Google Product Manager Estimation Interview: Paint Market

[@tryexponent]

Don't leave your product management career to chance. Sign up for Exponent's PM interview course today: bit.ly/3NA7qqC

[@gdr5804]

Really nailed it: The U.S. paints & coatings market size was estimated at USD 24.2 billion in 2019 and is expected to reach USD 24.8 billion in 2020.

[@anandprakashgoel7013]

The assumption on vehicles, what I understood, is to paint it every year. This does not make sense. Overall estimation is very good.

[@AlekseiVorobev-r3c]

And also it's worth to take into account the only American manufacturer cars market share

[@rusl12]

Really nice. The only thing I was thinking is that while the arts and crafts is a small amount of paint, the Units are way more expensive so it might actually not be that small. Having said that I didn’t event think about arts and crafts myself. The one can per room heuristic was very nice

[@georgejoseph7519]

How did you estimate the number of large buildings to ve 5M ?

[@vsjadon]

Is there any mistake while calculating cans required to paint industrial building? We have assumed 5M large building but I think calculation is done using 10 M large buildings figure.

[@aarbee96]

Inside and outside

[@pandeymani]

Arent, we considering the new houses getting built every year will change the estimates here; just wanted to call out!

[@leotini]

It will be great to keep sharing the document. It is easy to follow when you are reading the doc. You are showing the doc and then return to your camera. not good. but very nice case!

[@yummyday44]

Approach to the painter, with the assumption 90% of ppl will hire people to do the work. No. the painter labor is estimated with % of the overall population, times the average price of a can of paint, and assume a month of 20 cans of paints used which is 240 cans yearly. Multiply three of them will obtain a fair estimation market cap provided with solid data support of the number ~

[@vyche7882]

I wonder that why she did not divide the lifecycle (assuming residential houses may require 10 years to paint for new and residential industries for 20 years). Additionally, she also shared the 5M large building on residential industries but did not use this fact to estimate instead of using information from residential houses. Can you help to clarify?

[@MaaanasaP]

I didn’t catch it at first but if you go back she estimated 100M houses and then because the lifecycle was 10 yrs, she divided 100/10 and got 10M houses that needs paint every year

[@mycelleismybffl]

@@MaaanasaP but she didn't do that for the industries life cycle, she left that out

[@bhavyabansal484]

@@mycelleismybffl Yeah exactly.

[@MC-wp2vz]

For the car paint, I feel it is a different of paint and the price would be very different. It is hard to imagine that car paint is same as paint for residential houses.

[@pratsway961]

That was a great estimation.🙌🏼

[@superretroquest]

In my opinion the answer although executed well has two significant flaws: 1- commercial buildings are estimated based on the residential number, it should be significantly lower 2- number of cars is too high because not everyone gets a new car every year

[@MG-Alexandrovich]

Agreed, also what surprised me is lack of questions pre estimates. Perhaps they already discussed and staged it before video.

[@priyasinha8197]

Is she considering the rooms or the buildings are painted every year? As per my opinion for Households it should be 2-3yrs and for industrial buildings it should be 4-5yrs.

[@tryexponent]

That's a good point! You could probably bring this up in an interview as a tradeoff or exception you want to consider.

[@nikhilsingh3928]

Nice catch.

[@karthonk]

Oh hey you were my 186 TA lol

[@JorgeRTrejo]

So, after viewing some of your videos, i have a doubt: Why do all candidates assume they know everything (they do all the assumptions, without considering asking customers or getring actual data); also they asume there are not secondary sources of info (they create fantastic assumptions about a market they dont know and they assume its accurate). It literaly takes 2 secs to solve the question. Yet they explain a 15 mins process on making assumptions Instead of using 3 mins to get data and use it. Its like walking the long road to show you can walk it, while everyone in real life will just get the data and use the time on something more useful. Is this really expected on fantastic companies or is it just the romantic view of how interviews are?

