Thank you for amazing presentation. But I have any question. for 'CohortIndex' column, why the column was calculate with this (year_diff * 12 +month_diff +1)?
If we count without +1 the first cohort_index is 0. For example, when you need to count retention rate for 6th month, you'll use cohort_index 5. But if we add +1, indexes starts from 1 and if you need retention rate for specific month, you'll just take it from the same cohort_index
Hi Nick, thanks a lot for another great video! Do I understand correctly that for a cohort analysis, one always needs some kind of user/customer ID? I couldn't find and pull it from FB, Google Ads or GA. So, I guess I can't do any cohort analysis, unfortunately, except I get data with customer IDs from the client's backend?
It's perfect! Thank you very much!!
Here's a crown👑. You deserve it.
Thank you for the amazing presentation.
Great tutorial... very insightful
Thank you for amazing presentation. But I have any question. for 'CohortIndex' column, why the column was calculate with this (year_diff * 12 +month_diff +1)?
It is calculated in order to get the number of customer retained for each cohort month.
Sorry, from where it comes the invoice_year in the year_diff? That's not clear and is giving me a mistake!
In the create cohort section we create two variables using the function in the previous cell.
hi nick, I was a little confused about when you calculate corhort_index why +1?
If we count without +1 the first cohort_index is 0. For example, when you need to count retention rate for 6th month, you'll use cohort_index 5. But if we add +1, indexes starts from 1 and if you need retention rate for specific month, you'll just take it from the same cohort_index
Hello Nick, I have tried to install datetime to conda, but it seams there's isn't an option to MacOS. Do you know what I have missed? Tks!
I am attempting to do this in a weekly bases. But my cohort_Count looks wrong. I made the cohort index week_diff *7 days
Hello I get error on the step year_diff. The error comes because cohort_year is not defined. From where it comes this variable? Thanks!!!
Hi Nick, thanks a lot for another great video! Do I understand correctly that for a cohort analysis, one always needs some kind of user/customer ID? I couldn't find and pull it from FB, Google Ads or GA. So, I guess I can't do any cohort analysis, unfortunately, except I get data with customer IDs from the client's backend?
You need turn on user_id function in Google Analitics, export your data and after you can use pandas for your task.
Hi,
Could you please explain year_diff * 12 +month_diff +1 formula in creating CohortIndex ?
Liking your content a lot, thank you. Do you have a link to that python application spreadsheet? I'd love to copy pasta that if possible? Cheers
pls upload the data set also which is used in the video
If you follow the video you'll find the data...