How to Group & Analyse by Day of the Week in Amplitude

With this new feature, it’s now possible in Amplitude Analytics to group your user events by which Day of the Week they happened on, opening up a whole raft of new insights.

Are you an Amplitude user asking yourself whether your user engagement patterns follow a weekly cycle? Do you suspect that certain days of the week are more effective than others in engaging users or getting them to convert? Well now you can definitively answer those questions!

Setting up the Day of Week Event Property

To do this, you’ll need access to the “govern” add-on, which means you probably need a growth or enterprise plan. If you don’t have these, you’ll need to use a workaround to add your own day of the week property to your events.

If you have the add-on, head to govern and then derived properties.

Click derive new property and then name the property “Event Day”. All you need to do is paste the function “EVENT_DAY_OF_WEEK()” into the formula box, and hit save.

This derived property will now use this formula to add the day of the week event property to every event that is in Amplitude.

That’s done, so let’s head over to a chart and check out our new property. Open an event segmentation chart, choose an event (I’ve chosen Any Active Event), and then group by. Scrolling to the bottom of the list you’ll see “Event Day”, so group by that.

Now take a looking at the resulting chart — you should see lines for each day of the week, spiking on that day each week!

You might also have spotted two things.

1. The spikes are not perfectly limited to each day — there might be some events the day before or after. This is due to the difference in the timezone between the events and your Amplitude settings. You should either change your timezone to match, or just keep this in mind.
2. I’ve chosen the bizarre time period of the “last 28 days”. This is to ensure each day of the week is equally represented, because the default “last 30 days” will include an extra day for the two most recent days of the week!

Let’s dive into some use cases …

Use Cases

Unique Actives by Day of the Week

You might have previously noticed your product has a weekly cycle — seen as a repeating pattern on daily charts — well now you can confirm it!

Switch the chart to “bar chart” mode, and you’ll see the unique actives across the period grouped by day of the week. In this case, Thursday is the day that has seen the greatest number of unique active users over the time period.

Unique Actives Doing Some Action by Day of the Week

Change the event in the top left, and you can see the same thing but limited to a particular action — in this case “Play Song or Video”.

% of Actives Doing Some Action by Day of the Week

Switch to the Active % tab, and you’ll get the answer to the question “Are active users more likely to Play a Song or Video on certain days of the week?”.

Conversion Rate by Day of the Week

Head over to one of your funnel conversion charts, and break down by the Event Day property for step 1.

This tells us if the conversion rate differs based on the day that the user starts the funnel. In this case Fridays has by far the highest conversion rate!

Retention by Day of the Week of Signup

In a retention chart, we can again group by the property, and answer questions like whether retention rate differs by the day on which the user signs up.

Interpret carefully here — if you do notice a difference, it’s very likely to be due to patterns in your traffic, and may not be easy to action. Don’t ask your users to wait a few days to sign up on a “better” day of the week and hope their retention will be better — you will probably never see them to find out!

Closing Thoughts

This new dimension is something I’m really looking forward to trying out with all my Amplitude clients! Humans plan based on days of the week, both in their work and personal lives, and that will inevitably affect their engagement patterns with products. Let me know in the comments if you found a cool use case!

Through my agency Permutable, I help companies get more out of their analytics setup, interpret the data, and make decisions that lead to growth. Please get in touch if you’re interested, or subscribe below to get notified when I post more analytics tutorials.

30 Minutes Free Analytics Coaching!

I’m running a survey to understand more about my readers. In return for completing the survey, I’ll give you 30 minutes of free coaching. Here’s the link.