Product estimation mock interview: number of Google Docs created daily (w/ Microsoft PM)

[@leijin7429]

He focuses so much on the nuances (like 5 days or 6 days for the enterprise user vs. 7 days for the normal user). None of this matters because the error margin on his other estimates are bound to be much more than these. It's like you are already off by 50% (which is fine, because we are really just trying to get the order of magnitude right), but you keep trying to figure out that 0.5% number.

[@amaansajid01]

I would have tackled it in a slightly different manner. Starting with who are people who use any sort of documents-> People in Jobs and Students (there are small business owners who use it too but that number should be small and will account for them later) After this what is the percentage of People who use Google docs vs Microsoft word or something else? Then make categories as Heavy, moderate, and light users and assign a percentage and number of docs created by each category per day. solve and get an answer. Would love for someone to poke holes in my logic, so I can learn more.

[@SonOfMan-YT]

I wonder why the interviewer was so focused on DAU’s. Instead just focus on the number of docs likely created by an enterprise user per day (likely 0.25-1) and number of docs created by consumers per day (likely 0.01-0.1), then multiply those odds by the sum total number of users in each segment.

[@bk6103]

compared to other mock interview channels, I really like the flow of these. It helps to see interview notes posted while the interview is being conducted.. to validate my own thoughts about the answer.

[@LALITMITTAL1604]

I think docs creation can be considered under 2 buckets - Professional and students (enterprise) and Personal. IMO, enterprise is majority case compared to personal ( Dont think people do use docs or any other software in day to day personal life, Rather notes and other tools are more handy for writing something down). Now, Simple formula for Google Docs created can be => (Number of Professionals or students ( white collar job folks) X Average Number of Docs created per month X Google Market Share in docs market ( compared to Word, Confluence and other small players)) / 30 . This assumes weekdays and weekends are alike. For personal, replace Number of professionals with Number of people having access to Internet. 1) Number of Professionals can be something like => Total US Populations X Internet Penetration X %age people in white collar job . Similar, calculations for Student community can be done. 2) Average number of docs can be => Segregate users into heavy, medium and light and take some assumption of docs created per month. But definitely, this one is a tricky guesstimate question if done on the spot.

[@leihuang477]

His answer can be summarized in one line: 330 million * 70% * 70% *70% * 70% * 70% ....

[@devingu5554]

dude lol!

[@sheshangajwalia4606]

Lol😂

[@sophieshen6054]

Feel like a good demo of what a bad PM estimate interview looks like…

[@millenialmusings8451]

Yep

[@kaushikibhattacharya9976]

As per my understanding, we only need a Google account to access Google docs. Why did he talk about gmail users then?

[@kinshuknigam3367]

He needs to explain why 70%. A lot of assumed %s

[@manishbharti7475]

Felt he overcomplicated it

[@Brickkzz]

He's not very concise ...

[@cicciove178]

I wonder why he calculates the average between the 70% of internet users with a GMail (which I think it's his way to consider google accounts) and the 50% of professional accounts: as per my understanding, if 70% of all the internet users has a personal Google account, then this is a lower boundary; hence in the 50% of professional accounts there can be some without and some with a personal Google account, the 70% could even increase.

[@MsSmile90]

The last part does not answer how to sanity check that answer of 3M. He mentioned that he must have over estimated it a bit, but he didn't explain how to cross verify. Good approach overall and I got 3.25 as the answer myself. This was very helpful.

[@ravidubey2714]

taking weighted average could have done the job

[@this_is_anmol_jain]

He very much overcomplicated a answers