I start with quality. I'll take a quality product or service and a quality sales letter over just about everything else. You can have this huge affiliate payout, but that doesn't matter is the product or service is crap.
If you're going to get all of these refunds, it just isn't worth it. A sales page that converts is something that can't be understated. I don't care how good the product or offer is, they must have a sales page that converts. I'm not going to waste my time sending traffic to a sales page that doesn't convert.
Another major factor that comes into the equation is how much money that I think that I have to spend promoting it. It isn't about the amount that I will spend, it is about the amount that I plan on gaining, the net.
I don't care if I'm making $100 a sale or $5 a sale. It is all about the net gain and how many times that I can replicate that net gain. I'll give you an example of what I'm talking about.
Say that I'm using pay-per click traffic and I'm using landing pages to get to the affiliate sales page. If the payout is $100 and it costs me $90 to get a sale, I'm only making a $10 net. If the payout is only $25 and I pay $10 to get a sale, the $25 payout is more fruitful.
There are other things that would go into that equation. I may be able to scale up the $100 payout campaign and it doesn't cost me much more and I can make a lot more sales. The $25 payout may cost me a lot more to scale that campaign up and it may not be as fruitful as the $100 payout, in the end. It is all about testing and variables.
The last thing that goes into picking an affiliate program is my infrastructure in that niche. What I mean by that is how many assets that I already have in that niche market. If I already have a mailing list in that niche, it is pretty obvious. It also helps if I have a website or blog in that niche and if I have promoted a product or service in that niche before. So in a nutshell selecting the right program is not that easy even if the gurus have told you otherwise!