A Facebook giveaway comment count is not the number of entries. Facebook's comment badge counts every comment ever posted, including comments its spam and quality filters hide from everyone, so the badge can be three or four times larger than the list of comments anyone can read, pick from or export. To pick winners fairly, build the entry list from the comments Facebook actually serves, with the filter set to All comments, then de-duplicate by commenter and draw from that list.
This guide walks through a real case where a badge of 784 turned into 201 exportable comments, compares the ways of collecting entries, and gives a draw method that holds up when an entrant asks why they did not win.
The 784-versus-201 case, checked three ways
On 15 November 2026 a customer asked why their export of a Facebook giveaway post was "missing" most of the comments. The post was a Thai utility brand's tag-a-friend giveaway with 200 prizes. The numbers looked like this:
- Comment badge on the post: 784.
- Rows in the ZocialComment export: 201.
- Logged-in Facebook, filter set to All comments, scrolled to the very end: 202 comments, and no reply threads left to expand.
The comments in the browser that were not in the file had all been posted after the export ran. Every comment in the file was also in the browser. So the export was complete: it held every comment Facebook was willing to show to a logged-in person with the most complete filter. The remaining 580 or so comments counted by the badge were not visible to anyone, and no method of reading the post could reach them.
One detail made the gap matter more than usual. With 200 prizes and about 200 visible entries, nearly every visible entrant would win. A host working from the badge would have expected roughly one winner in four.
Where the missing comments go
The badge on a Facebook post is a running tally. Comments add to it when they are posted, and in practice they rarely come off it again. Several things remove a comment from view while leaving it in the count:
- Spam and quality filtering. Facebook hides comments it classifies as spam or low quality. Meta's spam standard covers repetitive and mass-posted content, which is exactly what a tag-a-friend giveaway produces: hundreds of near-identical comments, each made of a few names, often from new or rarely used accounts.
- Page moderation. A Page's own profanity filter and blocked-word list hide comments from everyone except the commenter and their friends.
- Deleted comments and restricted accounts. Comments removed by their authors, and comments from accounts that were later disabled, stop being served.
Tag-a-friend giveaways are filtered more heavily than any other kind of post we export. A question post or a product announcement with a similar badge usually exports a far larger share of its count.
All comments versus Most relevant: the view decides the entry list
Facebook opens most posts in Most relevant, a ranked view. Meta describes ranking as choosing what to show first based on predicted interest (see how Meta explains ranking), and the comment view follows the same idea: some comments are pushed down, collapsed or left out. Newest shows comments by time but can still hide some. All comments shows every comment Facebook will serve, including ones the ranked view would skip.
To change it, open the post, click the filter label above the comments and choose All comments. A host who scrolls Most relevant and copies names will miss entrants who are perfectly visible in All comments. In our experience the difference on a large giveaway is often dozens of people.
Four ways to collect the entries, compared
| Method | What it reads | Duplicate and tag checks | Record of the draw | Best for |
|---|---|---|---|---|
| Scrolling Most relevant | A ranked subset of visible comments | By eye only | None | Never for a draw |
| Scrolling All comments and copying by hand | Every visible comment | By eye, slow and error-prone | Screenshots only | Under about 50 entries |
| Built-in Facebook features (Page inbox, Insights) | Counts and notifications, not an entry list | None | None | Checking activity, not picking |
| Exporting with ZocialComment | Every comment Facebook serves, with replies | Sort and filter on author_id and text | A dated file you can share | Any draw over about 50 entries |
Which to use when: for a small giveaway with a few dozen comments, scrolling All comments and copying by hand works. For anything larger, or any giveaway where a disputed result would hurt the brand, export the comments and draw from the file. No method reaches the hidden comments behind the badge, so the method only decides how accurately and verifiably you handle the visible ones.
Drawing winners from the comments that exist
Start by closing entries at a stated time, then export. Paste the post link at the Facebook comment exporter and download CSV, Excel or JSON. The first 100 comments come with no signup; a larger thread is a one-time payment, never a subscription. Each row carries author, username, text, likes, replies, created_at, language, is_pinned, id, reply_to_id and avatar_url, plus Facebook extras such as author_id, reaction_count and the per-reaction breakdown.
