A seller I follow ran a giveaway last winter for a set of hand-glazed ceramic mugs. Good product. Good photos. The rule was the usual one: comment MUG and tag a friend.
The post pulled 4,300 comments. She was delighted for about two days.
Then she opened her inbox and found almost nothing in it. No orders. Barely any questions. Her follower count spiked, leaked back down over the next ten days, and left her with a screenshot of comments she never spoke to.
The prize wasn't the problem. The giveaway just ended at the place where it should have started.
A giveaway is a list, not a party
Here is what most sellers have backwards. The comment is not the result. The comment is a raised hand. Someone announced in public that they want something you sell. If you leave that hand in the air and only speak to the one winner, you traded thousands of warm conversations for a single mug.
The job of the giveaway is to move every one of those people out of a comment thread you don't control and into a DM thread you do. Instagram shows your post to whoever it feels like showing it to. Your inbox is yours. That is the entire trade.
So the score isn't comments. The score is how many commenters ended up in a real conversation, and how many of those bought something in the following month. A 900-comment giveaway with a DM behind it will quietly outsell a 6,000-comment one without.
The comment-to-DM loop, in the order it should happen
The mechanics aren't complicated. What breaks is the sequencing.
- Post with one trigger word. One word, spelled one way, said in the caption and in the first three seconds of the video if it's a reel.
- Reply to the comment publicly, automatically. A short human line: "Sent you a DM." That does two jobs. It tells the commenter to check their inbox, and it shows everyone else reading that the DM is real.
- Send the DM within seconds. Not in a nightly batch. Not tomorrow morning. The person is still holding their phone.
- Confirm the entry and ask one question. This is the step nearly everyone skips, and it is where the money is.
- Follow up after the draw. Everyone who lost is still on your list, and they still want the thing.
Steps one through three are what people usually mean by turning Instagram comments into customers. Steps four and five are what turn it into revenue.
Choose a trigger word that filters for you
"Comment YES" is a magnet for anyone scrolling past. So is a bare emoji. You will get volume and nothing else.
Add a small amount of friction and the word starts sorting people before you reply to anyone. Ask for the word plus a detail only a real prospect would bother typing: their size, their skin type, the colour they'd pick, the city they want it shipped to.
"Comment MUG and the colour you'd choose" costs a bot nothing and costs a bored scroller one second of thought. Better still, you now have a stated preference attached to every entry, which makes the DM you send next feel like it was written for that person. Because it was.
One warning. Keep the word short and hard to mistype. Pick something with a common misspelling and a chunk of your entries won't trigger anything at all, and those people will decide you ignored them.
The first DM decides everything
Most giveaway DMs read like a receipt. "Your entry has been recorded. Winner announced Friday. Good luck!" There's nothing to reply to, so nobody replies, and the thread dies with the conversation still closed.
Write an opening instead. Confirm the entry in one line, then ask something they can answer with one tap:
"You're in for the mug set 🎉 Out of curiosity, is this for you or a gift? I'll send the two we'd suggest either way."
Now the thread is alive. They answer, you send two options, and a slice of them ask a price question without you pitching anything at all. It's an unglamorous path and it beats a broadcast every time. For the shape of that conversation in detail, the DM selling playbook covers where these threads usually stall.
There is a catch, and it's the real one. This only works if somebody answers back. Nine hundred people replying "it's a gift" at eleven at night will bury a solo seller in a single evening. That's the honest reason giveaway DMs get set up as one-way announcements. Sellers know better; they just can't physically keep up. Which is a capacity problem, not a strategy.
Thinning out fake entries without insulting the real ones
Every giveaway attracts the giveaway crowd: accounts with forty followers, no posts, and a bio full of trophy emojis. They enter everything, win almost nothing, and unfollow you the day after the draw. They cost you prize money and they bend every number you look at afterwards.
Cheap filters that hold up:
- Require a DM reply for the entry to count. The comment gets them in the door; one answer in the DM confirms it. Bots don't answer. Neither do people who were never really interested.
- Ask for something specific in the comment (colour, size, city) rather than a generic word.
- Skip accounts created that week or with no posts at all, quietly, at the draw stage. You don't need to announce this anywhere.
- Say the winner is contacted by DM and has 24 hours to reply. Serial-entrant accounts often don't.
- Don't ask people to tag five friends. You get tags from people who will never look at your page and a comment section nobody can read.
You'll never reach zero fakes and chasing zero isn't worth your evening. Going from sixty percent junk to twenty percent is enough to make the follow-up worth doing.
The people who lost are the reason you ran it
One person wins. The other 4,299 are the actual return on the whole exercise, and most sellers never message them again.
The day after the draw, send one honest message. Not a consolation discount dressed up as generosity; people can smell that from across the app. Closer to this: "You didn't win the set. You said you'd pick the green one, and it's back in stock this week. Here's what it costs." Plain, slightly disappointing, useful.
Send it once. If they don't reply, leave them alone. A list is only worth something for as long as you don't burn it.
Four numbers worth watching
- Comment-to-DM rate. Of everyone who commented, how many actually received and opened a DM? Under seventy percent means your trigger word or your automation is dropping people on the floor.
- DM reply rate. How many wrote back? Ten percent is weak. A real question in the opener does considerably better than a receipt.
- Sales in the thirty days after. Attributed to entrants, not to the post. This is the only number that tells you whether the prize paid for itself.
- Follower retention at thirty days. If most of the spike is gone, your prize attracted prize-hunters instead of buyers. Change the prize, not the caption.
If those figures look grim next to your usual inbox, hold them against a realistic DM conversion benchmark before you write the giveaway off. Giveaway traffic converts lower than ordinary DM traffic. It is supposed to. It is also many times the volume.
Where this usually falls apart
Not in the setup. In the hour after the post takes off.
Switching on a comment trigger takes ten minutes. Answering the nine hundred replies it produces does not. That stretch is where an assistant earns its keep: one that reads each reply, answers price and shipping questions in your voice, remembers the colour that person mentioned, and hands you the awkward ones. Vardast connects to your Instagram, WhatsApp and Telegram and covers exactly that window, from the comment trigger through the first DM and the hundreds of conversations behind it, without you sitting up until two in the morning.
Run your next one this way. Same prize, same post, one extra question in the DM.
Then count the sales in the month afterwards instead of the comments on the day. That is the only review a giveaway needs.