95% of bookends clients come from the query slush. About 2% from referrals, 2% from conferences, 1% other (eg inheriting someone else's).
Paying an agent to shop your book around won't help you because it presupposes that the only reason an agent isn't already shopping your book is a lack of monetary incentive. The monetary incentive is already there, which is why agents sign books they think they will sell and sign clients they want to work with career-wise.
If you had a pay to play system in place, what would happen is that an awful lot of people would shell out money and never see a return. Which good agents know, and that's why only shysters do it. Having an agent isn't going to automatically open doors for you; editors only buy 1-3% of agented submissions landing on their desks. It's still tough, even with rep.
#
Longer answer, on why the process matters:
Learning to query isn't just about going through pointless motions of formality. Learning to write and pitch books is actually a microcosm of the whole business, and this is why querying is so difficult: you are learning *publishing* the industry. It is an absolute mountain to climb.
Writers who have mastered query letters have learnt to assess the commercial potential of their own art, how to capitalise and sell it in a business sense, and how to gauge when a book is ready. They have made the (wobbly, baby steps) leap from hobbyist to aspiring career writer.
I get the frustration of writers who are beating their head against the door but in most cases (exceptions always apply) a working manuscript + working query will yield requests, even if not offers, and that's generally how you know you're writing to standard (the rest is timing, personal taste, market trends, ie luck; that can only be controlled for through persistence.)