Gleam Giveaway Benchmarks 2026, Edition 3: User Generated Content

Ask people to send you something in a giveaway and, unless it's the only way in, very few will. We looked at every Gleam competition since May 2020 that asked entrants to upload a file, submit a photo or video, or submit a link (about one giveaway in 19), plus the ones that asked for a written comment. This report shows what a user generated content Action collects, who uses one, how it's set up, and what it costs in reach.

Source: Gleam campaign data, May 2020 to August 2026. 5,922 UGC competitions from 1,960 businesses and 7,876 comment competitions, out of 114,405 competitions with 100 or more entrants.

It covers every way a Gleam giveaway collects something an entrant made: uploads, media, links and comments, and, from a second export that carries the question texts, the open answers to Answer a Question. The question it answers is the one a marketer asks before adding the Action: if I ask people to send me something, how many will, and what else changes.

1. What we found
2. What counts as UGC, and what organisers collect under that name
3. How many giveaways ask for UGC, and which Action they use
4. How many submissions a UGC Action collects
5. Two ways to run it: UGC as the campaign, or UGC as one tile
6. Entrants resist uploads and write comments, measured inside the same campaign
7. Position in the widget doesn't move it
8. What organisers award, and why that isn't the lever it looks like
9. What a UGC giveaway costs in reach
10. Who runs UGC giveaways, by industry
11. Plan, and when in the year
12. How often businesses run UGC again
13. What to decide before you add the Action
14. Benchmark tables
15. Frequently asked questions
16. How we did this
17. Cite this research

Before any benchmark, the definition decides the number. Three defensible definitions give three answers on the same 114,405 campaigns.

Table 1 · Three definitions of a UGC giveaway, competitions with 100 or more entrants, May 2020 to August 2026

This report uses the strict definition. The broad one is dominated by X Posts, on over 15,000 campaigns, and the X Post Action usually ships with organiser-written text that the entrant posts unchanged. Calling that UGC would quadruple the headline with content nobody generated. Comments are the second tier of UGC and get their own benchmark in section 6, with their own sample, because they complete at several times the rate of uploads and belong in a different plan.

Answers to Answer a Question are the third tier and are benchmarked in the same section on a second sample, because only the featured-giveaway export carries the question text.

Inside the strict set, the Action's name says less than the title the organiser gave it. Every Action carries that title, and reading 8,285 of them sorts three quarters into a purpose.

Table 2 · What strict UGC Actions ask for, by the organiser's own title, 8,285 Actions on 5,922 campaigns

Half of the Actions Gleam files under uploads ask for content. A quarter ask for evidence. Classification from the organiser's Action title across 8,285 strict UGC Actions on 5,922 campaigns, May 2020 to August 2026. One Action in four carries a title too short to classify.

A quarter of what the platform calls an upload is verification, a third once the unreadable titles are set aside. The file is evidence that the entrant did something else: shared a post, bought the product, left a review, hit a score. That matters for everything that follows, because a proof-of-purchase Action and a photo contest behave nothing alike, and because any count of "UGC campaigns" that stops at the Action type overstates UGC by about a third. Where a section below depends on the distinction, it uses the creative-content rows only and says so.

The population export carries the title and nothing else. The featured-giveaway export described in How We Did This carries the organiser's instructions and the file types allowed as well, so the same reading was repeated on its 415 strict UGC Actions.

With the instructions in hand, 22 Actions stayed unreadable, 5%, almost all a bare "Upload a File" with nothing underneath. The other 393 sorted like this.

Table 2b · What the Action asked for once the instructions were read, featured public giveaways

The title alone placed 256 of the 415 Actions. The instructions placed 137 more. What the titles had hidden was mostly creative content, then receipts.

Table 2c · What the instructions revealed where the title said nothing, featured public giveaways

So the "title doesn't say" row in Table 2 is more of the same mix. The population split holds: about half content, a quarter to a third proof.

The configuration also says what the organiser let in.

Table 2d · File types and sources the organiser allowed, featured public giveaways

132 File Upload and 152 Media Submit Actions.

The photo is the default. Video alone is the exception, one File Upload Action in seven.

Seven campaigns, one or two per class, as entrants saw them. The action title in each is the organiser's own.

Creative content. The whole giveaway is the upload. Beginning Boutique asked for a photo with three actions on the widget, and every one of its almost 4,500 entrants uploaded. Minecraft asked for an image and a video with two actions, and drew over 17,000 entrants.

