Hello there !
I’m facing app limitations because of spam with @mentions.
In fact i’m running a website to help people find some items and be alerted when there is a restock.
A lot of them were asking for a new functionnality to be alerted only for specific items and we had to develop custom notifications on twitter to prevent them from turning notifications ON for all our tweets.
The way we did that is to mention them in comments on the tweets they are interested in. This way they are not spam with 200+ notifications per day.
The issue is that according to twitter support we are spamming and mentionning people without sollicitation. (Because our website is external and twitter support can’t know that user gave access from there).
Is there any way for users to whitelist an account ? Or is it possible to speak about the project with someone real and not bots to get rid of this limitation ?
Best regards and have a good day 
Paul from DropReference
I seriously doubt there’s any way to do this; it would be basically impossible for Twitter to enforce the automation rules in such a situation (and a whitelisting feature could easily be abused by malicious applications).
Imo you need to re-evaluate your use case. Automated twitter mentions aren’t really meant to handle this - as far as I know they’re meant more for synchronous conversations that occur relatively fast, rather than asynchronous notifications where the user’s message and your message might be weeks apart.
Even if you got users to request their stock update on Twitter rather than on your external website, I suspect it would be flagged up by Twitter’s systems, because it might be 1 day or 10 years before that product becomes available. It’s going to look to Twitter as though your app is sending out a bunch of unsolicited DMs (as it won’t look as though the user’s message, and your subsequent reply, are necessarily correlated).
3 Likes
Thank’s for the time you take to answer
I think that the only way to get rid of this limit is to be whitelisted by someone on the twiter side.
For now we moved our notifications system on an other tool to avoid twitter policy but our users were on twitter and they have to move to another platform 
No problem!
That would probably be the only plausible solution for doing this on Twitter, but as @andypiper mentioned in this post, it unfortunately isn’t possible.
The only way you can really do this on Twitter is via Direct Messages - you can try it, but I would advise you that it is (in my estimation) highly likely your use case would get flagged as spammy/unsolicited automation. The rules about what constitute “spam” or “aggressive automation” are deliberately vague to avoid giving malicious actors loopholes to exploit, so it’s impossible to say for certain if your app would get flagged up for DMing users for stock alerts - these are just my thoughts on it.
The main reason I think it would be flagged is that in automated systems like this, the user gives consent for your app to give an automatic reply - but in almost all use cases, that usually means an instant reply to a message, rather than an eventual reply based on uncontrollable external factors such as stocks of a product. At least to me, when a user gives consent for automated messages it’s usually implicit that the timeframe is not infinite or indefinite, but rather fixed and short.
There are two other issues to consider:
- When a product comes in stock is not within your control - a user might receive 1 message in one month, or 3 messages in two hours. Twitter is almost certain to the view the latter as spammy or ‘aggressive’ even if it’s what the user wanted. You’d have to limit them in some manner so that they could only appear with a maximum frequency, which given the nature of products you usually alert consumers to (GPUs) and the speed at which they sell out, would severely hamper your use case.
- Generally speaking, I think the intent of the automation rules is that if the user gives you one interaction (e.g. sends you one direct message), your app is expected to only message back once, or only message so long as the user continues the interaction. From Twitter’s perspective, a user would ask for a stock alert, and then receive any number of DMs over an unspecified time scale - which would look very much as though your app was just sending DMs to people who didn’t necessarily want them. Where the automation rules are concerned, “the user consents to an automated message” does not equal “the user consents to any number of automated messages until they say otherwise”.
1 Like
Yes i guess you’re 100% right
But as you said @mention was the only solution i found to notify them for some tweets and not others.
I can’t really private message because api rate limit speaking i’d be allowed to PM 15k users per day.
For now almost 500 people registered through twitter oauth for personnal notifications.
Some days very special items drop (4 in a row for example) and almost everyone asked to be notify for. It results that we would be PMing 500 people 4 times in a row, it almost already breaks rate limit.
With mention i was able to group @mention in comments and reply to my tweets. Let’s say i can tag 25 people per tweet i had to do 20 replys.
Rate limit is 300/3 hours, even with 4 huge drop it’s only about 80 status update and it would have been ok until we reach maybe 2k users with personnal alerts (We are aiming for that but are thinking about a mobile app development to enable push notifications by our own if we reach this number of users).
Anyway, i guess there is no solutions As i said we moved on a other platform and a lot of users are complaining about the fact we left twitter. But there is clearly nothing to do i guess
Thank’s for answering, everything is way more clear
ReborN for @DropReference
2 Likes
cutups
6
I’m running into this same issue - my app got flagged due to
“Unsolicited @mentions -
if you send large numbers of unsolicited replies or mentions;”
In my case it’s also the users requesting the mentions themselves inside of the app.
I understand Twitter’s need to minimize spam, but their guidance on this is a bit lacking.
Anyone know what “large numbers” here means? Twitter has not elaborated to me, at least so far three message exchanges that seem semi-automated. I’m just trying to decide whether I can rate limit this in my app to a small number, I just have no idea on the scale. Under 100 per day? Under 20? Under 3? Absolutely none?
If anyone has some in practice advice that would be a great help!
What do you mean by “requesting the mentions themselves inside of the app”? If it’s anything other than them tweeting the account that sends the automated mention, it won’t count (imagine Twitter having to get staff to comb through every app claiming it had users’ consent in an external app area).
Twitter isn’t likely to elaborate on the numbers, as doing so would allow malicious actors to find a loophole in the rules to exploit. That said, when it comes to mentions, I would assume that 1 unsolicited mention is too many.
cutups
8
In the case of my app, it’s event promoters adding their own twitter handle on my site, which is then referenced in a formatted tweet with some additional useful information about their event.
When an event is added to my site, the event is tweeted (once) and mentions the promoter’s twitter (if it exists). The account that sends the tweet is the account for the site - essentially a type of community event aggregation page.
That said, I understand the difficulty in validation of the use case for Twitter. I’m just trying to understand the information they gave me.
If 1 is too many, it would be useful to indicate that it’s just not allowed, rather than saying “large numbers” are not allowed.
I’m not saying there is an actual limit of 1, simply that when it comes to designing your application you should consider it as such (for replies and mentions that start with an @, as opposed to simply mentioning someone in a tweet). The latter - mentioning on its own - is more complex, but there is no real way to quantify any kind of limit to it - Twitter’s adaptive systems figure this out, so it will depend on a lot of factors. As such it’s entirely about making sure your use case is suitable, i.e. not relying on mentions and so forth.
From what you’ve told me, your use case is pretty much always going to result in your app getting suspended imo, as it will consist of purely tweets of what Twitter will see as unsolicited mentions.
Look at this from Twitter’s standpoint: it will see an account whose entire activity is posting tweets at seemingly random users, which won’t be distinguishable from e.g. a bot spreading dodgy health information or trying to sell a questionable product.
1 Like
cutups
10
Understood - thanks for your insight here. This project has been in a dev state for years with the same configuration, so I’m having to gain a bit more knowledge about the Twitter integration. For the most part the tweets that the app sends do not have @ mentions. It seems like it just happened to cross the unknowable threshold recently. I can still use the twitter integration without the mentions. I imagine Twitter is well aware of this, it just took me understanding it.
1 Like
system
Closed
11
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.
