It's About Ethics

Let me tell you a little story about ethics. One that doesn't involve games journalism (because seriously, if complaining about that is your bag, you suffer from a serious inability to see the forest through the trees).

The forum I built and operate is home to a small and tight-knit community of friends. It's funded by the affiliate marketing revenue I generate on Amazon Associates. I show product ads in the sidebar, and the users' Amazon purchases result in a small percentage for me.

Where do I get the ads? Well, that's where ethics come in.

The Amazon Associates system records every purchase made under a referral code. The data is right there for the taking: Product X is selling like hotcakes, Product Y is popular during November and December, Product Z was only purchased once but has a higher commission, and so on.

What's to stop me from using that data to funnel my users into buying the products that are popular with the other users on the site?

Empathy. (Hah! You thought I was going to say ethics!)

You see, when the user clicks through to Amazon and completes an order, I earn commission on the entire order. As a result, all the purchased items show up in the earnings report - not just the item that was clicked to from the forum.

Amazon sells EVERYTHING. People can buy all kinds of things, from milk to personal items all the way to uranium (yes, seriously). Although I can't see specifically who purchased which item, guesses can certainly be made, especially when everyone knows each other so well. In light of this, I force myself to operate under this assumption:

In a small community, if one person bought something potentially embarrassing, they might be MORTIFIED if it showed up for the whole site to see.

For this reason, I only rarely look at the full earnings report, and I have no plans to ever feed the data back into the ad selection tool.

So, with that avenue ruled out, where can the ads come from? Other sites, where user interaction is limited, might benefit from Amazon's other options, such as pre-defined offers or "aStore" pages based on Listmania wish-lists.

Luckily, because my site is an active forum, the users themselves tend to post Amazon links for products they find interesting and want to share. Rather than mining my users' raw purchasing data for computational recommendations, I can operate on a second assumption:

If the user chooses to share it, they accept that other users will see it.

It turns out that that's not the whole story, though. A while back there was an incident where a user was goofing around and posted a product link for a speculum. More than a few users were embarrassed at the thought (if you don't know what a speculum is, you'll probably understand why once you look it up). It's also fair to say that zero - or nearly zero - users would be interested in buying one. Luckily, I had anticipated an issue such as this.

When dealing with user-supplied ads, there needs to be some kind of feedback process to give bad ads the boot. I set mine up so only two down-votes block the ad from the rotation. This is the third assumption:

The users must be able to protect themselves from ads they don't want to see.

Three assumptions. Driven by empathy. Guiding my choices.

Nothing - and no one - is perfect, mistakes can happen, and one solution doesn't fit everyone. These are just the guidelines that have worked for me so far.

It's tempting to mine the raw purchasing data, because it's just right there and reflects the tastes and preferences and interests of the users - but in my opinion, doing so would be a violation of my ethical responsibility to the community.

How do you approach your own ethical quandaries?


Web Application Developer, Amateur Photographer, Blogger, Recovering Incompletionist.

comments powered by Disqus