Doesn't seem like a terrible deal for Amazon. Amazon effectively paid $0.036 per click, which in the grand scheme of things is not that much. It's a premium on top of what Reddit charged ($0.02/click), but then Amazon didn't have to pay any humans to write and set up those ads, nor research where to put the ads. Basically they paid a premium to have a third-party figure out the advertising for them.
Plus, the associates program has other benefits for Amazon. As the article points out, authors use affiliate links to promote their books for sale on Amazon so they can get a higher cut of the profits. Seems like that scheme, not even limited to books, would motiviate a lot of people to sell their wares on Amazon.
And then there's the benefit of squeezing out other affiliate link programs. An ad can only have one affiliate link, so by making the Amazon program attractive they out-compete other stores in the ads with affiliate links marketplace.
Great point regarding providing motivation for people to sell their wares on Amazon. I did not think about that.
This especially makes sense to me because the majority of Amazon's sales are filled by third-party merchants, including nearly 100% of their sales when they launch new product categories.
It's a bad deal for Amazon if, as the author assumes, the people aren't buying the product featured in the ad, but rather just happen to buy something else from Amazon within the 24 hour window without any connection to having seen the ad.
I wouldn't be surprised if in future amazon changes the system so that it only pays the fees when the original product from the link is purchased (or perhaps switches to this only for prime users.)
It is likely Amazon advertised several other products on the landing page for that Zimbabwe currency, so that no doubt increased the effectiveness of these ads for Amazon and might have made that referer fee worth it.
I've been in the Amazon Associates program for several years - at one point, I had an LLC set up specifically for it, because my state at that time was having a bit of a legal battle with them.
My job is to get people on Amazon's site. Why they're on the site almost doesn't matter, because my job ends when the user's eyeballs hit Amazon's page.
Amazon's job is to convert eyeballs on their pages into sales - and they're really good at their job, as is evidenced by their position in the market.
If you add an item to a wishlist or Amazon adds you to a retargeting list, they can still get a sale at a much later time. Most affiliate programs are much stingier in this regard, and it makes it harder to justify participating.
The arbitrage is where you purchase clicks for less money than you gain in commissions. It can be fairly lucrative if executed well.
The problem in the long run is that you're constantly working to find the next thing before the current tactic gets saturated. For example, the ad in the article is no longer possible on Reddit and most sources of cheap clicks to Amazon will get you banned from the program (they have stricter guidelines than you think for traffic sources)
Unless you stay ahead of the pack, your avenues for cash flow will rapidly close most of the time.
Pay $0.02 per click, send that click to Amazon, earn $0.04 per click. Earn $0.02 per click margin for funneling cheaply acquired traffic to an entity that is willing and able to pay more for that click.
The cost per click is dependent on an ad with a high CTR, in addition to a low cost per thousand impressions. Reddit ads, as mentioned in the post for example, are one place with low priced ad inventory.
You pay reddit per impression, not per click. Getting an ad that works out to $0.02 per click is not as easy as it sounds, especially if you're just trying to make them buy something on Amazon. Clickbait doesn't lead to sales very much, and advertising to capture attention of people who'd actually buy it leads to a very high cost per click.
Amazon does not pay associates per click. You are essentially purchasing sales leads from reddit. This is not arbitrage. Buying lottery tickets from a gas station and bringing the winners to the bodega next door is also not arbitrage.
I built an app to help make Amazon and iTunes affiliate links from iOS, because I was playing with some similar schemes. Might be useful to others interested in this. The app is currently free, though I'll likely add an in app purchase of some kind soon.
I've thought about that, but I haven't figured out how to clearly communicate what's happening and make sure it's not percieved as scummy/untrustworthy behavior.
This seems similar to cookie stuffing, which is prohibited by amazon. The idea is to send lots of low quality traffic to skim commissions from the session
A major problem is that amazon has a 2-month lag in payments (feb payment are made in late april), so it would really suck to spend $1000's on ads and not get paid.
It's not really cookie stuffing. In the end, people have to click on the ad to get the cookie.
Cookie stuffing would be, for example, HN embedding a hidden iframe with their Amazon code and placing cookies on users.
