To protect our users from spam and other malicious activity, this account is temporarily locked. Please log in to https://twitter.com to unlock your account

[twiz718]

My tweets go out every 15 minutes during working hours, no more than 5 at a time. All do have links and hashtags. Content is of similar nature, for example, one bot is all about NBA news, another is about MLB news, etc. If we’re breaking some rules, it would be nice to know about them even post mortem.

[cv_tmj]

I agree. i was wondering if we happen to be using a hashtag that was trending at a certain point. They seem to want to protect against automated usage of trending tags. I have no evidence this is the case, just a thought.

[gavinwatson]

I’ve just had one blocked for the third time, this particular one tweets every 30 minutes, hardly breaking any limits.

I’ve seen a very large number of our customers have this same issue with no changes on our platform to account for this sudden issue.

At the moment I cannot see a pattern unfortunately. Some of the accounts post every 5 minutes, some every 15, 30 60 minutes but no pattern so far.

I’ve reported it twice to Twitter now looking for an answer, but no official response so far.

Please report back with any findings so that we can share it and look, where possible, to work with any change that they may have made.

Thanks.

[cv_tmj]

Will do. I am leaning more toward content over a certain amount of time or number of tweets being the issue rather than the rate at which we are tweeting. They already have rate limits in place for that and the rate for these “spam locked” accounts is not super frequent. The content is a tough one to diagnose. It could be some overall similarity algorithm or maybe it has more to do with hashtags, either specific hashtags or repetition of the same hashtags in a set of recent tweets. Or it could be tweets with links, maybe if every tweet has a link. The trouble is we have accounts that have the same tweet behavior that have not been locked. But then I don’t know if they just haven’t been evaluated yet. Very frustrating. They made an api and apps so that developers can do what we do. We are not spamming (or at least don’t think we are) and it is very bad for businesses that rely on this to work.

[twiz718]

I messaged (through LinkedIn) an ex co-worker of mine who works at Twitter right now in hopes he may be able to direct this thread to the right person internally. crosses fingers

[cv_tmj]

Awesome!

[gavinwatson]

Interesting, I note that this thread has been moved to ‘Rules and Policies’ from it’s original REST API location. Someone at Twitter (or I guess a moderator) must have seen it already, hopefully we can get a reply.

One thing that surprises me is that, if this is a widespread issue, there are not more people reporting it here?

We’ve now seen a large number of accounts locked, all the same reason, and they’re being locked again, so it isn’t a ‘one-off’ thing that has happened.

I’m still awaiting a reply from Twitter. Will post back if I hear anything.

[andypiper]

There is a chance that one of our safety algorithms was tweaked if there’s a specific change in behaviour here. For example, this might happen if the links being shared are detected as potentially suspicious, or if a lot of accounts appear to be Tweeting similar or identical content. Our antispam system (“Botmaker”) is constantly evolving so it is not possible to provide advance warning for these kinds of situations.

I appreciate that, as developers, this affects your applications; but unfortunately this is an account support issue rather than something we can help you with on the developer platform side, so the correct route is to report this via the account support forms. For URLs that are marked as unsafe or spam, there is guidance here. For suspended applications, use the platform support forms.

Thank you.

[twiz718]

FYI, I found this based on Andy’s comments:

Fighting spam with BotMaker

I tweeted at Raghav asking him to look at this thread, he may be able to provide more information on what we can do.

[gavinwatson]

@andypiper Thank you for your comment here, it is much appreciated.

My particular issue is that no-one from Twitter is responding to tell me what has changed or how to inform Twitter that the behaviour you are flagging as malicious is not. We’re talking hundreds of our users having their accounts locked all of a sudden How can I engage with someone at Twitter to find out what is happening? I’d like to have the opportunity to find out what is going on so that I can guide my users accordingly.

Currently users are changing their password, which works, but the accounts are being locked again straight away with no explanation as to why except for the message in the title.

I understand that you say that this is for account support, but ultimately something has changed on Twitters platform but we do not know what or why, or how to now comply with these changes…
Do you have any further advice on how to get a response or how it would be possible to submit any information with a view to getting a better understanding of what is happening?

