Ask HN: Dealing with abusive users

We have this on a daily basis.

We deal with it like this:

- first we write the email that we would like to write telling them exactly how we feel

- we then delete this email

- we then write an overly friendly email which takes care of the users complaint and adds sugar on top as well as a pony.

Usually they get the message and they'll apologize for their behaviour

One exception is made for people that threaten legal action, these are without exception barred from the service and requested to follow through. I really can't stand it when people pull out their 'lawyer gun' at the first opportunity, and it raises an immediate red flag, these are real trouble makers, and will continue to be so in the future, so we see no reason to have them in our userbase.

To date nobody has ever followed through on our invitation to press suit.

I really wonder what is wrong with those types, by the time I threaten to sick a lawyer on to someone I'm 100% prepared to go through with it, and I reserve that for those times where such action is really warranted, such as outright fraud.

Wackos are a fact of life, as the profile of your site increases you'll have to deal with some of that anyway, better be prepared and use your support queue as the selector determining which 'wackos' are only temporarily deranged and which are really firing on 3 or less pistons. Those you can do without, even if they try to make more trouble later.

A nice example from HN is the user 'arrington', a complete douchebag that made it his mission to destroy HN after being kicked off. It was 'before my time', but I've seen some of the fallout from it and it wasn't pretty, but it did go away after some hacks that made the site more robust against that sort of thing. What doesn't kill you makes you stronger!

Thanks. I've been following some of this policy, but a little altered. I make it clear that they didn't reach some customer service drone, but instead the guy who runs the site. I try to solve the problem, but lay down the law at the same time making it clear that we don't allow any crap in our community.

I also make a distinction between panic and rage. Panicked users will say, "It's broken! Why won't you let me log in!" while the enraged ones will just toss out a string of profanity.

We've never had anybody threaten legal action, but it's good to know to call them on it. I'll remember that.

I had my first lawyer gun threat today, actually! I feel so special. This is a user that repeatedly broke terms of use.

I told him to grow up and get off my lawn.

Thanks for providing ready to use introduction for user support staff manual.

By the way, I can corroborate the 'assholes are never paying users' statistic, it's the freeloaders that seem to engender the jackasses. Paying members are almost always articulate and somewhat forgiving of errors on our end.

I have a personal theory about this. I think it is a personality type.

I had a friend who later in life became a free loader (and drug addict). I tried to help him and his family out by letting them stay a while while he tried to kick his habit.

Anyway long story short, from the moment they walked through the door, there was an attitude of entitlement because they where being victimized (in their mind).

When you confronted them about things, they would become very hostile like you where the one victimizing them.

I believe that this is a learned behavior that a certain segment of the population uses as a survival mechanism. It preys on the fact that a larger portion of society likes to avoid conflict and therefore will seek to sooth the situation.

The worst part about my personal situation was that his kids where also learning this behavior, and where actively being used against me by playing up the putting kids on the street argument. Which I eventually had to do for the sake my own family.

Anyway, I think the non-paying complaining assholes are a lesser form of this personality type.

If they're paying, they're deriving value from your service. If they're deriving value, they'll probably come to rely on that value. And thus, if it stops working, they won't want to piss you off so badly that you take that value away. Game theory, really.

All of our abusive ones are forwarded to me; I write sugar sweet emails and try to fix their problem like any other support email. But sign it like this:

Tom

XXX Product Manager / Team Lead

P.S. Please take a quick look at this friendly FAQ <link> on how we suggest users contact staff members. We try our best to answer queries and fix problems as fast as possible; however, I inform staff they have a right not to suffer verbal abuse, which is why I have personally dealt with your email"

Usually you get a really bashful apology (and anyone that follows up the abuse gets a terse "you had a last chance, sorry" email and a perma ban :))

i.e. I know people get pissed and write emails they later regret (everyone does it) - but twice uses up their credit :)

If your main worry is that they will start trolling/flamming your community, you could try http://www.urbandictionary.com/define.php?term=hell%20ban

At some point, they will realize that this is happening, and there will be outrage when they communicate about it to their friends via an outside channel. Be prepared for a lot of splashback.

Make it a 85% hell ban - they will never figure out what happened, yet they will not be able to continue flaming.