I did something similar a couple of years ago with a _very_ cheap traffic source I found. I bought traffic and sent it straight to the Amazon homepage. I had a good run for about 4-5 months. On my last month I really ramped up traffic, spent like $250 and made around $6k in a single month. But amazon caught on and like you said, cancelled my account and didn't pay me.
It's a sucky thing to do I guess. I know of massive operations where guys are pulling in $10-20k per month. If someone wanted to really put their brains in to it, cookie stuffing is very hard to detect.
If you have access to cheap traffic that's not complete garbage, giving it to Amazon is one of the dumbest things you can do.
I really tried AA on a few of my sites, and eventually gave up and switched back to Adsense. I even created a custom ad system which tried different Amazon products, tracked impressions and clicks, which gave me the data on which ads were more likely to be clicked on. I went up from the initial 0.5% CTR to 5% CTR, which I felt really proud of. But then you have to realize that once these people land on Amazon, they have to convert again in order for you to get paid, and the conversion rate is like 1-2%.
So you end up with shit like this: 250 clicks, 4 items ordered, $2.40 earned (straight from the earning report).
Now think how much Adsense would pay you for 250 clicks. I average $0.432 per click (just checked). So $108 is what I'd expect to earn with Google. That's 45 times more than Amazon.
I run several sites, one of which is mostly gun reviews. I know my audience on that site very well, and I can hand-select Amazon products that will sell.
On some pages, I have ads for things like a P-38 can opener. They're around $3 each or $10 for 10. Something like 5% of my traffic clicks that ad, and 70% buys the product. Because it's so cheap, I make only a couple of pennies per sale... but the number of items sold bumps me from the default 4% commission for the month to 6.5% or more without working for it.
On other pages I might link to a $400 rifle optic, or a $50 gunsmithing book. Because of all those little sales, I make $26 on that optic instead of $16 - it makes a big difference.
I've put Adsense on those pages in the past, and it averages around 10% of the revenue I get from Amazon.
It really depends on what you are promoting. Niche, high priced products do well. I have a niche site that earns $80+ per sale averaging 30 sales per month. Adsense on that same site earns less than $200/mo.
I do not believe this is a valid comparison. The clicks are not happening on my website, they are happening on Reddit. I am paying for the clicks.
If I were to use Adsense in my experiment, I would've most likely earned between $10-20 from 4,800 pageviews, around $2-3eCPM. By comparison, I earned about $175 with Amazon Associates from 4,800 clicks. As LyndsySimon commented, Adsense would earn only about 10% of my earning potential with Amazon.
That's only because Reddit let you post direct affiliate ads, which they don't anymore. I'm surprised Reddit adops don't do that kind of arbitrage themselves, it's like free money.
If we were to dive in further, we can also expand on the ad-commission arbitrage and go ad-ad arbitrage. By finding price differences between ad exchanges, one can find arbitrage opportunities. Granted this would apply to display inventory on a CPM basis. Here is an article that sheds some light: http://www.bloomberg.com/news/articles/2014-11-07/high-speed...
I have an Amazon Associates website and I've considered trying to use ads to generate traffic. This adds another step in the process vs what the article describes. Users have to click on the ad, click on my affiliate link, then make a purchase.
Users of my site click on an affiliate link about 40% of the time and those that do make a purchase about 2.5% of the time with an average commission of $5. This means I make about $0.05 per visitor so I would need ads with a CPC less than that. In my niche this does not seem remotely possible.
Luckily I'm getting over 1,200 visitors per day without paying for them.
To anyone who thinks this is a great idea, read and understand the Amazon Associates Operating Agreement https://affiliate-program.amazon.com/gp/associates/agreement and the Linking Requirements. Buying search ads and using redircting links may lead to account closure. Usually ad links are redircted and even redirects within your own domain for tracking purposes can cause trouble. When in doubt, ask the associates team before implementing questionable advice you found on the internet.
32 comments
[ 4.7 ms ] story [ 75.7 ms ] threadPlus, the associates program has other benefits for Amazon. As the article points out, authors use affiliate links to promote their books for sale on Amazon so they can get a higher cut of the profits. Seems like that scheme, not even limited to books, would motiviate a lot of people to sell their wares on Amazon.
