Streamers

How to Run Transparent Giveaways on Stream

A clear giveaway starts before the drawing. Agree on the rules, check who is entered, show the result, and give winners a reliable way to claim.

Suppose you have three game keys to give away during a community stream. A moderator collects entries while you finish a match. When you open the picker, one username appears twice, someone asks why they are missing, and another viewer wants to know whether subscribers get better odds.

The wheel has not moved yet, but the important decisions are already waiting. Who belongs in the pool? Can someone win twice? Which key goes to the first winner? What happens if that person misses the announcement?

This is an illustrative scenario, but it captures the work a drawing tool cannot do for you. A good giveaway makes those decisions understandable before anyone wins. The animation can be entertaining; the procedure needs to remain clear when an entry is disputed or a connection fails.

Decide what you are promising

Start with the prize and the people who can actually receive it. For game keys, check redemption regions, supported platforms, and any restrictions attached to the keys. For a physical prize, establish where you can deliver it and how fulfillment will work before describing the giveaway as worldwide.

Then separate four questions. Eligibility determines who may enter. Selection probability determines each valid entry's chance. Transparency concerns what participants can understand and observe. Fulfillment concerns whether the promised prize reaches the verified winner. A visible wheel addresses only part of that chain.

For an ordinary community drawing, a manageable default is free entry, one valid entry per eligible person, and equal chances. A quiet viewer is not automatically a bot, and a subscriber is not automatically more eligible than everyone else. Any restrictions need a reason, clear advance notice, and a way to administer them consistently.

A judged creative competition is a different format: the award depends on published criteria and judging, rather than a random draw. Calling something a contest, raffle, or giveaway does not by itself settle its legal classification.

Set the rules before opening entries

For the three-key example, decide which key is awarded first and whether each person can receive only one. Describe entry opening and closing times with a time zone, the entry method, age and geographic restrictions, and the number and nature of prizes. Explain selection, notification, the claim deadline, eligibility verification, and how an alternate becomes entitled to a prize.

Put the full rules somewhere entrants can access and retain before entering. State the conditions that materially affect participation in the announcement itself. Include the organizer's contact method, expected delivery timing, and how personal data will be used. A rules link is useful only if it works when viewers follow it.

Keep payment outside the default drawing

In common US explanations, a prize, chance, and consideration—something of value required to participate—raise lottery questions. The FTC's consumer guidance warns against payment to enter sweepstakes or improve the odds. That is a sound reason to avoid casually attaching tickets to subscriptions, donations, purchases, or Bits.

The legal details still depend on the promotion and jurisdiction. For example, Washington's statute permits certain purchase-linked additional entries when a no-consideration alternative exists. It does not establish a rule for every state or certify a particular Twitch promotion. Subscriber-only arrangements need review in their actual circumstances; adding a free form is not a universal compliance fix.

In Great Britain, the Gambling Commission distinguishes wholly free draws from compliant paid/free arrangements. The latter require a properly promoted free route that is no less convenient, with no distinction between routes when allocating prizes. Genuine prize competitions have a separate skill threshold; a simple question rarely suffices.

These are bounded examples, not worldwide rules. Free entry also does not resolve every registration, advertising, tax, or prize obligation. For a valuable, sponsored, or cross-border promotion, obtain advice for the places and participants involved before opening entries.

Check the platform as well as the law

Under Twitch's current Promotions clause, you are responsible for the promotion, including its rules, administration, winners, prizes, and applicable approvals. You must not imply Twitch sponsorship. When promoting, administering, or conducting it, display or read the required disclaimer, replacing the name placeholder:

This is a promotion by [Your Name]. Twitch does not sponsor or endorse this promotion and is not responsible for it.

If the stream includes branded content, check the disclosure obligations in those Terms too. Separately, Twitch's Channel Points policy prohibits gambling redemptions and restricts exchanges for value outside Twitch. Do not assume Points are an unrestricted substitute for payment.

YouTube's contest policy requires free entry and official rules consistent with its Terms, including a Community Guidelines link and disqualification of noncompliant entries. Rules must identify YouTube as a nonsponsor, require a release of YouTube from contest liability, and include a compliant privacy notice. Run and award the contest as described. The policy also prohibits manipulating engagement metrics to misrepresent genuine engagement; do not assume every like-or-subscribe mechanic is permitted.

