Welcome to WebmasterWorld Guest from 188.8.131.52
If your only concern is with conversion, why not test it? Every site and every site's visitors are different, after all.
As an added bonus, you get an extra 2.5% for direct links, meaning if you send them to a product and they buy THAT particular product, you get an extra 2.5% referral fee.
Put those two together, and I think you end up making more money.
However, if you're shopping for some other type of item -- electronics, for instance -- you're probably going to want to compare many different items, and in that case Amazon's default landing page might be best.
We haven't done any research on how using "ref=nosim" impacts the direct-link premium, but in previous tests linking directly to the detail page using ref=nosim has had a negative impact on conversion as compared to the Similarities Explorer page that shows the featured product and other related products.
Myself, I rarely use 'ref=nosim' because even though I have a program that checks AWS and alerts me if there is any change in the 'Availabilty' status, I'd rather them see something to buy rather than nothing when the direct item is out of stock, etc.
I've toyed with the idea of having the program email me, but instead I just run on demand since I'm not going to leave a party or something to change a link!
I get the impression, the landing page inspires a lot of customers to pick up another item or two also. Unfortunately, I can't prove that unless I use a massive amount of tracking ids. Many have implored Amazon to come up with a report that shows what items constituted a given order.
Here is a sample URL for a link -- insert the item's actual ISBN and your actual Associate ID:
You would use this link if you had worked to convince your visitor that they should buy a particular item, and you don't want to have them distracted or confused by having a page full of allegedly similar items thrown up at them.
If your visitor buys an item after clicking a direct link to it you can get a bonus--ref=nosim may increase the chance of getting the bonus by increasing the chance that they will buy that item and not some other.
I would not take the Amazon statement as being very useful. They are talking about averages, and I would think there are Associates who use ref=nosim more effectlvely than others.
Also, purely aside from the question of whether or not it brings you more money, you may like to use it to enhance your visitor's experience. On many of my pages, when someone clicks on an Amazon link, they expect and want a specific item, not a bunch of similar stuff.
One final point: don't overuse it. If you've got a lot of links in a list, and you're not providing much other info., I don't see any point to using it. Use it only if you're proving real information about something....