7 ms·

I'm going to do some cocktail napkin math and see if it makes any sense. The total amount of new vehicles sold in the U.S. in 2015 was 17.5 million [1]. Since

I'm going to do some cocktail napkin math and see if it makes any sense. The total amount of new vehicles sold in the U.S. in 2015 was 17.5 million [1]. Since the average is $33666, we can assume half are above, and half are below that price. In reality, there are far fewer than half above that price, since an outsized sale price pushes that average up. Half of 17.5 million is 8.75 million. The U.S. Census estimates the 2015 population was 323,995,528 [2], with 116,211,092 households [3]. With every household replacing their car every five years or so (total guess), that leaves us 23.2 million households that should have been buying a car in 2015. That means that about 37.7% of the eligible U.S. households bought a car worth over $33666. That is bizarre enough that it leads me to believe there is a serious fault in my methodology or it just so happened that car sales were abnormally high in 2015 due to an enormous rise in consumer confidence. Either way, I'm more confused now than when I started scribbling on my metaphorical cocktail napkin. [1] https://www.google.com/search?q=site%3Awsj.com+U.S.+Car+Sales+Set+Record+in+2015 https://www.google.com/search?q=site%3Awsj.com+U.S.+Car+Sale... [2] https://www.census.gov/popclock/ https://www.census.gov/popclock/ [3] https://www.census.gov/quickfacts/ https://www.census.gov/quickfacts/

rjsw 10y ago
Does your 17.5 million include cars owned by rental companies. I guess some will be taxis too.
- linkregister 10y ago
  Maybe? It's total car sales, this is cocktail napkin math. Wouldn't the amount owned by taxis and rental companies have to be absorbed by the U.S. population anyway? If anything it takes away the amount of eligible households, making the issue even stranger. This is because utilization of rental cars is going to be lower than purchased cars.
  - cloudsymphony 10y ago
    Without looking up the stats, my guess is that this number includes leases. Which artificially inflates the number for "new car sales". Not only that but manufactures have been offering insane financing 72 months .9% apr. Which brings down the cost to own for people who probably shouldn't be buying a new car. Car manufactures remind me more of financial companies then car companies.
    - czinck 10y ago
      They really are (or subsidiaries are). Ally Financial (or Ally Bank until a few years ago) started as the part of GM financing their auto loans. I'd bet Ford and Chrysler have similar groups, and foreign automakers as well.
    - woobar 10y ago
      Leases do not 'artificially' inflate new car sales. They are not really different from regular financing. In both cases the new car is sold and paid for by a financial institution. And the customer can use the car as long as all financial conditions of the deal are met. For instance, in both cases you can pay off your balance earlier to own the car outright. Or you can sell car before the end of the lease/loan terms and use proceeds to payoff the balance.
  - rjsw 10y ago
    Rental cars get sold on at fairly low mileage. My point was that "people" don't have to buy new every car that other people buy used. I have bought an ex-rental car in the past that was 1 year old with 10000 miles on the clock.
smallnamespace 10y ago
This is a classic market sizing question that you would get asked in any consulting interview. I didn't understand why you split the market into 'above-average' and 'below-average' prices though? I highly doubt the distribution is symmetric -- most of the probability mass is probably in cheaper cars, and there are smaller numbers of very expensive luxury vehicles pulling up the average price [1]. In that case, a much smaller number of households buy the 'above-mean-price' vehicles. Also, wealthy households tend to replace their cars more frequently; luxury car dealerships do much of their business in rolling people over every 3 years into a new vehicle lease. In any case, 116mm / 17.5mm cars per annum households is 6.6 years per car per household, which seems totally reasonable to me since many households own multiple vehicles, and many vehicles are owned outside of households. [1] http://fortune.com/2015/02/17/luxury-car-sales/ http://fortune.com/2015/02/17/luxury-car-sales/ - 'Top-shelf brands, with cars generally selling for more than $40,000, represent a big and growing slice of the car business. They now account for 11% of total new car sales, when measured by number of units sold'
- linkregister 10y ago
  Thank you! It's a good thing I haven't gone for any business consulting interviews lately. This context helps to explain what's going on, especially the cars per year metric you added. It would be nice to have more information, but it appears that 11% luxury segment can have an outsized effect on the mean sales price of the vehicle sales.
- pault 10y ago
  How does one size a market? I've always been curious about this. Where do you go looking for the data? Are there any biz-dev people that could point me towards some introductory material?
  - smallnamespace 10y ago
    One common technique is to use basic demographic estimates: https://www.preplounge.com/en/bootcamp.php/case-cracking-toolbox/identify-your-case-type/market-sizing https://www.preplounge.com/en/bootcamp.php/case-cracking-too... In a real interview, you normally wouldn't have access to a calculator or a computer. The point is to show you can arrive within a factor of 2-3x of the actual real-life answer just by smart guesstimating and doing arithmetic in your head. One reason is because consultants often get asked these sorts of on-the-fly questions by clients, and they need to provide reasonable answers without looking them up.
rdl 10y ago
People also weren't buying new cars in 2008-2011, so there are fewer used cars resulting from that on the market. Plus, interest rates are really low. So a new car bought in 2007 + 5-7 years would be getting replaced.
- woobar 10y ago
  Over the last 10 years average car and light truck sales were 14.5M/year. Even in 2008-2010 annual sales never dipped below 10M. Detailed stats - https://www.bea.gov/national/xls/gap_hist.xlsx https://www.bea.gov/national/xls/gap_hist.xlsx
Retric 10y ago
First, households have more than one car. If you assume on average cars last 15 years and the average household has 2 cars that's. 116,211,092 * 2 / 15 ~= 15.5 million new cars per year which is within the ballpark of 17.5 million new cars. Next, averages tend to be well above median. One BMW might bring the average up relative to 20 Honda Civics. Further, the low end of the market is more likely to buy used than new. So ~1/3 of the market is buying new cars, 15% of that is buying the expensive new cars and your looking at top 5% income earners at $167,000+k/year. Granted, not everyone making 150k is buying a 50k car, and some people making less than 150k are buying 50k cars. But, if you think in terms of that kind of income then a 50k car is less ridiculous than you might think.
- linkregister 10y ago
  Good analysis. The used car sales aren't captured by this metric so it automatically pushes the mean up. However, it still accounts for an exceptionally high sales year, since none of the used car sales are included in that 17.5 million figure. Maybe car sales are highly variable from year to year?
- duaneb 10y ago
  The average household has two cars? That seems ridiculous! Do you know anyone with three cars? The two car average implies quite a few!
  - ceejayoz 10y ago
    There are plenty of 3+ car garages in affluent suburbs.
    - Baeocystin 10y ago
      I'd wager there are even more 3+ car less-affluent families. Having multiple, older vehicles that are 1. fully paid off, 2. dirt cheap to insure, and 3. old enough that you can actually work on them yourself is a good way to make sure you have something usable at any one time, even if the general maintenance needs are higher than average.
jonknee 10y ago
People, not households, buy and use cars so that's a weird way to calculate. You're also not counting any work vehicles which are very numerous and often expensive vans or trucks. A quick search shows there were 260,350,938 registered highway vehicles in 2014. The truth is US auto fleet age is older than it has ever been so car sales are not abnormally high and could actually accelerate (assuming robots don't drive them all!). http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_11.html http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...
brownbat 10y ago
Enormous respect for someone rolling up their sleeves, showing their back of napkin math, and ending with, "Huh, that's weird (and very possibly wrong)."
taneq 10y ago
> With every household replacing their car every five years or so (total guess) I think the issue is here. Yes, most households don't hold onto a car for longer than 5 years, but not all households upgrade to a brand new car every time. (Heck, I've never owned a brand new vehicle ever.) So one new car sale could result in 2-3 households upgrading their vehicle, as cars are passed down the line.
totalZero 10y ago
> Since the average is $33666, we can assume half are above, and half are below that price Maybe you're not stating a symmetry assumption about the distribution of prices about this mean, but in general you can't actually assume that.

