Friday, January 18, 2013

Interview: How many cows are in Canada?

How many cows are in Canada?

I was surfing around stumbled on this interview question. At first, this appeared to be one of those break down the problem to smaller more manageable parts, like "How many quarters would you need to reach the height of the Empire State building," or "Estimate how many windows are in New York." But then I quickly realized, I do not have the faintest idea what the base case should be. Below is my thought process to answer the question with my current knowledge without looking up facts.

I think the premise of the question is really how I troubleshoot or go about the question more than the correct answer like the questions mentioned above. There are many factors that cannot be used because I do not know the start or end points of such data, like how much milk is produced or how much beef is consumed. I do not know how much is consumed domestically nor how much is exported for the former data point. I do not know how many people eat beef nor even the frequency.

To start, I think my solution would be some sort of range that I slowly refine as much as I can with smaller error. This will cause a wider range, but at least it is some information that can be used. Assuming there are cows that are kept for recreational purposes, I will assume this number is negligent. With that assumption, cows are raised for the purpose of consumption or to produce goods for consumption (ie milk).

Starting with calculating the upper range and consumption, I will assume cows (around 500 - 1000 pounds) weight about 750 pounds [from memory]. A person who normally eats beef will eat about 1 pound per week. Because most hamburgers and steaks are served in units of quarter pounds to one pound, and on rare occassion 32 ozs. Anything larger is assumed negligent [round up for upper range, down for lower range if necessary]. 1 pound seems like a lot of beef. So for 50 weeks (slightly less than a year but makes it easier to calculate), a person will eat about 50 pounds per year. So a cow will feed at most 15 people.

I do not know the population of Canada, but the US is about 300M (I actually know this number). Canada is about the same size but a good amount of the land area is not populated. I'd guesstimate that is is about a third of the US, so the population is 100M (also a nice round number). 100M / 15 = 6.67M, so approximately 6.67M cows are needed for consumption assuming all people eat approximately 1 pound / week. Another assumption is that Canadians eat as much or less than Americans.

Now, I have to compensate the number for actual population because cows do not grow instantly. Thus, 7M is a good starting low range for number of cows in Canada. Now I have to figure how long cows live for.... and I have no idea. Hmmmmmmmmmmm.... Cows eat grass... You can buy beef for approximately $5-20 per pound... I'll use $10 (pure guess and an easy number to use). Thus, $7500 worth of grass, feed, etc. Minus profit of middle stores, probably a third or even a quarter... so $2000. This is going now where quickly...

At this point, I will realize I am taking too much time, so I'll guess that it is about 5 years. My answer would be between 7M and 35M cows. With some sanity check that it is less than a billion cows and more than hundreds of cows.

As good practice, I would email research afterwards to show that I am also capable of following up and doing research. Provide online resources if you need to demonstrate that you can web search.

After minimum research, there are approximately 13M cows in Canada. Canada has a population of 35M. Approximately 91M cows in US. US has a population of 311M. US consumes 26B pounds of beef. 0.012 pounds of beef per person per year. Although the final number fell into the range I provided, using the calculations above, I should have come between 2.3M and 11.5M. A cow's life is between 1-2 years for meat and 15 years for dairy. An easier way to remember is that there are about a third of the country's population. I'd estimate less for other countries unless they export cows, but that's enough research for a blog.

Prologue

Also see post on other methods I would answer the question
After doing a few more practices of other forms of this question, I've found that I do have some idea approximately what the number should be. There may be some tendency to fidget which metrics to use to estimate the final number because my final numbers tend to be within a range that surprises even myself. So just remember, typically these questions is just your process of thinking. You may come up with the worst guess but may have the best explanation, and it is the latter that has greater weight and for some interviews may be the only weight.

Update

7/31/2014 - Checking on some old subjects and stumbled across a new link on how this question is not useful. I agree that there are more effective questions (as I've noted in My Take on Odd Questions). It does have its use but has too many variables to consider. Either way, a candidate should still be prepared for these questions because there will be some companies that will ask this type of questions. For me, I still enjoy thinking of different variants. It kind of has a nice intention to test one's imagination.

Reference

http://publications.gc.ca/site/archivee-archived.html?url=http://publications.gc.ca/collections/collection_2011/statcan/23-012-X/23-012-x2010002-eng.pdf
Google
http://www.greenkeyllc.com/blog/2013/06/google-no-longer-cares-how-many-cows-there-are-in-canada/

8 comments:

  1. Hi

    Tks very much for post:

    I like it and hope that you continue posting.

