How to prove your giveaway was fair

On this page: what provably fair means · why an RNG isn’t proof · the four things a fair draw needs · levels of proof compared · what to publish · what to look for in a tool · FAQ

What does “provably fair” mean for a giveaway?

A giveaway is provably fair when the organizer locks in the entries and the method before the result exists, publishes something that binds them to it, and lets anyone check afterward that the announced winner matches what was locked in.

The important word is before. Committing to your inputs ahead of time means you can’t quietly change them once you see who won. That’s the whole idea: not that you promise you behaved, but that you gave up the ability to misbehave without it being detectable.

Two properties make a result provable:

• You can’t change the answer after seeing it. The entry list and the method were fixed before the draw ran.
• You can’t deny what you committed to. The commitment was published, so nobody has to take your later description of it on faith.

Why isn’t a random number generator proof?

A random number generator addresses randomness. It doesn’t address integrity, and integrity is what your audience is actually questioning.

Nobody suspects your software of picking a bad random number. What they suspect is one of these:

• You ran the draw several times until you liked the result
• You added or removed entries before running it
• You quietly excluded someone
• You picked a name and worked backward

A random result you can re-roll in private is not verifiable. Ten honest draws and one dishonest one look exactly the same from outside if all anyone sees is your announcement. So “we used a certified random algorithm” answers a question nobody asked.

This is why screenshots don’t settle it either. A screenshot proves a result was displayed. It doesn’t prove it was the first one, or the only one.

What does a verifiable draw actually require?

Four things, and all four have to be in place before the draw runs.

1. The entry list is fixed and provable. Something published beforehand has to pin down exactly which entries were in the pool. Otherwise “we drew from 4,312 entries” is just another claim.
2. The method is fixed. How many winners, any weighting, any exclusions, how substitutes get chosen. Deciding after the fact that you meant to exclude a category is indistinguishable from rigging.
3. The result is published with a timestamp. Winners, when it ran, and under which configuration.
4. Anyone can check the result against the commitment. Not just you, and not just a regulator. The person who entered and lost is the one with the doubt.

Miss the first one and the rest doesn’t help. A perfectly documented draw from a pool you could edit is a perfectly documented nothing.

How much proof do you actually need?

More than most organizers offer, and less than you’d think. Here’s the ladder, from weakest to strongest.

Most organizers sit on rows two to four and are surprised when that isn’t enough. The jump that changes the conversation is the last row, because it’s the first one where checking doesn’t depend on believing you.

For a small giveaway among people who already know you, a published winners list may be plenty. Proof costs effort, and matching it to what’s actually at stake is reasonable. The threshold moves when the prize is valuable, when the audience is strangers, when a regulator or client is watching, or when a single accusation would cost you more than the prize did.

What should you publish?

Publish two things at two different times.

Before the draw:

• The entry pool, or a fingerprint of it that can be checked later
• The number of winners, and any rules that change someone’s odds
• Any exclusions, stated as rules rather than as names
• When the draw will run

After the draw:

• The winners, at whatever level of anonymization you promised
• The exact time it ran
• A way to confirm the result matches what you committed to beforehand

Anonymization is worth deciding early. Publishing full names is often unnecessary and sometimes unlawful depending on where your entrants live. Initials or a partial identifier usually gives the audience enough to see the draw was real without exposing the winner. Say which you’ll use in your rules, before anyone enters.

What should you look for in a giveaway tool?

Ask one question: What can a skeptical entrant check without contacting you?

That single question sorts the market quickly. Some concrete follow-ups:

• Is there a public results page, or does verification mean emailing support for a PDF?
• Was the entry pool committed to before the draw, or is the record generated afterward from whatever’s in the database now?
• Can the organizer edit entries after the pool is sealed? If yes, what does the record show?
• Does the audit record cover instant win and judged contests too, or only random draws?
• If the tool disappeared tomorrow, would the published proof still mean anything?

That last one separates real verification from a vendor badge. A certificate that only the issuing company can interpret is a promise wearing a lab coat.

Common mistakes

Treating fairness as a post-incident response. Proof has to exist before the draw. Once someone has accused you, anything you produce looks assembled in response, because it was.

Publishing the winner but not the rules. The winners list is the least informative part. What convinces people is that the rules were fixed and public before the draw.

Changing something mid-campaign without saying so. Extending a deadline or adding a prize is fine. Doing it silently is what turns an ordinary decision into evidence of bad faith.

Excluding people by name rather than by rule. “Employees and their families are not eligible” is a rule. Removing a specific person after entries close is not, however good the reason.

Assuming nobody will check. Usually true, and irrelevant. The value of proof shows up on the one occasion someone does.

Frequently asked questions

Does a fair giveaway have to be verifiable? No. Plenty of giveaways are run honestly with no proof at all. Verifiability matters when your audience has no particular reason to trust you, or when the cost of being accused is high. The two are separate: fairness is what you did, verifiability is what you can show.

Is a live video draw good enough? It’s better than a screenshot and worse than a published commitment. Video shows the draw happening, but it doesn’t show what was in the pool beforehand or whether earlier attempts were discarded. It also asks your audience to watch in real time, which most of them won’t.

What if I have to disqualify someone after the draw? Write the disqualification rules before you open entries, publish them, and apply them visibly. Removing a winner is defensible when the rule existed beforehand and the reason is stated. It’s indefensible when the rule appears afterward.

Can I still keep entrant names private? Yes. Verification and privacy aren’t in conflict. What has to be checkable is that the result matches the commitment, and that works with partial identifiers or initials. Decide the level in your rules before entries open, so nobody is surprised.

Do instant win campaigns work the same way? The principle carries over, but the timing differs. Each instant win result happens privately on one person’s screen, so the thing to commit to beforehand is the prize schedule: how many prizes exist and when they can be won. There’s more on this in how to run an instant win campaign.

Isn’t this overkill for a small giveaway? Often, yes. Match the proof to what’s at stake. A follower giveaway with a mug as the prize doesn’t need what a five-figure sweepstakes needs. The mistake is deciding you don’t need proof and then discovering, after an accusation, that you can’t produce any.

Raflia seals every drawing before it runs and publishes a results page your entrants can check for themselves. See how verification works, or join the waitlist and we’ll email you when Raflia opens.