I've thought about this, but it basically means I'm programming features for assholes. I don't want to get in a war like that, but I may have to, I guess.

HN is full of programming features for assholes, it seems to have served the community well, even if it isn't always 100% transparent and in the case of false positives it requires manual undoing.

That sounds interesting. Are those features in the public source code?

No, PG keeps a lot of the secret sauce out of the public release in order to make it a bit harder for those that want to game the system.

I love this concept; an alteration would be to instead change their background to care bears or barney, or something overly obnoxiously loving.

Hmm. Has anyone tried a purgatory ban (I just made that up). Basically give individual users the ability to never see posts or messages from other individual users, but no indication to the ban-ee.

I've done that. They figured it out and twas no fun. Don't fight with pigs, you'll just et dirty and they like it (or how does that saying go?)

I personally wouldn't ban them unless they were actively making it difficult and unenjoyable for other users. I don't mind laughing a bit at a disparaging message, but I have a thicker skin than most. However, what I wouldn't mind doing for the really abusive is grabbing a beer and cooking up something that makes them feel abashed. A personal homepage for their account with something like 'do unto others as you would have them do unto you' plastered across the top. Man, maybe even make the banner randomly draw from a list of trite sayings...'Consideration for others is an important virtue'.

We get a few rough emails/suggestions come through, but haven't had to deal with anything profane. Usually it's something like "This site is terrible, it doesn't work worth crap and I can't believe you have any users". It's almost always something really hilarious that causes it (javascript being turned off, they are using IE6). I respond with an explanation, query for more and try to help. Everytime the person turns their attitude around and says thanks, that we have good service. I am sure that will change when our userbase grows, but it is working well so far.

On a sort of related note, I did have a user put in a name of F YOU (but fully spelled out, yes all in caps). Best part is they were an active and seemingly happy member. I am not for censorship, however, we are currently a small site, so his name might come up semi-often for other users. Enough to drive down the value for other users, which is when I draw the line. I decided to change it to LOVE YOU. It's still there, even though the person has used the site many times since I made the change. They must have a good sense of humour :)

I have a forum that allows anonymous posts. I have a list of banned IPs, sets of IPs that must register to post, various banned words (spam, mostly), banned user agent strings, etc but there are still griefers that get a new IP and return. Any tips on how to block them more aggressively?

I run a little script that tails the httpd log and that figures out which IPs show 'unrealistic' user activity.

Unrealistic are bots (high frequency requests without interaction, as well as requesting links that are invisible to normal users) and unrealistic is a new IP that immediately wants to sign up without seeing some pages first.

Use flash cookies. Set them to recreate themselves to sync when the user clears their browser cookies. That has proven tremendously effective for me. Flash cookies get a bad rap -- but they're a very effective way to keep people out.

Is your code open source? I googled about this once, but couldn't find anything. I posted on stackoverflow and I was instructed to build one myself using flex. Fair enough, but I never did.

We've always dealt with abusive users in a polite manner. A lot of times they just love the attention and drama.

But time and time again, once we fed them the attention they craved, they became some of our biggest allies in the community.

I really strongly disagree with all the suggestions that involve punking, shaming, or otherwise battling the offenders. Like I said, they are never paying members. It's not my job to teach them anything, and they're not going to learn it anyway.

If I spend time punking worthless trolls, it's not time spent enhancing the site for the users I do care about. It can be fun to get down in the mud and fight with someone, but considering that they will never become a paying subscriber, it's really just a waste of my time.

Instead, I just want to minimize the amount of trouble they cause.

You may want to experiment with adding an "Emotional State" option to your support form following the Wufoo approach. They say that a large number of people fill it out and the byproduct is that people use calmer language because they are conveying how they feel through the dropdown instead of having to do it in the text.

edit: Here's the link to their blog post: http://particletree.com/features/on-asking-users-for-their-f...

Patrick's (patio11) advice in dealing with such frustrations from long back:

http://news.ycombinator.com/item?id=666171

Not Always Right (http://notalwaysright.com/) has shown me that some customers cost you more than they'll make you, ans has convinced that keeping users who consistently waste time and resources is not the best strategy.

In addition to already suggested hellban I suggest exponential back-off in response time as a way of dealing with anyone you don't like. They will either calm down or leave in frustration.