And then there's the benefit of squeezing out other affiliate link programs. An ad can only have one affiliate link, so by making the Amazon program attractive they out-compete other stores in the ads with affiliate links marketplace.
This especially makes sense to me because the majority of Amazon's sales are filled by third-party merchants, including nearly 100% of their sales when they launch new product categories.
I wouldn't be surprised if in future amazon changes the system so that it only pays the fees when the original product from the link is purchased (or perhaps switches to this only for prime users.)
I've been in the Amazon Associates program for several years - at one point, I had an LLC set up specifically for it, because my state at that time was having a bit of a legal battle with them.
My job is to get people on Amazon's site. Why they're on the site almost doesn't matter, because my job ends when the user's eyeballs hit Amazon's page.
Amazon's job is to convert eyeballs on their pages into sales - and they're really good at their job, as is evidenced by their position in the market.
The problem in the long run is that you're constantly working to find the next thing before the current tactic gets saturated. For example, the ad in the article is no longer possible on Reddit and most sources of cheap clicks to Amazon will get you banned from the program (they have stricter guidelines than you think for traffic sources)
Unless you stay ahead of the pack, your avenues for cash flow will rapidly close most of the time.
https://geo.itunes.apple.com/ca/app/rivet-effortless-affilia... http://rivet.link
A major problem is that amazon has a 2-month lag in payments (feb payment are made in late april), so it would really suck to spend $1000's on ads and not get paid.
That's probably to dissuade people from ordering items then doing refunds/charge-backs after collecting commission on their orders.
Cookie stuffing would be, for example, HN embedding a hidden iframe with their Amazon code and placing cookies on users.
I did something similar a couple of years ago with a _very_ cheap traffic source I found. I bought traffic and sent it straight to the Amazon homepage. I had a good run for about 4-5 months. On my last month I really ramped up traffic, spent like $250 and made around $6k in a single month. But amazon caught on and like you said, cancelled my account and didn't pay me.
It's a sucky thing to do I guess. I know of massive operations where guys are pulling in $10-20k per month. If someone wanted to really put their brains in to it, cookie stuffing is very hard to detect.
I really tried AA on a few of my sites, and eventually gave up and switched back to Adsense. I even created a custom ad system which tried different Amazon products, tracked impressions and clicks, which gave me the data on which ads were more likely to be clicked on. I went up from the initial 0.5% CTR to 5% CTR, which I felt really proud of. But then you have to realize that once these people land on Amazon, they have to convert again in order for you to get paid, and the conversion rate is like 1-2%.
So you end up with shit like this: 250 clicks, 4 items ordered, $2.40 earned (straight from the earning report).
Now think how much Adsense would pay you for 250 clicks. I average $0.432 per click (just checked). So $108 is what I'd expect to earn with Google. That's 45 times more than Amazon.
On some pages, I have ads for things like a P-38 can opener. They're around $3 each or $10 for 10. Something like 5% of my traffic clicks that ad, and 70% buys the product. Because it's so cheap, I make only a couple of pennies per sale... but the number of items sold bumps me from the default 4% commission for the month to 6.5% or more without working for it.
On other pages I might link to a $400 rifle optic, or a $50 gunsmithing book. Because of all those little sales, I make $26 on that optic instead of $16 - it makes a big difference.
I've put Adsense on those pages in the past, and it averages around 10% of the revenue I get from Amazon.
I targeted the products to my audience too, and still didn't get above 2%.
If I were to use Adsense in my experiment, I would've most likely earned between $10-20 from 4,800 pageviews, around $2-3eCPM. By comparison, I earned about $175 with Amazon Associates from 4,800 clicks. As LyndsySimon commented, Adsense would earn only about 10% of my earning potential with Amazon.
Users of my site click on an affiliate link about 40% of the time and those that do make a purchase about 2.5% of the time with an average commission of $5. This means I make about $0.05 per visitor so I would need ads with a CPC less than that. In my niche this does not seem remotely possible.
Luckily I'm getting over 1,200 visitors per day without paying for them.