rjsw 10y ago

Does your 17.5 million include cars owned by rental companies. I guess some will be taxis too.

linkregister 10y ago
Maybe? It's total car sales, this is cocktail napkin math. Wouldn't the amount owned by taxis and rental companies have to be absorbed by the U.S. population anyway? If anything it takes away the amount of eligible households, making the issue even stranger. This is because utilization of rental cars is going to be lower than purchased cars.
- cloudsymphony 10y ago
  Without looking up the stats, my guess is that this number includes leases. Which artificially inflates the number for "new car sales". Not only that but manufactures have been offering insane financing 72 months .9% apr. Which brings down the cost to own for people who probably shouldn't be buying a new car. Car manufactures remind me more of financial companies then car companies.
  - czinck 10y ago
    They really are (or subsidiaries are). Ally Financial (or Ally Bank until a few years ago) started as the part of GM financing their auto loans. I'd bet Ford and Chrysler have similar groups, and foreign automakers as well.
  - woobar 10y ago
    Leases do not 'artificially' inflate new car sales. They are not really different from regular financing. In both cases the new car is sold and paid for by a financial institution. And the customer can use the car as long as all financial conditions of the deal are met. For instance, in both cases you can pay off your balance earlier to own the car outright. Or you can sell car before the end of the lease/loan terms and use proceeds to payoff the balance.
- rjsw 10y ago
  Rental cars get sold on at fairly low mileage. My point was that "people" don't have to buy new every car that other people buy used. I have bought an ex-rental car in the past that was 1 year old with 10000 miles on the clock.

