Average Days between Purchases

[jtlrlfrc]

Hi all,

I've scanned endless posts trying to find a solution to this, but the ones I've found don't quite adress what I'm looking for. For some reason this is the one thing that continues to stump me. If you have a solution or a link to a thread I missed it would be really appreciated!

Here's a very simplified dataset I'm working with:

CUSTOMER ID	DATE
007	01/10/2016
001	01/09/2016
006	01/09/2016
003	01/08/2016
001	01/06/2016
005	01/04/2016
004	01/02/2016
003	01/02/2016
002	01/01/2016
001	01/01/2016

I'm trying to find a way to calculate how long each customer has been making purchases with us, which would be the time difference between the first and last time each Customer ID appears in the first column.

My more complicated issue is with trying to find the average time between purchases to see how long it takes on average for each customer to shop with us again. In the dataset above for example, Customer 001 takes an average of 4 days to repurchase. 

[ImkeF]

You can do this in the query editor:

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("ZczBDcAgDATBXu6NkM8QJ9Ri0X8b4F8sf0fadceLBkoXdhUadnPwkqxEVmkEfTW0RE/QTDSDtL4yaRDr/kf7AA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type text) meta [Serialized.Text = true]) in type table [#"CUSTOMER ID" = _t, DATE = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"CUSTOMER ID", type text}, {"DATE", type date}}),
    #"Grouped Rows" = Table.Group(#"Changed Type", {"CUSTOMER ID"}, {{"First", each List.Min([DATE]), type date}, {"Last", each List.Max([DATE]), type date}, {"All", each _, type table}}),
    #"Added Custom" = Table.AddColumn(#"Grouped Rows", "AverageTimeBetween", each if Table.RowCount([All])>1 then Number.From([Last]-[First])/(Table.RowCount([All])-1) else ""),
    #"Removed Columns" = Table.RemoveColumns(#"Added Custom",{"All"})
in
    #"Removed Columns"

Pls let me know if you need help implementing this code.

[jtlrlfrc]

Hi ImkeF, thanks so much for your help!

Are you aware of any way to do this in DAX? I don't have a strong enough background to adapt this code to my dataset in the advanced editor, I'm still very much a newbie.

[ImkeF · Accepted solution]

There is no need to do this via the advanced editor, all can be done through the UI. It's just that you might be able to study the steps if you past the code as it is into the advanced editor.

Have a look at the steps in this video:

[jtlrlfrc]

Hi Imke, that worked like a charm! In hindsight I made it much more complicated than it had to be, but I've been fiddling with PowerBI only a few days. Thanks so much for taking the time to explain it, I learned a little more today 🙂