[@CaseyMartin]

It’s expected. It’s not really about getting the right answer as much as it’s about determining one’s ability to pull apart a big problem and see the multitude of smaller things that can affect that big problem. By getting more and more granular, we can start thinking about the problem differently and opening our minds to potential solutions we might not have thought about when viewing the problem from a bird’s eye view. Or at least that’s how I see the value in these sorts of questions.

[@qz3024]

Google's idea of a PM is someone who could be a leader of a product, any product. They need to be so adaptable so that when they get thrown into a new industry they have never heard of before, they could use the same approach to break down the problem and delegate to the data team or engineering.

[@j10001]

I agree that this is in many ways silly. At minimum I’d like to see the interviewee say at the beginning and at the end that they are not an expert and could be way, way off in their numbers and assumptions about the structure of the industry. I’d also like to hear them articulate _how_ they would get data and which assumptions they would _start_ with testing (e.g., by interviewing experts or reading online). As someone who has spent years building complex models in spreadsheets, I agree that this is extremely basic thinking. It’s just laying out some cells in the sheet and typing basic formulas to link them (“we would multiply these and add these”). So my instinct is that this exercise doesn’t reveal much. *However,* can you imagine hiring someone who _cannot_ think through a problem like this? Someone who just freaks out mentally and stalls when confronted with complexity outside their comfort zone? Someone who cannot even begin to dig into a problem and pull it apart into bite-size researchable tasks? That is the person you do not want to hire. So this exercise is just screening out people with no intellectual initiative and no ability to break down a complex, unknown topic into pieces and do _something_ about it. When I was a young associate in a strategy consulting firm, I was assigned to supervise a young woman who asked to be staffed on a project where she could learn to build a financial model. She was a liberal arts major (I had an accounting degree) and had no clue what she was doing, but she was friends with the staffing people who said she had to be allowed to bike the model. It was a disaster. I had to walk her through every cell. It’s fine if you don’t know finance, but she couldn’t even do the kind of thinking shown in this video. She could comprehend breaking a problem into pieces and then rolling them back together again. Absolute disaster. She kept calling me while I was traveling between client sites and asking me what to do next on the spreadsheet, how to type each formula, etc. She seemed completely void of a capacity for analytical thinking. So yeah-there’s value in properly screening out people who can’t think their way out of a paper bag.

[@JorgeRTrejo]

Thanks everyone for their comments. I still think this question along with a couple more, will become the new "so what are your strenghts" and "tell me of a time you made a mistake". Repetitive and studied anwsers for repetitive and studied questions, you get nothing significant out of there, Jr. HR recruiters doing them by process without knowing what to expect, and extending recruitment processes looking for "perfect candidates". Just my point of view, of course Google and the rest know way better.

[@michaeljaf5734]

Can anyone explain why multiplied by 10^6 at the end?

[@bhavyabansal484]

Did I miss it or the 20 years for the industries was not taken into account, and therefore this came to a close approximation?

[@bhavyabansal484]

And for vehicles too, we didnt take into account the paint lifecycle!

[@itachi2023]

I dont get the calculation for the Industrial one. 5M buildings to be painted every 20 yrs and say 100 cans per building, I do not see it coming 1000M at all

[@ajinkya1443]

There are 5M buildings but she assumes that each building is 10 times the size of a house. So basically multiply the amount required for all houses by 10. That's where she gets 1000M (100M x10). She basically uses her first calculation and scales it for the other 2 calculations so that she doesn't have to start from scratch for each component. In my opinion, its a smart way to do it, unless the basic assumptions are varying lot. Hope this helps.

[@abhishekrbhat8919]

@@ajinkya1443 why did she not divide it by 20 since it needs to painted once in 20 years?

[@mycelleismybffl]

@@abhishekrbhat8919 agreed i don't understand why she didn't divide by 20

[@ashutoshandhari6597]

Shouldn't the amount for Industrial buildings be 500M and not 1000M? As 5M buildings * 100 cans/ building?

