How can we detect (and deal with) spam?

Question

The premise of Hardware Recommendations is that Person A is looking for one or more product(s) that will satisfy certain criteria or complete a certain task. Persons B, C, D, etc. may write answers giving the details of products that can help Person A.

The problem is that there is a bit of a hole here where spam can get in. If Person E is affiliated with the company that makes a certain product that might be related to what Person A wants, s/he can post it in an answer and get attention for that product. If there is enough similarity that it could be a valid answer, the answer will not be deleted, and Person E gets what s/he wants. The cycle can repeat.

Person E can also self-answer, for added efficiency and a greater likelihood of putting an answer that matches the question well, thereby decreasing the risk of deletion.

How can we detect and deal with spam? The first part may well be impossible, but the second is important. Going forward, it could be an issue.

Answer 1

On Software Recommendations, as a moderator, this is my approach: when I see something low-quality, I check it, and the user, for spam. Also, if I see a user promoting the same product or brand repeatedly, I'll check for spammyness.

If I see signs of association, I'll either destroy outright or, if they show signs of becoming a good contributor, I'll send them a message and suspend.

We can't detect all instances of undisclosed self-promotion. But that's okay: we don't need to. If we stick to a simple policy of deleting low-quality answers, every time, then we only have high quality, informative answers left. Also, we check for users promoting the same thing over and over and take care of those.

Now, if we only have informative posts, voting will take care of the rest. The spammed product doesn't work? It'll get downvoted into oblivion and won't really be a good advertisement. But if it does do the job... why should we care? That's exactly what we want.

Moderators, and users, on this site will get very good at detecting spam here.

Answer 2

Spam will be hard to catch on this site, even by spam filters (unless it's some Korean spam, of course :P). I think that we will mostly have to catch spam by hand. This means we will have to have a dedicated user base, who is willing to try to find spam. There are ways we can help make finding and reporting spam easier, though.

The hardest part will actually be detecting the spam. If you see a post that looks remotely spammy:

• Check the user who posted's profile - you may find hints of affiliation between the user and the product they recommend in their profile. If their profile picture is the logo of the product that are promoting, if they have a link to the site of the product in their profile, or if their Twitter account is affiliated with the product, then they are most likely a spammer, and you can follow the steps below

• Check other posts from the user - if other posts from the user are about the same product or products from the same company, the user is most likely a spammer. Again, follow the steps below

Here are some guidelines you can follow if you ever find spam on this site:

• Flag it as spam - this should be the first thing you do, as it will put a flag in the moderator flag queue, so if a moderator is online, they can make it go kaboom! Also, if enough users flag it as spam, it will be deleted

• Report the spam - luckily, SE is a wonderful place with wonderful chatbots. One of these is Smoke Detector who flags spam and rude posts from around SE. You can report spam to Smoke Detector by going to either the Charcoal HQ or the Tavern on the Meta and report the spam using !!/report [spam post link]. If you are not able to do that (because you don't have the privilege to use !!/report, you can ask someone else in one of those rooms to report it for you. If you see certain link being posted in a lot of spam posts, you can also ask someone to blacklist that link