Thank you.

[michichan]

I have lost several applications because of this blocking so yes, I consider this very much an API /Developer issue because if you all are encouraging us to develop on the Twitter platform which provides value added services to Twitter and gives people a reason to look at Twitter (including the advertising and other sponsored tweets) but then when we come up with innovative services to enhance the Twitter experience for YOUR users who CHOOSE to follow our tweets, we get shut out as spam bots. I guess those of us who develop content for Twitter are not that important. This issue needs to be escalated but I am glad I was not the only one experiencing this.

[twiz718]

All of my bots are currently blocked again, it keeps happening… I change their password and maybe 24 hours later they are again suspended. This new algorithm is breaking our applications which are not spamming anyone. Please escalate this @andypiper to someone internally. We need help!

[gavinwatson]

Another day passes and another bunch of customers accounts locked due to ‘malicious activity’ and no response from Twitter.

I really don’t know how to get Twitters attention other than reaching out here. We’ve spent 3 years developing our platform and we now face losing everything if we can’t find a solution or someone responds either here or via reaching out on Twitter.

Surely there must be a way to alert Twitter from a development point of view?

[gavinwatson]

Could it be a repeat of this?

Twitter Suddenly Locks Thousands Of Accounts [Update: Now Fixed]

[twiz718]

If I had to guess why this is happening, I would wager it’s related to this:
http://www.nasdaq.com/article/twitter-bot-scandal-puts-further-pressure-on-ceo-jack-dorsey-cm763217

[gavinwatson]

Hi @twiz718 thanks for sharing that, I hadn’t seen it.

I guess until any official comment is made, we’ll just be guessing what the problem is and how to address it. One thing for sure, if I cannot find a solution or compromise (which is impossible without knowing what has happened or how to correct it), my entire platform could be history very soon.

If the post you mention is the cause, he (Jack) also needs to remember this post http://www.theverge.com/2015/10/21/9586084/jack-dorsey-twitter-ceo-apology-developers

My frustration is that at the moment the algorithm has been become oversensitive and there is no way of reporting that it isn’t spam nor malicious because they are just not responding.

Anyway, I’m just repeating myself now. I guess we’ll just have to sit back and wait…

@andypiper sorry to reach out to you again, but is there any way you could get any more info about this? I’d sure appreciate it. I have tried the links you gave me to no avail (no response). Thank you.

[andypiper]

Hey all.

The first thing to note is that a number of the links posted in this thread are themselves highly speculative so I’d avoid making assumptions based on those. There are various ways in which the platform algorithms work and as I’ve mentioned on other threads, for safety reasons we can’t go ahead and list all of them out in detail as it would just lead to even more of an “arms race” with those people who do want to spam the platform - note that I’m not labelling anyone on this thread in that manner and do totally understand the frustration you’re facing, and apologise for that.

We also can’t comment on individual accounts or apps on a public form for account owner and developer privacy reasons.

Beyond that, I will note that I did some further investigation and found a couple of potential avenues to get more info on. I’m hoping that as a result, this behaviour should settle out. Note that you may still need to unlock / reset an account again going forward, but hopefully not on a repeated basis.

Finally, one more link to take a look through for future reference is the automation rules, which as bot developers I’m sure most of you are already familiar with. Thanks.

[gavinwatson]

Hi @andypiper, thank you for your information and assistance, much appreciated. Any feedback would be welcomed.

[michichan]

Thanks. We need to get this fixed. I just got locked out again…

[andypiper]

@michichan can you confirm that your app is active and not suspended or write-restricted? Is your account only using a single app to post these updates?

I suggest the best way forward is to open a ticket via https://support.twitter.com/forms/platform with details of your app ID, and a number of the @handles that have been locked as a result of posting Tweets from your app.

Please also review the Tweets being posted for e.g. violations of the Twitter Rules (among other things, pay attention to “if your updates consist mainly of links, and not personal updates”, as well as any use of affiliate links), and the automation rules.

As mentioned much earlier in the thread, if these situations all relate to accounts being “locked” then we are unable to help from a developer platform side, and you will need to pursue these issues with the account support team.

Thank you.