Proof of a social action, and proof of purchase. The file is evidence. Career ACTivate asked entrants to post to their Instagram Stories and upload the screenshot as the mandatory first step. NQR made a receipt showing a $30 spend the only action, and 8,612 Entrants uploaded one.

Review or rating, and proof of gameplay. AORUS asked for a screenshot of a review and a receipt on a three-action widget. Kakele asked for a store review as a Submit URL among seven actions, and almost one entrant in five sent one. Scan Computers asked for the most skilful and the funniest gameplay moment as two actions. Most of its entrants sent the first, and over half sent the second.

About one giveaway in 19 asks for UGC (5.2%), and that hasn't changed since 2021. The yearly share has stayed within a point and a half.

Source: Gleam campaign data, May 2020 to August 2026. 5,922 of 114,405 competitions with 100 or more entrants, run by 1,960 of 17,217 businesses.

Table 3 · Share of competitions with a strict UGC Action, by start year

2020 holds 1,061 campaigns from 24 May onward and is left out of the yearly table.

Table 4 · Strict UGC Actions by type, campaigns carrying each, May 2020 to August 2026

Campaigns can carry more than one Action, so the rows sum to more than 5,922. Pinterest Pins is one business and is excluded from every per-Action figure below. File Uploads and Facebook Uploads sit above the 10% dominance line, and Table 7 gives their rows with and without that business.

Two product changes sit inside these counts. Instagram Uploads, which let an entrant pick a post from their own feed, last appeared on a campaign that started on 21 November 2024. Instagram Comments arrived on a campaign that started on 24 March 2025, and it was on 758 campaigns in the first seven months of 2026. Neither moves the strict share, which is why we publish the strict share as the trend line.

This is the number you want before anything else, and it's small. We estimate submissions as the entries an Action earned divided by the entries it was set to pay, so only Actions with a fixed award count. An Action that earned nothing counts as zero (about 3% did).

Table 5 · Submissions per strict UGC Action, full distribution, May 2020 to August 2026

Half of upload Actions collected fewer than 10 submissions, and a quarter collected more than 70. Submissions per strict UGC Action, 8,278 Actions on 5,689 campaigns from 1,946 businesses, May 2020 to August 2026. Each Action is one observation.

The distribution is wide because it is two populations laid on top of each other, and the next section separates them. A quarter of Actions collecting fewer than two submissions is the half of the picture that a benchmark built only from successful photo contests would never show.

Asking for a photo or video specifically doesn't help. An optional creative-content Action in a widget of four or more Actions collected under 5 submissions at the median. Under 1% of entrants sent one. Receipts did even worse. The typical optional proof-of-purchase upload collected almost nothing, with 0.01% of entrants sending one.

Most UGC giveaways bury the ask. Over half (55%) put it in a widget of nine or more Actions. About 15% built the giveaway around it, with three Actions or fewer, and just over 5% ran it alone. UGC giveaways also ask more and run longer. The typical one had 9 Actions where a giveaway without UGC had 6, and ran 21 days rather than 14.

Table 6 · Share of entrants completing a strict UGC Action, by setup

Fixed-award Actions only, with an Action that earned nothing counted at zero. The 100% row is mechanical: when the upload is the only way in, every entrant uploaded. The rows that carry information are the middle ones.

Add two Actions beside the upload and a quarter of entrants still do it. Add eight and fewer than one in a hundred does. Median share of entrants completing an optional strict UGC Action, by number of Actions on the widget. 7,232 Actions, 1,711 businesses, May 2020 to August 2026.

The drop is steep and it happens early. With two or three Actions on the widget, a quarter of entrants uploaded. With four or five, one in thirty. From six Actions on, about one in a hundred or fewer, and the count stops falling because it has nowhere left to go. An optional upload in a long widget is competing with a dozen ways to earn an entry in one click, and it loses to all of them.

Making it mandatory is the one setting that reverses this. About one UGC giveaway in six does. On those, a typical 76% of entrants sent something in, whatever the widget length, because an entrant who declines doesn't enter. That is a product behaviour, which is why it is the one recommendation in this report that doesn't depend on comparing different businesses. It also changes what the campaign is, as section 9 shows.

