Unsolicited @mentions - Account Restricted from performing write actions


Hi. My company matches volunteers to non profits and today we turned on a small feature that would tweet that volunteers are needed for a volunteer opportunity whenever a non profit lists their opportunity on our site. The tweet would simply list the name of the event, give the location, put a link to where people can register and then mention the non profit that posted it. After a few hours there had been 25 new opportunities listed and for 11 of them the non profit had entered their twitter handle so we would add @twitterHandle in the tweet

We were quickly denied write access via the api for having too many unsolicited mentions. I’ve been searching the documents and rules and guidelines and can’t for the life of me figure out:

  1. Why is 11 considered too many?
  2. What is unsolicited? If the non profit enters their twitter Handle and we are trying to get them volunteers, why would we be banned for mentioning them
  3. How does a mention become “solicited”?

Note that our site has no ads, no spam, is free to use, etc

Do I have to follow the person for it not to be unsolicited?
Do they have to follow me for it not to be unsolicited?


I think the basic idea here is that if you are @mentioning users regularly with no interaction from that user i.e. they don’t themselves follow you, reply to the Tweet, or @mention your handle first, then our automated antispam system is likely to deem this as potentially abusive. I understand your use case, but unfortunately for every good idea like yours, there are a lot of bad folks who want to generate noise or even to deliberately harass others, so the system is geared towards safety.

The automation rules generally have more information on this, but I do understand how frustrating it is when you’re trying to build something like this, and it is difficult to document every case that would cause an app to trigger our antispam system, because it is an adaptive set of rules that has to respond to new attacks every day. If you’re interested in learning more about it (it is called “Botmaker”) then you can watch a video from the team from our Flight conference last year.

Help with replies, doesn't show on users timeline

Just speculation based on what i’ve seen before, but: It should definitely help if run your account like you would your personal one for a little while and do things that spam accounts rarely do - like adding a phone number to the account, interact with others more (participate in conversations, get the accounts you’ll be mentioning to follow you etc - same way that @givepulse is doing)

Even more speculation: It may also help to recreate the application or even account again - manually setting things up and getting started without using the API first, then automate things - I suspect that accounts & apps that get flagged for spam are remembered in some way, and are more likely to get flagged again.


Thanks Igor. I guess we’ll just take it slow in the beginning and not turn it on till we’ve built up more of a following.


Thanks Andy. It is frustrating, but hopefully we can get our account unblocked and be back up and running soon.

So if we can’t mention people directly unless they are following us (for example), that makes me wonder if we need to split our twitter app into multiple apps, one for each major city or state since if we’re posting volunteer opportunities in Hawaii, Texas and NYC, it might be too much noise for most people to follow (they only care about local events). Note that we do send lat/long, but not sure if users can filter a twitter feed based on the lat/long of the tweets?

Would you recommend creating 50-500 apps for the majors states and cities (and countries) so that our feeds are more follow-able for people in specific locations?

e.g. GivePulseAustin, GivePulseNYC, GivePulseChicago, GivePulseTX, GivePulseCanada, etc


I would not recommend that, because you’ll be rate-limited on app creation (a few per day), and on overall number of apps that your account can support. Generally we wouldn’t recommend this kind of pattern.

At the moment, I’m not sure this is the best way of trying to promote these opportunities, given the platform rules around automation and the safety system.


We will stick with one app and use hashtags then. Our thought was that it would be great if civic-minded people could follow a twitter feed that listed volunteer opportunities in their community. They would be able to learn about new volunteer opportunities and see what local non profits are doing and make it easier to get involved. Twitter seems like a great medium for this… but since we can’t @mention non profits, and people might not want to follow a noisy channel, they will have to just follow/search for the hashtags (#VolunteerAustin or #GetInvolvedChicago), not ideal, but a semi workable temporary solution


Starting off manually tweeting & curating accounts is a good idea, also adding location tags is a great and should help people searching.

As Andy said, creating multiple applications for the same purpose will get you flagged for spam too. You’re much better off just sticking with 1 app & 1 dedicated account to start out and create accounts for different areas later as you grow.

For a good example, check out VolunteerIreland - they do exactly this kind of thing really well on twitter: VolunteerDublin covers the city & surrounding county, VolunteerCarlow covers Carlow & surrounding areas, and there are a few others i think. Maybe it will help get some ideas from those accounts and how they’re run: in terms of what kind of conversations and types of posts, encouraging the groups / charities looking for volunteers to tweet themselves and just RT / Quote them to announce etc.