivers. However, it has created a multi-billion dollar business and claims around 10% of the global ground transport market, just 7 years after its founding.
Putting it all together
Price optimization should account for reasonable, variable costs. These include:
- Depreciation, per mile
- Commission, for fees to an OTA or GDS
- Labor, per rental
- Maintenance, per mile
Price optimization algorithms should never include fixed costs, because this will artificially inflate your prices and result in a death spiral, eventually destroying your business–a business that might have been profitable if pricing were done right. The death spiral will happen regardless of your competitive environment–but it will happen much faster if you have competition.
Remember, your price optimization algorithm should not be a "cost plus" model, taking your costs and adding a margin. Instead, you should factor your costs into a profit optimization, which at its core accurately forecasts demand at different price points. Only with this method can you maximize your profit.
We hope this article is helpful. If you want to learn more about price optimization, just reach out to us.