[@ashikag8775]

Yes, I think she made a mistake.

[@ajinkya1443]

There are 5M buildings but she assumes that each building is 10 times the size of a house. So basically multiply the amount required for houses by 10. That's where she gets 1000M (100M x 10). She basically uses her first calculation and scales it for the other 2 calculations so that she doesn't have to start from scratch for each component. In my opinion, it's a smart way to do it, unless the basic assumptions are varying a lot. Hope this helps

[@diptahi]

She made a mistake + the whole structure of the answer is very rudimentary

[@nehaom3829]

Even I am confused. She said the size is 10 times of the house and if it will take 10 cans to paint a house then it will take 100 cans for the industrial building and multiply it by 5M, you get 500M. I think she had just multiplied the cans needed using the residential but she forgot that residential gets painted every 10 years vs commercial every 20 years, maybe that is why the calculation did not make sense? or am I missing something and did not get it right?

[@rushkabab]

She didn't make a mistake. She is doubling the estimate to take into account both the interior and exterior. Same way she did for residential buildings.

[@ruchikajangid9408]

How did she end up deciding 5M industries ?

[@Hakunamatatism]

Error in cans needed for industrial buildings? 5 mn industrial buildings need 10x the amount of paint than a residential building = 5mn * 10 cans per residential building *10 = 500 mn. Not 1000mn.

[@sudhanz1]

I didn’t understand how are there 5M industrial buildings ?

[@gplayj9984]

It think in estimation questions they only see how we approach to the number rather than how accurate the number is. I am sure no PM uses such estimation in real work.

[@visheshbhaskar3315]

she assumed that there are 100M industrial buildings, that get painted on average of 1 time every 20 years. Which means that per year about 5M industrial buildings would get painted! Hope that clears your doubt :D

[@thepoetandtheprimrose3790]

@@visheshbhaskar3315 Okay, but the whole reason she got to the number of residences was because she assumed three people per residence. So she's also allotting three people served per industrial building, which makes no sense.

[@abhishekrbhat8919]

@@thepoetandtheprimrose3790 yeah exactly, the number of industrial buildings in her estimate is way too much

[@poojasinha5868]

Cars won't be bought every year. Assuming vehicles are retired after 5 years, it should have been 40M cans per year for vehicles.

[@narendraparyani]

That's a very good observation

[@LildAnakin]

very helpful

[@sudeepamitra]

Nice

[@kaushalsad6953]

Hey I'm preparing for APM roles, is there anyone who wants to join to take interview of eachother.

[@gmronit]

I can Kushal. Where are you based out of?

[@PJ-cu4cl]

@@gmronit Did you guys team up? I'm new to this I'm looking for someone to prep with.

[@devenjain7079]

First Asian I saw who is terrible at maths

[@agatabanach2766]

This seemed like an inaccurate, rudimentary estimation that got fairly close to the right result by accident. I think research should be encouraged before any estimation is done. Architectural and decorative paint has 57% share of the market, and these include not only buildings, but all general infrastructure. She missed most of the non-architectural market: wood, marine, protective, industrial. In the estimation she did, she got too large estimate for architectural because she missed the percentage of infrastructure that is not up-kept or that does not require paint, through use of other material (wallpaper, cement, brick etc.). In addition, her assumption that a car is painted every year is outside of any common sense.

[@agatabanach2766]

Adding to that, even if we're taking in account a raw number of sales, price for each segment should have been adjusted.

[@kevindeloscities]

That's not really the point of the interview tho. Estimation questions aren't about getting the answer right or about touching on each and every use case. It's about clearly articulating your thought process, how your approaching the problem, and outlining your assumptions all while engaging the interviewer. If you've interviewed at Google, Meta, etc they clearly articulate that that's exactly what they're looking for.

[@NitinSharma-ly9so]

the top down approach is inaccurate

[@rishiwork5922]

she forgot paintings for electric devices maybe?