In the space of two days, two different platforms removed something I had published. One was an
article, taken down by an automated filter about two minutes after it went live. The other was an
answer in a community, deleted under a spam policy a few hours after I posted it.
I appealed both. Both appeals worked, in the sense that I got a real answer rather than a form
letter, and one of the two pieces is back up.
I am not going to name the platforms, because the point of this is not that they were wrong. In
one case they were right. The point is what I put in the appeals, because the same move did the
work both times and it is not the move I would have made a month ago.
The move: name the thing you would be accused of, before anyone accuses you of it
Both appeals open by stating the most plausible reason I had been removed, in the platform's
favour, before making any argument of my own.
For the deleted answer, that was the posting interval. I had posted two answers roughly ten
minutes apart that night. I wrote: that is a reasonable thing for an automated filter to stop on,
and I will not do it again regardless of how this appeal is decided. Then I said I had already
written the rule down for myself, one answer per session, no matter what the daily limit allows.
For the removed article, the most plausible reason was its format: a numbered list of six causes
under an "N reasons" title, which is exactly the shape of the mass-produced content their filter
exists to catch. I said so, and I said a filter that stops on that shape is doing a reasonable job.
Neither of those was a rhetorical concession. In both cases I believed it, and in the first case I
had already changed my behaviour before writing the appeal.
The second move: offer the rival explanation, in one sentence
This is the part that surprised me.
In the article appeal, after arguing that the format was not the problem, I added one paragraph
offering a completely different reason I might deserve removal. The post was consumer-facing
troubleshooting on a platform whose audience is developers. I wrote that if that mismatch was the
real issue rather than the format, I would rather be told so plainly, and I would keep that kind of
subject off the blog.
The reply came back and said, in almost exactly those words, that the post fell outside their
developer-focused scope. Not the format. Not the code of conduct, which they said explicitly it
was not. The thing I had offered as an afterthought.
If I had not written that paragraph, I would have received a decision about the format, accepted
it, and corrected the wrong thing. I would have gone on writing consumer troubleshooting and
wondering why it kept getting removed.
A rival explanation written down costs one paragraph. It sometimes comes back as the answer.
What I did not do
I did not argue that the rule was unfair. I did not compare myself to worse offenders. I did not
ask for an exception.
I did make one factual argument, and it was the only one I would call a defence. For the deleted
answer, I compared it structurally to nine of my answers that were still published: word count in
the same range, same number of links, same disclosure paragraph in the same place and the same
words. If the content pattern were the violation, those nine would be violations too. That is
checkable by the reviewer in about a minute, which is the only kind of argument worth making to
someone with a queue.
The ending that made both appeals easier to grant
Both appeals end the same way: if a reviewer looks at this and still considers it a violation, I
accept that and will not repost it.
I think this matters more than it looks. It tells the person reading that granting the appeal does
not commit them to an argument with me, and that denying it does not start one either. It also
made the request smaller. I was not asking them to agree with me. I was asking for a human to read
the thing.
One of the two came back with the content restored. The other came back with a clear explanation
and a decision I accepted, and I have not reposted that piece anywhere it would be the wrong sort
of content again.
The practical version
If something you published gets removed and you think it should not have been:
Write down the most likely reason you were removed, in their favour, first. If you cannot
think of one, you are probably not ready to write the appeal.
Fix the thing before you send the appeal, and say that you have. "I will not do it again
regardless of how this is decided" is only worth writing if it is true.
Offer the rival explanation you would find least convenient. One sentence. It came back as the
actual answer once out of two.
Make one checkable factual argument, not five persuasive ones. The reviewer has a queue.
End by accepting the outcome in advance. You are asking for attention, not agreement.
And the boring one: keep enough of a record that you can make the factual argument at all. I could
compare the deleted answer to nine live ones because the word counts, link counts and dates were
written down at the time. Without that, my appeal would have been a paragraph of feeling.
Disclosure
I build BlueTicks for Gmail, a Chrome and Firefox extension that shows WhatsApp style ticks in your
Gmail sent list, one tick sent and two blue ticks opened. It costs 4 dollars a year and there is a
free tier. Everything above happened while publishing about it on platforms I do not own, which is
most of what my distribution work actually consists of. You can find it at blueticks.io.
The removals were both reasonable calls made quickly at scale. Nothing here is a complaint about
either platform, and one of the two taught me something about my own writing that I would not have
learned otherwise.
Top comments (0)