Hey Diego! Thanks for this video. Mario gave some great insights, especially from the interviewer's perspective. What I'd like to add to make this guesstimate more precise is divding the the user base further into premium and non-premium users (guessing a 30:70 ratio). It will require an additional calculation but will give a fairer estimate and also show the the interviewer that one has thought about it.
Thank you very much Diago for you good works you put out here. Your prestigious YT channel has been of great advantage and help to me and my career transition as a whole. i can't stress that. ❤
Hey Diego. Thanks for the video. It was very helpful. Since this being an internet related question , can we also take the following assumptions? - 1. Internet penetration in the given region? - Since there can be regions with low internet availability, thus restricting the viewership of YT in those regions. 2. Access to YT premium - Since ads are not shown for individuals having access to YT premium. Thanks.
Overall great effort of explaining how to break such estimation questions into smaller manageable nuggets; one small thing that can be done to further simplifying it is by rounding up/down some large number for easier calculations. e.g. 198M can be rounded up to 200M
Apart from the demographic segmentation we also have to segment that 60% more into understanding who has a premium version of UA-cam. I guess this will make a solid case to back your assumptions.
Hey Diego .Thanks for the video. I have one question though. In one of my interviews i was asked to arrive at the traffic footfall for a big city international airport. My equation was = 2 *( no. of flights(arrived/departed)* passenger/flight) . my interviewer was Ok with the equation and how i scoped down the problem. We were even aligned on how to calculate passenger/flight . But both of us had different ideas on how to calculate the no .of flights. What to do in such cases when you can't arrive at a number with your understanding and interviewer is not aligned with your approach of coming at number.
Love this! Have been waiting for this for so long! Thank you so much! Mario mentioned that he took 60% because senior citizens and kids don’t use UA-cam but wouldn’t mentioning that young adults often use other platforms such as tiktok which is trending nowadays in the US and we also have senior citizens who do not use UA-cam as a better justification for the 60%? Also, quick query. What if I decomposed the problem further into paid and free users? I’m not sure if ads in the form of banners are present on premium subscription. But can I decompose the problem further?
Hi Anath, I think your logic is perfectly valid for the 60%. I just gave one reason, it doesnt have to be the only right answer. The important thing is to demonstrate you have some logical reason for 60% in this case. Also, its fine if you want to break the problem into premium and free, however, now you have to provide a number for both. You’ve set the bar higher for yourself by doubling your assumptions, calculations, etc. Its not a bad move, but consider you have a limited amount of time to solve the problem. Thanks for the question! Hope this answer helps
@@mariogarza5580 Yes! Thank you for getting back to me. I understand that the main essence of these questions is to break down the question into segments that I can answer without bringing in any complexity..
First of all, thank you for sharing your practical approach to solving estimation questions. Overall, it is very informative and helpful. However, I do believe that your approach is missing an important last step - i.e. validating if the estimate is realistic. In fact, looking at UA-cam's actual revenue for 2020 (~$20B), your estimate (~$30B per day!!) is far off.
Hi! Thank you so much for the feedback and the missing piece on the approach! You are right, at the end when Mario and I talked about double checking the numbers at the end we missed emphasizing the importance of “reality check!” Thank you so much for the idea and I hope people come into the comments to see yours! 😊
Hi Diego! I have one question to ask. In case someone is not familiar with the revenue system of UA-cam or for that matter any X,Y,Z platform, how shall we go about it? Shall we ask the interviewer directly how it works or tell that idk the algorithm? Or is there any better approach for this?
Hi Akanksha, you can always say something like: "Let me share my understanding of UA-cam's revenue structure, I believe that it is broken into (whatever your understanding is). Is this approach correct? or is there anything else that you think I'm missing?" And let the interviewer "help" you. My recommendation is to always show an initial idea of yours vs just asking for help.
Hi Murphy! While yes, a sanity check is definitely something I recommend doing, the number of 30M is not off, in fact, it should be higher. In 2020, UA-cam reported an annual revenue for 19.7 Billion dollars, or 19,700,000,000. This number, divided by 365 (daily) will give you 53.9 Million dollars every day. $30M x 365days = 10.95 Billion Dollars or 10,950,000,000 Let me know if this helps ☺️
@@PMDiegoGranados Ah I see so you meant 30M right? I thought 30k M because I confused the , and the . which is the other way round in my country. So we would then assume 57% of all revenue is in the US. I would agree this is reasonable since it's a US company but any advice how to sanity check such a thing in the interview?
Yes! I'm sorry for the typo there - I'll see if I can edit the video somehow. Thanks for catching it! Sanity checking these things in an interview can be hard... so what you can do is look at which of your assumptions is "the weakest" one. For example. maybe you could say that your assumption about how many people watch UA-cam is too high, so reduce that... or maybe the money made / ad. There's no straightforward way to do this, but double checking your assumptions always helps!
I believe $10 in revenue per video ad is way too off. That has to be corrected I guess. I think $0.05 or $0.1 is the ideal number. That will bring the estimates to a very realistic number. While the approach and everything is right, the number has to be corrected.
