Draw a rectangle 23 squares wide and 45 squares tall. Counting the squares one by one would take forever, so cut it along place-value lines into a 23-by-40 piece and a 23-by-5 piece. The first piece is 920 squares, because 23×4=92 and 40 is ten times 4. The second is 115. Together they make 1,035 squares. Cut both sides and you get four pieces, 20×40=800, 20×5=100, 3×40=120 and 3×5=15, which also add to 1,035.
The standard algorithm is the same cutting, written compactly. You multiply 23 by the 5 in the ones place to get 115, then multiply 23 by the 4 in the tens place. That 4 means 40, so its row is 920, which is why you write a 0 (or shift one place left) before writing 92. Adding the two rows gives the total.
Large numbers alone do not call for multiplication. If one shipment brings 23 markers and another brings 45, the total is 23+45=68: two amounts joined once. Multiplication needs a repeated equal group, such as 23 boxes with 45 in each box. Ask whether one amount is repeated, and how many times.
Think of it like this. Multiplying 23 by 45 is like counting the tiles on a floor 23 tiles wide and 45 tiles long. Instead of counting tile by tile, you measure big rectangular sections, like 20 by 40, and add the sections. However you cut the floor, the number of tiles stays the same.
Core idea. Multi-digit multiplication is ordinary multiplication with place-value bookkeeping.