I’ve seen so many trolls and snarky comments online that I think I’m fairly immune to them at this point. Some people are seriously emboldened by anonymity, and it makes them say things online that they would never say to someone face to face. That can be a costly mistake.
What most people don’t reckon with is how long a comment outlives the moment that produced it.
Years ago, a kid told my daughter on Instagram: post about my man again and I’ll gouge your eyes out. He regretted it and deleted it. That didn’t help him much, because we had screen captures showing the account it came from.
That’s the part worth understanding properly.
What’s actually true about permanence
I used to tell this story by saying that platforms are legally required to keep everything forever. That isn’t right, and it’s worth correcting because the real position is more useful anyway.
Screenshots are the permanent part. The moment anyone captures what you wrote, it exists entirely outside your control and outside the platform’s. No delete button reaches it. That is the mechanism that actually makes online comments permanent, and it requires no law at all.
Platforms keep what they keep. Retention is governed by each company’s own policy, not by statute. Deleted content frequently persists in backups and internal logs for a period that varies by platform and that you have no visibility into.
Preservation is real but narrower than I thought. Under 18 U.S.C. § 2703(f), a provider must take all necessary steps to preserve records when a government entity asks — for 90 days, extendable by another 90 on a renewed request. If no legal process follows, it’s the provider’s call whether to delete.
So a threat made online can absolutely be preserved and subpoenaed. It just happens because someone involves law enforcement, not because the data was never going anywhere.
The practical conclusion is unchanged: treat anything you post as permanent, because the parts you can’t control usually are.
Why this matters more when you own a business
A teenager who posts something vicious faces consequences that are real but mostly personal. An owner’s online conduct is attached to a company, which changes the arithmetic considerably.
Your comment is the company’s comment. Not legally, in most cases. But in practice nobody separates the founder’s temper from the business’s character, particularly in a small company where you are the visible face of it.
It’s discoverable by exactly the people who matter. Prospective clients search for you. So do potential hires, partners and journalists. They aren’t reading your best work — they’re scanning for anything that gives them pause.
And the audience for a heated reply is never who you think. You’re writing to the person who annoyed you. It’s read by everyone else.
I’d add one specific case: replying to a bad review while angry. It is the single most common way small businesses damage themselves publicly, because the review was one person’s opinion and the reply is evidence of how you handle criticism. Prospective customers read the reply far more carefully than the complaint.
A workable standard
Not “be inoffensive” — having opinions is part of being worth reading, and a business with no point of view is forgettable.
The test I use is simpler: would I be comfortable if this were screenshotted and shown to a client, an employee, or my children?
That allows for disagreement, strong positions, and criticism. It rules out the thing you’d regret, which is almost always tone rather than substance.
Three practical habits:
- Never post while angry. Write it if you need to. Don’t send it. Almost nothing genuinely needs answering within the hour.
- Assume every private channel is public. DMs, group chats, internal Slack. Anything that can be screenshotted will be, if it ever becomes worth someone’s while.
- If you’re wrong, say so plainly and quickly. A correction posted fast reads as integrity. A deleted post reads as a cover-up, and someone has the screenshot anyway.
And if it’s aimed at you
Being on the receiving end is different, and worth separating out.
Document it before you respond. Screenshots with dates and the account visible. If it escalates, that record is what gives you options.
Decide whether responding serves you. Usually it doesn’t. Responding rewards it with attention and puts your reaction on the record next to their comment.
Involve the platform or the police where it warrants it. A threat of violence is not a rude comment, and it should be treated as what it is. That’s the situation preservation requests exist for.
What I said to my daughter at the time still seems right. You cannot control what other people put online. You can control what you put online, and you can make sure that if this ever ends up in front of someone who matters, your half of it reads well.