Before drawing, compare the row count with a logged-in All comments scroll. If the two agree within a few late comments, the file is your entry list. Then decide, in writing, what counts: top-level comments only, or replies too. reply_to_id is empty for a top-level comment and points to the parent for a reply, so either rule is one filter away. If you would rather not touch a spreadsheet, the Facebook comment picker draws winners straight from a post link.
One entry per person: de-duplicating by commenter
Comment-to-enter giveaways attract repeat entries: the same person commenting five times with five different friends. Unless the rules say each comment is an entry, keep one row per person.
- Sort the sheet by
author_id, then bycreated_atoldest first. - Add a helper column with
=COUNTIF($M$2:M2, M2), where M is theauthor_idcolumn. The first comment from each person reads 1. - Filter the helper column to 1. What remains is one entry per commenter.
Use author_id, not the name. Two different people can share a display name, and one person can change theirs during the giveaway. The numeric ID stays the same. Remove the Page's own account and any staff accounts at this step too.
Checking the tagged friends
If an existing giveaway asked for tags, you still have to judge the entries it received. Facebook writes a tagged friend into the comment text as a plain name with no @ symbol, so a tag looks like ordinary text in the file. Three checks cover most rule-breaking:
- Empty or one-word comments. Filter
textfor very short values. A comment such as "me" or a single emoji usually tagged nobody. - Too few names. If the rules require two friends, a comment with one name fails. Count capitalised words or separators as a rough first pass, then read the borderline rows.
- Tags that are not people. Before announcing, open each winning comment on Facebook and confirm the tagged names are links to real profiles rather than typed text or tagged Pages.
Check winners and a few reserves this way rather than every entry. It is quicker, and it is the part entrants will actually question.
A draw record people can check
A transparent draw is one that someone else could repeat. Keep three things:
- The exported file, untouched, with its download time in the file name.
- The cleaned entry sheet, showing which rows were removed and why (duplicate, missing tag, staff).
- The draw itself. Add a column with RAND in Google Sheets or RAND in Excel, copy and paste it as values so it stops changing, sort, and take the top rows as winners, then the next rows as reserves. Screen-record those steps.
When the visible list is close to the number of prizes, say so in the announcement: "202 eligible comments, 200 prizes." Entrants who saw 784 on the badge will otherwise assume most entries were ignored.
How we know this: the ticket behind the numbers
This guide comes from operating the exporter, not from reading about Facebook. The 784-to-201 case above is a support ticket we worked on 15 November 2026: we compared the customer's file row by row with a logged-in browser session on the All comments filter, scrolled to the end, and found nothing in the browser that was missing from the export apart from comments posted after it ran.
The same pattern repeats across the Facebook posts people export with us. The badge is always the larger number, the gap is widest on tag-a-friend and comment-to-enter giveaways, and the rows in the file match what a person sees in All comments. Our earlier audit of zero-comment results found the same thing from the other side: giveaway posts with comments turned off still show a count but serve nothing. The column names, the no-signup first 100 comments and the one-time payment model described here are the live product values at the time of writing.
What Meta's promotion rules ask of the host
Meta's Pages, Groups and Events Policies allow contests, giveaways and sweepstakes on Pages, with conditions. The host must follow the law where the promotion runs, give entrants official rules, eligibility requirements and offer terms, require participants to release Meta from liability, and state that the promotion is not sponsored, endorsed or administered by Meta.
One line in that policy matters directly for the case above: a promotion must not require or incentivise participants to share, repost, tag others or otherwise publicise it. A "tag two friends to enter" rule is exactly that, so the format is outside Meta's rules even though it is common. It is also the format Facebook's filters hide most heavily. The compliant version asks for a comment answer, such as a favourite product or a word from the caption, and draws from those comments; it also exports cleaner, because the comments carry real text instead of lists of names.
Whatever the format, add a line to your rules saying that only comments visible on the post at closing time are eligible. That protects the host when the badge and the entry list disagree. If commenters are in the EU, the GDPR also applies to the entry file: keep it for the draw and for prize delivery, then delete it.
Download your Facebook giveaway entries →
Related reading: picking a Facebook comment winner, exporting Facebook commenter user IDs, downloading Facebook giveaway comments.
