I think you'd have to select your readers carefully and you'd have to have limits either on volume or on performance. If you read a lot, then you're familiar with different quality of work and you know whether something is good or not, whether you agree with it or not. The question to ask you as a reader... if you have an article a day to recommend, would you recommend the one you've read?

But here's where I want engineering folks like your good self to test the idea further. If your one recommend is for a hidden piece that is so good that everyone loves it - then you may only make one recommendation a month. That recommendation however, would net you more in that month than the other 99 chosen readers who aren't as discerning as you are and who pick mediocre pieces which fizzle out. 10% of awesome needs to be better than 10x10% of mediocre.... that way you'd start to see a shift towards better writing.

I'd also put a performance limit in. If you start doing junk referrals and your nominations for boost drop below a limit... say 50%. If you can't pick a good story better than a flip of a coin, then you get removed from the programme for six months. Though if it were up to me I'd have people head-hunted for readership-expert roles rather than putting themselves forwards.