Choose a drawing workflow

Choose the simplest procedure that meets the promise you made. A chat bot, a numbered list, and a wheel can all present a single drawing. Their appearance does not tell you how accurately they collected entries or how the winner was calculated.

The comparison below describes procedures, rather than interchangeable software features.

Workflow Useful when Check
Single equal-chance draw One prize, with a quick result Close and validate one pool before selection.
Multiple winners without replacement Several prizes, with no repeat winners Track remaining entries and the announced prize order.
Finalists and elimination rounds A scheduled segment benefits from stages Explain finalist selection and every advancement rule.

One prize from one eligible pool

Close entry, validate the list, confirm the count, and select from that fixed pool. If there are 160 valid entries and selection is uniform, each entry has a 1/160 chance. A duplicate belonging to one person changes that person's opportunity even if the randomizer treats every row equally.

For a small group, viewers may be able to check a displayed list. With hundreds of names, use a searchable confirmation method or a moderator-assisted correction process. Tiny labels around a wheel are not a practical entry receipt. Show a readable text result and announce the selected account aloud.

Several prizes without repeat winners

For the three keys, announce the award order first. Draw the first provisional winner, record that account against the first key, and remove it from the eligible pool before selecting for the second. Repeat for the third. This is sampling without replacement.

With 160 equally eligible entrants and uniform selection at each step, every remaining entrant has a 1/159 chance at the second draw, then 1/158 at the third. Before any draw, each original entrant has a 3/160 chance of receiving one of the three prizes, assuming all three are awarded to distinct eligible people without disqualifications. That does not mean everyone has the same chance of every prize after the first result is known.

Check the actual remaining list between selections. Before relying on a tool's no-repeat setting, rehearse drawn entries, manual exclusions, the last remaining entrant, and an exhausted pool with dummy data. Reset only for an announced new round. A greyed label alone does not establish which entries the selection algorithm excludes.

Finalists followed by elimination rounds

For a planned feature segment, you might select 16 finalists from the full eligible pool and then reveal successive head-to-head results. Publish both stages before entry. A random matchup advances one participant by chance; it is not a test of playing skill simply because it looks like a tournament.

Use a fixed bracket and make clear whether a selected name advances or is eliminated. Preserve each result. A 16-person single-elimination bracket has 15 match decisions, so budget for transitions and explanations as well as reveals. Extend the segment only if it serves the event; no particular duration is a fairness standard.

Understand what weights and stages change

In a proportional weighted draw, a participant's chance is their weight divided by the total eligible weight. This assumes positive weights and an algorithm that actually samples in those proportions. The following is a mathematical illustration, not a recommendation to award bonus entries.

Entrant Weight Chance
A 1 12.5%
B 1 12.5%
C 2 25%
D 4 50%

The total is eight. D has four times A's chance, not a guaranteed win. A proportional wheel could represent those weights with angles of 45°, 45°, 90°, and 180°, but correct geometry does not prove correct random selection. Nor does correct weighting establish whether the entry rules are appropriate or permitted. For the equal-chance workflow here, keep every eligible entry at weight one.

Stages require the same care. Suppose exactly 16 distinct finalists are selected uniformly without replacement from 160 eligible entrants. Each finalist then starts in a complete 16-person bracket, plays four rounds to win, and has a conditional 50% chance in every matchup, regardless of the opponent or preceding results. Under those assumptions:

Chance of becoming champion = 16/160 × (1/2)^4 = 1/160 = 0.625%.

The stages have changed the presentation without changing an original entrant's overall chance. This calculation does not describe a skill tournament or establish the behavior of any specific app.

Now imagine one finalist advancing from a group of ten and another from a group of one hundred. If each group selects uniformly and the finalists face an equal-chance final, an entrant in the smaller group has ten times the overall chance. Byes, re-entry, unequal match outcomes, and inconsistent eligibility checks can also change the calculation. Equal chances within an individual step are insufficient to establish equality across the full process.

