Have a look at this workout.
First I made a data model with Purchases/Stock, Sales, (FactTables) Product and Date (LookupTables)
I assumed that stock per product can be cumulated, at least I calculated a cumulative stock per product, but in your example, it does not matter, as you have only 2 stock values for 2 products.
5 measures made, some complexity was to avoid calculations of sell-through without sales taken place in the visuals.
It is setup that it calculates Sell through Rate over periods, you could e.g. select periods of years.
3 Visuals on 3 pages, have a look at the attached PBIX file.
Learning experience:
If you fastly make 3 tables out of your 1 Excel table (which can easily be done in the Query Editor), as a source for the data model, ambiguity appears, and one to many relationships from the lookup tables to the fact tables (waterfall principle) are not possible ( learned another thing today ). Apparently the one Excel table is still detected as source for the 3 tables. To avoid this I decided to make 3 “normalized” input tables in Excel.
I appreciate your comments ! Glad I could help you out.
Minor thing;
to avoid an error calculation for products with zero stock, I slightly adjust the last line of code in the measure Sell through:
IF( and( NOT( ISBLANK(SellThroughH)),CumPurchasesQ <> 0),SellThroughH)
It was before IF( NOT( ISBLANK(SellThroughH)),SellThroughH)