smallnamespace 10y ago

This is a classic market sizing question that you would get asked in any consulting interview. I didn't understand why you split the market into 'above-average' and 'below-average' prices though? I highly doubt the distribution is symmetric -- most of the probability mass is probably in cheaper cars, and there are smaller numbers of very expensive luxury vehicles pulling up the average price [1]. In that case, a much smaller number of households buy the 'above-mean-price' vehicles. Also, wealthy households tend to replace their cars more frequently; luxury car dealerships do much of their business in rolling people over every 3 years into a new vehicle lease. In any case, 116mm / 17.5mm cars per annum households is 6.6 years per car per household, which seems totally reasonable to me since many households own multiple vehicles, and many vehicles are owned outside of households. [1] http://fortune.com/2015/02/17/luxury-car-sales/ http://fortune.com/2015/02/17/luxury-car-sales/ - 'Top-shelf brands, with cars generally selling for more than $40,000, represent a big and growing slice of the car business. They now account for 11% of total new car sales, when measured by number of units sold'

linkregister 10y ago
Thank you! It's a good thing I haven't gone for any business consulting interviews lately. This context helps to explain what's going on, especially the cars per year metric you added. It would be nice to have more information, but it appears that 11% luxury segment can have an outsized effect on the mean sales price of the vehicle sales.
pault 10y ago
How does one size a market? I've always been curious about this. Where do you go looking for the data? Are there any biz-dev people that could point me towards some introductory material?
- smallnamespace 10y ago
  One common technique is to use basic demographic estimates: https://www.preplounge.com/en/bootcamp.php/case-cracking-toolbox/identify-your-case-type/market-sizing https://www.preplounge.com/en/bootcamp.php/case-cracking-too... In a real interview, you normally wouldn't have access to a calculator or a computer. The point is to show you can arrive within a factor of 2-3x of the actual real-life answer just by smart guesstimating and doing arithmetic in your head. One reason is because consultants often get asked these sorts of on-the-fly questions by clients, and they need to provide reasonable answers without looking them up.

rdl 10y ago

People also weren't buying new cars in 2008-2011, so there are fewer used cars resulting from that on the market. Plus, interest rates are really low. So a new car bought in 2007 + 5-7 years would be getting replaced.

woobar 10y ago
Over the last 10 years average car and light truck sales were 14.5M/year. Even in 2008-2010 annual sales never dipped below 10M. Detailed stats - https://www.bea.gov/national/xls/gap_hist.xlsx https://www.bea.gov/national/xls/gap_hist.xlsx

Retric 10y ago

First, households have more than one car. If you assume on average cars last 15 years and the average household has 2 cars that's. 116,211,092 * 2 / 15 ~= 15.5 million new cars per year which is within the ballpark of 17.5 million new cars. Next, averages tend to be well above median. One BMW might bring the average up relative to 20 Honda Civics. Further, the low end of the market is more likely to buy used than new. So ~1/3 of the market is buying new cars, 15% of that is buying the expensive new cars and your looking at top 5% income earners at $167,000+k/year. Granted, not everyone making 150k is buying a 50k car, and some people making less than 150k are buying 50k cars. But, if you think in terms of that kind of income then a 50k car is less ridiculous than you might think.

linkregister 10y ago
Good analysis. The used car sales aren't captured by this metric so it automatically pushes the mean up. However, it still accounts for an exceptionally high sales year, since none of the used car sales are included in that 17.5 million figure. Maybe car sales are highly variable from year to year?
duaneb 10y ago
The average household has two cars? That seems ridiculous! Do you know anyone with three cars? The two car average implies quite a few!
- ceejayoz 10y ago
  There are plenty of 3+ car garages in affluent suburbs.
  - Baeocystin 10y ago
    I'd wager there are even more 3+ car less-affluent families. Having multiple, older vehicles that are 1. fully paid off, 2. dirt cheap to insure, and 3. old enough that you can actually work on them yourself is a good way to make sure you have something usable at any one time, even if the general maintenance needs are higher than average.

jonknee 10y ago

People, not households, buy and use cars so that's a weird way to calculate. You're also not counting any work vehicles which are very numerous and often expensive vans or trucks. A quick search shows there were 260,350,938 registered highway vehicles in 2014. The truth is US auto fleet age is older than it has ever been so car sales are not abnormally high and could actually accelerate (assuming robots don't drive them all!). http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_11.html http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...

brownbat 10y ago

Enormous respect for someone rolling up their sleeves, showing their back of napkin math, and ending with, "Huh, that's weird (and very possibly wrong)."

taneq 10y ago

> With every household replacing their car every five years or so (total guess) I think the issue is here. Yes, most households don't hold onto a car for longer than 5 years, but not all households upgrade to a brand new car every time. (Heck, I've never owned a brand new vehicle ever.) So one new car sale could result in 2-3 households upgrading their vehicle, as cars are passed down the line.

totalZero 10y ago

> Since the average is $33666, we can assume half are above, and half are below that price Maybe you're not stating a symmetry assumption about the distribution of prices about this mean, but in general you can't actually assume that.