I don't know for sure.
I don't think there is a way to graph the rate of return on money invested in promotion. It's akin to the broken window strategy in law enforcement: nail the guys who jump the turnstiles (a small "victimless" crime) and you'll cut the crime rate (guys who jump turnstiles often have outstanding warrants for other petty crimes.)
While publicity and petty crime aren't in the same ball park (we hope!) they both have difficulty measuring results related to effort.
But back to bookmarks.
I recently attended Bouchercon in Albany. I ran into all sorts of authors, new, known, pubbed, unpubbed. Several were authors I'd known for awhile who now had book deals. Big huzzahs all around. Passing out of bookmarks very often.
Fast forward a week and I'm tidying my desk and what surfaces but a bookmark for A Killing at Cotton Hill by my friend Terry Shames. I've known her for years, had many pleasant conversations with her, had been glad to hear of her book deal at Bcon. Had I bought the book there? No I had not.
So, here was the bookmark to remind me, and today I ordered it.
And now I'm writing a blog post about it.
Net effect of that one book mark (building on years of attending Bouchercon, a snazzy website, and a lot of effort) is pretty good. Did she hand out bookmarks to people who didn't buy? Sure. But it worked for me, and now that you've read this blog post, maybe it will work for you.
So, do bookmarks work for promo? Yes. And they sure as heck are better than nothing.
My working philosophy for publicity and marketing is do everything you can, and one thing more.