    Let me show other source that may be good for community.

    Source: Hr sample interview questions

    Best rgs
    David

    ReplyDelete
    Replies
    1. Thanks. The site is a really good source of questions. Many of them have been used in my interviews.

      Delete
  2. I started off like you but then realized I was making up too many numbers. So I simplified.

    I guessed offhand 10M based on about 1 beef cow + a fraction of a dairy cow per 10 people and <100 millions in Ca.

    So, according to this article,
    http://www.statcan.gc.ca/pub/23-012-x/23-012-x2011002-eng.pdf
    there were 12.5M cattle in Ca. om Jan 1 2012.

    ReplyDelete
    Replies
    1. Thanks Loring.

      Just make sure in the interview that you "talk out loud" your other guesses. Try to tie your other guesses to your simplified guess.

      For those who only gave a simple equation:
      You cannot avoid guessing but you still want to reduce the guess work. If your equation basically boils down to (guess) + (guess) / (guess) = (guessed answer), this really does not say much about you as a candidate except that you can do algebra. I'd would just figure that you just worked backwards. 100M people / 10M cow = 10 people / cow.

      The key to the question for most situations is to recognize this is a bizarre question (not meant to be an answerable question) and not a math or guessing question. The importance to this type of question is not the final answer but how you answer. Someone could provide a completely wrong answer but may have a better answer because the interviewer likes the out-of-the-box answer.

      But, what you have is not bad as a start either. I would have liked to know how you would expand on that. You could reflect how it would be impacted if it were another country like US, Mexico, Iceland. Or you could go answer how 1.25 cows is associated with 10 people.

      As an interviewer, what am I learning about you from your response? Although this is a static response, in an actual interview I may interrupt you for multiple reasons. One is simply to get clarifications. The other is how you deal with disruption.

      Or from a different point-of-view, what type of person do you think I am from my post? Is that the type of person you want to be? Maybe I jump around too much? Maybe my story isn't ordered properly? Did my grammar make it harder or easier for you to understand? Did I make it more complex than it needed to be? Were you able to follow my logic from assumption to final answer?

      Delete
  3. It's actually 1/4. "Canada" contains a "C," but no "o," "w," or "s." There are 4 letters in "cows." "Canada" contains 1/4 of "cows."

    ReplyDelete
    Replies
    1. That's a unique way of seeing the problem. I am curious how the interviewer would react to that answer. To build on that, I'd say 1/8 because c is about half the size of C... assuming case-sensitivity.

      Delete
  4. I think this is the kind of question when you should say:

    "well, I don't know the answer. However, in the 21st century, you don't have to know all the answers. What you need to know is where to quickly find the right answers and what are the best tools to use to find those answers". I would also add: "give me a computer or a mobile device with internet, using Google search engine, I can find the answer in less than 0.11 seconds". I'm sure this would have impressed the people from Google.

    ReplyDelete
    Replies
    1. I have read quite a few alternative solutions, and they all sound like reasonable responses. Personally, I am not a big fan of the short answers as witty as some of them are. To me, the interview is already short as it is and I want to make use of that time to impress the company. I would rather respond with an answer that demonstrates my strengths and value to the company.

      Perhaps using google search is valuable and could save time. But if you apply to an engineering position, I guarantee you that practically everyone knows about web searches... at least those I'll be competing with. If the response is unique for the industry that you're applying into, then that would be a fine solution. It may work with Google if everyone thought the same way I think thus no one would use it, but I find it unlikely. I think the time is better spent having an answer that highlights your strengths.

      But let's say I used an answer like that at a Google interview and the interviewer has the look like he is expecting more. I could expand on that and say that there may be several results. It would be important to be able to track the sources of the information so that you pick the correct information. I may even end up with some statistics that is 2 to 5 years old. How would I refine my search query? Perhaps, I can create a web scraping application that will take in all the search results, check the sources, and check the age of the information. The application will then be able to spit out the answer automatically. If you're applying for a law, audit, or something dealing with due diligence related position, you may also throw in the fact that it is in Google's terms of service that they do not actually allow automated screen scraping. Then come up with a different solution.

      Thanks commenting. My views on answering these types of questions has changed a bit after reading some of the alternative responses. I have been answering only from my perspective, but now I realize that I can change my answer depending on how the interview is going or the position that I am applying for.

      Delete