Prepare the list and close entries deliberately

Keep a private working record connecting each entry to the account or identifier you can verify. Use display names or entry identifiers in the public view, never shipping details. One account per row is useful bookkeeping, but it does not establish one human per entry.

  1. Collect through the announced route. An illustrative !enter command needs a bot or moderator process that actually records it; a command in chat is not automatically an entry in a separate picker.
  2. Review duplicates and eligibility. Normalize accidental spacing or case differences, but inspect possible matches before merging people. Apply only the criteria announced beforehand.
  3. Check the imported list. Compare counts and sample entries, including the first and last rows. Confirm equal weights and distinguish people with similar names.
  4. Allow corrections within the announced timetable. Give viewers a way to report a missing entry before selection. Decide material roster errors consistently rather than making exceptions for friends.
  5. Close, validate, and preserve. Announce closure, finish the stated validation process, and save the final pool. Stop adding names once drawing begins.

Retiroll's Sandbox wheel offers bulk paste, spreadsheet/CSV import, and weight controls. For spreadsheet or CSV import, put a header such as Participant in the first row and names beneath it; the importer treats the first row as column headings. Review its preview and the final count. Confirm whether you are replacing or appending to an existing list so sample names or duplicate imports do not remain.

In a shared Retiroll Wheel room, the controller prepares the entry list and selects Use these options before drawing. Joining a room does not automatically add a person to that list. A room display name also does not verify a Twitch account or legal eligibility. Keep the entry and verification records appropriate to your promotion separately.

Rehearse the broadcast before the real draw

A dedicated browser window is a practical starting point: you operate the page you can see, and OBS captures that window. Avoid capturing your whole desktop if private messages, account pages, or fulfillment records might appear there. OBS's Sources Guide explains the available capture types; on recent macOS versions, window capture uses the macOS Screen Capture source.

  1. Prepare a clean draw window. Load dummy names first, close unrelated tabs, and silence notifications. Keep the winner administration window outside the captured area.
  2. Add and frame the source. In OBS, use the Sources dock's plus button and choose the appropriate window-capture source. Select the prepared window and check the crop from the audience's view.
  3. Test interaction and readability. Trigger a rehearsal draw. Confirm that the selected name and entry count remain readable in a small video player, and that the capture includes the result rather than just the animation.
  4. Check sound separately. Confirm the intended audio source reaches the output once, without duplicated desktop capture or monitoring echo. Record speech over the loudest reveal and listen back.
  5. Test the complete scene. Include the game, camera, alerts, and a realistically sized dummy roster. Recheck after switching scenes and before replacing dummy data with final entries.

An OBS Browser Source loads the page inside OBS. Treat it as a separate browser context: a URL does not automatically copy the roster or login from your ordinary browser. Verify its data and use the source's interaction window when needed. Check whether scene changes refresh or unload the page; do not discover during the drawing that its state has disappeared.

Choose its viewport for the space the draw actually occupies. A full-screen 1080p scene may justify 1920 × 1080; a smaller overlay may not. Test 30 or 60 FPS against the output you need and the resources available. OBS's performance guidance recommends minimizing browser-source dimensions and unnecessary sources.

For frozen or black browser widgets, Streamlabs recommends testing hardware acceleration on or off and restarting. This is troubleshooting, not a universal setting. Source names and audio options vary between OBS, Streamlabs Desktop, operating systems, and versions.

Lower effects enough that the winner announcement remains intelligible. An OBS limiter, placed last in the relevant filter chain, can restrain peaks; its documented defaults are −6 dB and a 60 ms release. Those are starting settings, not a guarantee about the combined stream mix. Check the final recording rather than assuming a source filter prevents all clipping.

Select provisional winners and handle claims privately

Before drawing the first key, remind viewers that entries are closed, identify which prize is next, and show the final eligible count. Select once and record the provisional winner. Explain that the account still needs to satisfy the published eligibility checks.

In the three-key example, record each selected account against its key and update the remaining pool before proceeding. If the rules allow alternates, record them in the announced order from the remaining eligible pool, keeping them distinct from prize winners. Do not introduce a fresh, undisclosed selection procedure after someone fails verification.

