It seems simple enough – a high profile campaigner gets some publicity and starts receiving some pretty horrible abuse online as a result. Twitter should therefore install a magic “report abuse” button that will make all this go away.

As with many simple solutions, it’s also wrong. And unfortunately not just slightly wrong but dangerously so for those most in need of protection.

The scale of the problem

What’s being asked for is a single button to report an abusive tweet, rather than the existing web form. However, one of the issues in this case was complaints that Twitter took too long to respond. I’m not clear on how having a button rather than a form that will inevitably attract more complaints, because it’s easier to click is supposed to speed things up.

There are over quarter of a million tweets per minute. If one in a hundred thousand tweets are abusive enough that someone clicks a button, that’s about three tweets every minute that need looking at.

If you need a quick response, you need someone at their desk 24 hours a day, 7 days a week. To allow for training, holidays and sick cover you need five full time employees for one person on shift at all times. Finally, assume that twitter are immensely generous by service provider standards and will dedicate 5% of their staff (20 people) to just handling abuse reported via this button. Four people on shift at any one time, dealing with three tweets a minute. Each person is having to evaluate an abusive tweet and come up with an appropriate response in 80 seconds if they do nothing else for their entire shift on duty, with generous assumptions about staffing levels and number of reported abusive tweets.

Oh, I’ve conveniently ignored all the problems associated with differing cultures and languages in use worldwide.

Unless you can somehow reduce the volume of complaints massively – to below one in a million tweets, giving a “generous” 5 minutes to assess each complaint with a very well funded department – this just doesn’t work.

Some sort of automated filtering is required, perhaps looking at volume of complaints as a threshold for action.

…and then it starts to backfire…

We’ve seen this before. A large portion of my timeline on twitter is taken up by members of marginalised communities that are subject to fairly routine, day to day abuse and know what works and what does not. People are quite vocal in their opposition to the twitter abuse button idea, because we know how it will pan out in practice.

The original article has Caroline Criado-Perez expressing concern about “victims without a high profile” getting help, whilst apparently ignoring those same victims saying this is a bad idea.

If you are a high profile individual with contacts, you can get lots of your followers to complain about tweets directed at you. This will attract enough attention that you get some response – but you could have just gone to the police in the first place if you have that much influence and it’s actionable abuse. (I have complained about death threats online in the past and getting much more than a crime reference number and a short appeasement visit from your local bobby is quite some achievement)

But the flip side of this is that it’s not just the good guys that can raise large numbers of reports. Any group seeking to silence an individual or community can just as easily rally the troops to try to get people kicked off. Either adopt a scatter-gun approach, complaining about anything and everything, or find one tweet that, taken out of context by someone with only a few minutes to assess the situation and take action, might be considered abusive.

Potentially, small groups or even one determined individual can create enough accounts to appear to amount to a significant number of unique complaints. In this age of near-ubiquitous free webmail, WiFi hotspots, 3G internet access and cyber cafes it’s impossible to conclusively tie multiple online identities together quickly without access to the kinds of legal resources only available to major police investigations.

How are genuine victims without a high profile supposed to generate enough noise to get any useful response, often to tweets that individually amount to little but constitute a clear pattern of abuse if you know the background? About the only option is to publicise the matter to try to make it high profile. But if you try this without already being popular, you’ll likely be accused of being one of those people engaging in the exact same silencing behavior I’ve just described. Which is what happened with the Burchill/Moore saga.

Unsurprisingly, Suzanne Moore is one of those backing the Twitter abuse button idea.