Hey Diego! Thanks for this video. Mario gave some great insights, especially from the interviewer's perspective. What I'd like to add to make this guesstimate more precise is divding the the user base further into premium and non-premium users (guessing a 30:70 ratio). It will require an additional calculation but will give a fairer estimate and also show the the interviewer that one has thought about it.
Thank you very much Diago for you good works you put out here. Your prestigious YT channel has been of great advantage and help to me and my career transition as a whole. i can't stress that. ❤
Hey Diego. Thanks for the video. It was very helpful. Since this being an internet related question , can we also take the following assumptions? -
1. Internet penetration in the given region? - Since there can be regions with low internet availability, thus restricting the viewership of YT in those regions.
2. Access to YT premium - Since ads are not shown for individuals having access to YT premium.
Thanks.
Amazing - Mario explained it brilliantly !
Overall great effort of explaining how to break such estimation questions into smaller manageable nuggets; one small thing that can be done to further simplifying it is by rounding up/down some large number for easier calculations. e.g. 198M can be rounded up to 200M
Apart from the demographic segmentation we also have to segment that 60% more into understanding who has a premium version of UA-cam. I guess this will make a solid case to back your assumptions.
Hey Diego .Thanks for the video. I have one question though. In one of my interviews i was asked to arrive at the traffic footfall for a big city international airport. My equation was = 2 *( no. of flights(arrived/departed)* passenger/flight) . my interviewer was Ok with the equation and how i scoped down the problem. We were even aligned on how to calculate passenger/flight . But both of us had different ideas on how to calculate the no .of flights. What to do in such cases when you can't arrive at a number with your understanding and interviewer is not aligned with your approach of coming at number.
Love this!
Have been waiting for this for so long! Thank you so much!
Mario mentioned that he took 60% because senior citizens and kids don’t use UA-cam but wouldn’t mentioning that young adults often use other platforms such as tiktok which is trending nowadays in the US and we also have senior citizens who do not use UA-cam as a better justification for the 60%?
Also, quick query. What if I decomposed the problem further into paid and free users?
I’m not sure if ads in the form of banners are present on premium subscription. But can I decompose the problem further?
Hi Anath, I think your logic is perfectly valid for the 60%. I just gave one reason, it doesnt have to be the only right answer. The important thing is to demonstrate you have some logical reason for 60% in this case. Also, its fine if you want to break the problem into premium and free, however, now you have to provide a number for both. You’ve set the bar higher for yourself by doubling your assumptions, calculations, etc. Its not a bad move, but consider you have a limited amount of time to solve the problem. Thanks for the question! Hope this answer helps
@@mariogarza5580 Yes! Thank you for getting back to me. I understand that the main essence of these questions is to break down the question into segments that I can answer without bringing in any complexity..
First of all, thank you for sharing your practical approach to solving estimation questions. Overall, it is very informative and helpful.
However, I do believe that your approach is missing an important last step - i.e. validating if the estimate is realistic. In fact, looking at UA-cam's actual revenue for 2020 (~$20B), your estimate (~$30B per day!!) is far off.
Hi! Thank you so much for the feedback and the missing piece on the approach! You are right, at the end when Mario and I talked about double checking the numbers at the end we missed emphasizing the importance of “reality check!”
Thank you so much for the idea and I hope people come into the comments to see yours! 😊
Hi Diego! I have one question to ask. In case someone is not familiar with the revenue system of UA-cam or for that matter any X,Y,Z platform, how shall we go about it? Shall we ask the interviewer directly how it works or tell that idk the algorithm? Or is there any better approach for this?
Hi Akanksha, you can always say something like: "Let me share my understanding of UA-cam's revenue structure, I believe that it is broken into (whatever your understanding is). Is this approach correct? or is there anything else that you think I'm missing?"
And let the interviewer "help" you. My recommendation is to always show an initial idea of yours vs just asking for help.
Thanks Diego
Wouldn't a quick sanity check in the end help here? 30k M per day would be 11 Trillion revenue per year which must be off by a factor of 1000x.
Hi Murphy! While yes, a sanity check is definitely something I recommend doing, the number of 30M is not off, in fact, it should be higher.
In 2020, UA-cam reported an annual revenue for 19.7 Billion dollars, or 19,700,000,000. This number, divided by 365 (daily) will give you 53.9 Million dollars every day.
$30M x 365days = 10.95 Billion Dollars or 10,950,000,000
Let me know if this helps ☺️
@@PMDiegoGranados Ah I see so you meant 30M right? I thought 30k M because I confused the , and the . which is the other way round in my country. So we would then assume 57% of all revenue is in the US. I would agree this is reasonable since it's a US company but any advice how to sanity check such a thing in the interview?
Yes! I'm sorry for the typo there - I'll see if I can edit the video somehow. Thanks for catching it!
Sanity checking these things in an interview can be hard... so what you can do is look at which of your assumptions is "the weakest" one. For example. maybe you could say that your assumption about how many people watch UA-cam is too high, so reduce that... or maybe the money made / ad.
There's no straightforward way to do this, but double checking your assumptions always helps!
I believe $10 in revenue per video ad is way too off. That has to be corrected I guess. I think $0.05 or $0.1 is the ideal number. That will bring the estimates to a very realistic number. While the approach and everything is right, the number has to be corrected.