Give people a workable way to respond

A live acknowledgment and a final claim deadline serve different purposes. A short on-air countdown can keep a segment moving, but 90 seconds is not a universal standard or a test that distinguishes humans from bots. Stream latency, accessibility needs, blocked messages, and connection trouble can affect a legitimate winner's response.

For a planned giveaway, choose and publish a realistic deadline, notification method, and follow-up process before entry. An acknowledgment in chat can contain only an agreed phrase; eligibility and delivery details belong elsewhere. Do not reveal a must-be-present condition after selecting a quiet viewer. Within its UK advertising remit, CAP guidance expects reasonable efforts to contact winners and says one attempt is insufficient.

Verify that you are communicating with the selected account, rather than someone using a similar display name. Twitch Whispers can be blocked by recipient settings, so publish an official fallback contact route. Never ask someone to post their address, email, identity documents, or private messages in public chat.

Collect only what the prize needs: a digital key ordinarily does not require a home address. Restrict fulfillment records to the people handling them, explain their purpose, and delete information when it is no longer needed, subject to applicable recordkeeping duties. This also reflects the UK's data minimisation principle. Keep forms and submitted responses out of capture.

Handle interruptions without improvising a new result

If the broadcast freezes but the drawing recorded a result, preserve it and investigate the capture failure. Rerolling merely because the reveal was missed gives the original winner an unannounced disadvantage. If no result can be established, pause and follow the technical-failure procedure published in the rules.

When verification fails, note the specific eligibility criterion and follow the stated alternate process. When contact fails, record the attempts and deadline. These records should explain the decision without exposing private evidence to the audience. Close the loop with delivery confirmation and a public update appropriate to the rules and applicable disclosure requirements.

Keep evidence that answers useful questions

Keep a snapshot of the rules, the final eligible pool, the selection method and settings, results in prize order, and the reasons for any alternate award. A local recording of closure, the count, selection, and announcement can document what appeared on stream. Store sensitive verification and fulfillment records separately with restricted access and an appropriate retention period.

A clip can help answer “What did you announce?” It cannot establish that every valid entrant was included or that the algorithm was impartial. Within CAP's remit, a computer draw needs evidence of verifiably random results; otherwise an independent selection or supervision arrangement is required. A visible animation alone does not meet that evidentiary burden. CAP's selection guidance explains the distinction.

For specialist events needing reproducible selection, NIST describes public-randomness workflows that commit to the procedure before future random input becomes available. An ordered entrant-list commitment, a specified unbiased mapping, and preserved inputs can allow later checking. A hash alone does not validate the list; public randomness does not establish eligibility or legal compliance. This requires suitable tooling and expertise, rather than being an extra setting assumed available in a wheel.

For the three-key stream, useful records let a moderator answer concrete questions: whether an entry arrived on time, which pool was used, which account received each prize, and why an alternate was contacted. Keep enough evidence to explain those decisions without turning the public archive into a collection of participants' personal information.

A preflight check before going live

  • Rules: Entry, eligibility, prize order, repeat winners, claims, alternates, and interruptions are explained before participation.
  • Roster: The valid pool matches the records; imports, duplicates, weights, and closure have been checked.
  • Scene: Names are readable, audio has been recorded and reviewed, and private administration stays off screen.
  • Results: Someone is recording provisional winners and keeping the remaining pool accurate between drawings.
  • Contact: Winners have a published private route and fallback, with a workable deadline and verification process.
  • Records: Rules, entries, results, delivery, and necessary evidence have an appropriate storage and retention plan.

The host can then concentrate on the people watching. Announce each result clearly, explain any pause, and follow through on the prize. A well-run giveaway remains understandable after the animation ends.

Sources and research notes

Official sources were checked on 15 October 2026. Platform rules can change; revisit them before your own promotion. The examples and production checklist are editorial guidance, not a legal assessment or an audit of a particular drawing tool.

The probability examples are direct calculations under the stated assumptions. The Retiroll references describe inspected roster and presentation controls; they do not certify random selection, no-repeat behavior, or promotion compliance. Visibility can help people follow a procedure without proving its underlying implementation.

📑In this article