Skip to main content

Hi everyone, do you know how I can create a chart to see WAU/MAU? Only been able to find DAU/MAU on the community. Thank you!

Hey @my88finds 

I haven’t personally dived into this metric yet, but can loosely translate the DAU/MAU formula to a WAU/MAU. This

Since our numerator is WAU, we will have to start with a weekly chart interval and do a rolling over the last one month defined in weeks. So a custom formula like UNIQUES(A)/ROLLWIN(UNIQUES, A, 4) with a chart interval of Weekly . Here’s a demo chart using this formula - https://analytics.amplitude.com/demo/chart/new/duv4m8x

This approach could be slightly inaccurate since there could be caveats in here - 4 full weeks need not translate to one full month always and 5 weeks overshoot the 30 day time period. Not quite sure if Amplitude allows to use of a 30 day window in the denominator, but still sticking with a weekly aggregation.

Would love to hear thoughts and alternate / better approaches from others one this one. Hope this helps to some extent.

 

 


Thank you, Saish! I will leave this open for others to chime in on how to produce this chart. 


Hi @Saish Redkar, thank you so much for your contribution to Amplitude Community!

 

Hi @my88finds, thank you for writing in and welcome to Amplitude Community!

@Saish Redkar’s answer is correct. I just wanted to add two workarounds for your consideration.​​​​​​​

  1. If your chart looks at a daily interval, you can use the custom formula: ROLLWIN(UNIQUES, A, 7)/ROLLWIN(UNIQUES, A, 30) to view the ratio of the number of users in the last 7 days  and 30 days. I’ve made an example chart in our Demo to illustrate. 
  2. For a date range based on calendar dates, a workaround would be to create two charts (one with DAU, the other with MAU), and use our Dashboard REST API or CSV export to export chart data for calculation outside of Amplitude.

Hope this helps! Let me know if you have any further concerns!

Best,

Thao


If you should be totally precise, I believe it’s possible something like below. But to be honest I would go with Saish’s approach!

  • create derived event properties for yearweek and yearmonth (assuming you can access derived properties)
  • create an event segmentation with two “any active event”, and group them by yearweek and yearmonth
  • set the chart to bar chart and set the time picker to last 12 months
  • use a custom formula to get the ratio

Reply