This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
I would like to create a Monte Carlo simulation using the simulation sampling tool and I'm a little confused how to set it up. I have anonymized my data so that we can discuss the issue, and hope that by working this through on the forum it will be of help to others.
The example is a little contrived, so please bear with me!
I have three customer profiles. Assume my website serves each of these profile groups in proportion. I have a dataset containing 250,000 rows. I have analysed these customers coming from four different channels (A, B, C and D) and looked at the relationship between the number of page visits versus the total revenue. Specifically, I wish to examine the percentage of page views by channel versus the percentage of revenue by channel.
Please find attached a simple tableau representation of the full data, this details:
Page visits by channel in terms of the three profile groups revenue by channel in terms of the three profile groups performance, i.e. the percentage of total revenue divided by the percentage of total page visits
We can see that for the entire dataset, in channel A customer profile group 1 outperforms by 11%, but group 2 underperforms by 7%.
I would like to use the Monte Carlo simulation to construct levels of confidence around these percentages. I'm thinking that one way we can do this is by taking a sample of data and making the same calculation, and then repeating this exercise again and again. I would expect the mean outperformance in channel A for customer profile group 1 to be 11%, but what's the standard deviation around this? This is where the simulation tool would be useful if I knew how to use it!
Hopefully somebody out there will find this interesting!