Bob, you probably know that Google also loves to test estimation questions with its PMs, which is how many ping-pong balls fit in let's say, Empire State Building, and so on and so forth. I'll give a unique one because that one's a little tired and overused. Let's pick a city of your choice. What city would you pick? St. Louis. All right. Let's imagine St. Louis, and I want you to estimate how many piano tuners operate a business there? Okay. If you need a piece of paper, and typically, within a Google interview, you will actually have your own folder and some pieces of scratch paper, feel free to do that Bob. Otherwise, if you're really good at mental math, which I'm not, you're also welcome to do that. There's no right answer here. This is more about the process and talking through your assumptions. Let me think about this a bit. Piano tuners in St. Louis... Okay. In this particular case, I don't know if you're looking for an exact answer, but the process that I would go by to try and get at an answer, would be to look at a couple of different data points. I'd probably start out by trying to identify if there is a data that tells me about how many customers, piano tuners, generally, can service in a given day. I'd want to look at how many customers a piano tuner needs to have in order to have a sustainable business. Then I would look for data that would tell me about the percentage of the population that actually have pianos. I would then look at the population of St. Louis, and I would use that to identify, of the population of St. Louis, what percentage would likely have a piano? How many customers of that population, does a piano tuner need in order to be able to have a serviceable business? Because I would make the assumption, that if they did not have enough customers to service their business, they would not stay in business. Based on that, my estimation would be, the number of piano tuners would be, a factor of the percentage of people in a population that likely play the piano, a population of St. Louis, and the number that could support a going concern of piano tuners. I'd probably add in some degree of buffer in that, to assume that a piano tuner could also pick up some additional business that would be similar to that. Because generally, tuning of pianos can also go into other types of musical interests. I would buffer in a little bit of slack on that. But generally, I would estimate it based on those factors. Got it. Okay. Thank you for walking me through the assumptions. For the sake of time, I'll just describe typically, how in depth a Google inner mere might look for. All right, let's start with a city. I haven't really looked at how many people live in St. Louis. Let's just say it's going to be really under 10 million, right? We assume 10 million people, how many people within 10 million play piano? I'm a little over-optimistic, and I'll say 10 percent, which means that about one million people play the piano. One million people play the piano, out of that, how many people actually own a piano? Let's say half, so 500,000 actually own pianos. Out of those 500,000, let's say 10 percent of the pianos need to be serviced at any point in time. That's about, and again, 50? Yes, 50,000. Let's say on a given day, and this is where you went to the rate-limiting, I'm just starting your answer, a given piano tuner can serve, let's say, 10 pianos on a given day. How many days would it take to service 50,000? At this point, I would do some math on a paper and I would say, let's say, it takes about a year for one piano tuner to service all 50,000. Clearly that's velocity would be too slow, so let's parallelize. At which point, let's say, add n equals 10 or n equals 20, does the rate of piano servicing actually meet people's demand? There you would make an assumption and say, let's say we end up with 20 or 30 piano tuners. Typically, that's the overall process in a nutshell. No right answers, you could also just say like, St. Louis has five people or 50 people.
