are dropped. Shipping people get the count wrong. Some copies end up in the wrong box and go to a bookstore that didn't order them and isn't invoiced for them and end up on a sale table or in the trash. Some copies are returned as damaged. Some are returned as unsold. Some come back to the store the next morning accompanied by a whack of buyer's remorse. Some go to the place where the spare sock in the dryer ends up.
The numbers on a royalty statement are actuarial, not actual. They represent a kind of best-guess approximation of the copies shipped, sold, returned and so forth. Actuarial accounting works pretty well: well enough to run the juggernaut banking, insurance, and gambling industries on. It's good enough for divvying up the royalties paid by musical rights societies for radio airplay and live performance. And it's good enough for counting how many copies of a book are distributed online or off.
Counts of paper books are differently precise from counts of electronic books, sure: but neither one
An interesting and rational look at why an author would willing give out his book, free and in its entirety, on the internet