Across all UGC giveaways with two or more Actions, the UGC Action brought in a median of about 2% of the campaign's entries.

Each campaign is its own control here, and the upload loses inside almost every one. In campaigns with six or more Actions and an optional upload, a typical 0.8% of entrants completed the upload. The typical other optional Action on the same widget got about 17%. The upload came in below the campaign's median Action 93% of the time. It was the least completed tile in about half of campaigns, and the most completed in only 2%.

Source: Gleam campaign data, May 2020 to August 2026. 3,886 campaigns from 1,352 businesses.

Table 7 · Median completion by optional Action type, on the same 3,886 campaigns

Optional Actions on widgets of six or more. Seventeen of the 25 Action types with 300 or more rows are shown. The eight left out are further visit, view, follow and join Actions that sit between 12 and 34%. One business holds 12% of the File Uploads rows and 18% of the Facebook Uploads rows. Without it, File Uploads reads 1.3% on 983 rows and Facebook Uploads 0.4% on 419. Viral Share completion counts the referrer's completion, which requires a referred person to enter, and isn't comparable with the one-click rows.

The four least completed Actions on the platform all ask for an upload. Median share of entrants completing each optional Action type, on 3,886 campaigns with six or more Actions and an optional UGC Action, May 2020 to August 2026.

The ranking is the whole menu laid out by effort. A visit completes at a third to two fifths. A follow at a fifth. An email at a ninth. A written comment at a tenth. A link at one in thirty. A file at well under one in a hundred. The one written-content Action, Blog Comment, completes at the level of a social follow, so the resistance is to producing media. Writing is a different ask, and that distinction is worth keeping, because "UGC" blurs it and the two should be planned differently.

Comments are UGC that entrants will actually send. A comment Action (Blog Comment, Instagram Comments, Comment Article or one of the comment imports) sat on about 7% of giveaways, a little more often than an upload. The typical one collected 34 comments, against 9 submissions for the typical upload. (The middle half ran from 10 to 114 comments.)

Source: Gleam campaign data, May 2020 to August 2026. 7,876 competitions from 1,069 businesses.

Table 7b · Share of entrants completing an optional comment Action, by widget length

Fixed-award Actions only, an Action that earned nothing counted at zero. The 2-to-3 row sits at 73 businesses, above the 50 floor, and is shown. Blog Comment is 6,167 of the 7,876 campaigns and one business holds 10.2% of them. Instagram Comments is 1,449 campaigns and one business holds 20.7%, so the Instagram row is indicative until the next edition.

Set the two tables side by side and a comment beats an upload at every widget length. With two or three Actions it's about 33% against 23%. With four or five, it's about 25% against 4%. From six Actions on, a comment still gets 5% to 10% while an upload falls to about 1% or less. From four Actions up, comments complete at seven to twelve times the rate of uploads.

A comment is a sentence typed into a box the entrant already has open. An upload is a file they have to find or make. The resistance in Table 7 is to the second, and a brand that wants words rather than pictures should plan against Table 7b.

Answer a Question is the third tier, and its name tells you the least about what it collects. The typical optional question was answered by about 23% of entrants and collected about 90 answers (13,932 competitions from 2,647 businesses carried one). That figure puts a box asking "what board game would ruin your friendships" in the same row as a box asking for a postcode. The population export only holds the Action's title, and two titles in three don't say what the question asks.

The question texts exist in a second export, the giveaways accepted into Gleam's public giveaway directory (described in How We Did This). We read every question Action in it and sorted each one by what it asks the entrant to type.

Source: Gleam featured-giveaway export, 23 October 2026. 1,187 question Actions on 692 of 3,588 featured competitions, from 224 businesses.

Table 7c · What Answer a Question asks, featured public giveaways, starts to October 2026

1,187 question Actions on 692 featured giveaways from 224 businesses, every template of the Action. Among open text boxes alone, 691 Actions, the opinion share is 32%, the contact share 25% and the quiz share 20%. The largest business holds 10% of the opinion rows. Featured giveaways are a selected sample, so these are the shares inside that sample and not a rate for Gleam campaigns in general.

One question Action in five, and one open text box in three, asks for something the entrant composes. The rest are a quiz, a form field or a poll, and the contact questions are the ones organisers make mandatory, two in five, which is the setting of a gate rather than a request for content. That's the same shape the upload tier showed in section 2. The Action's name promises UGC, and half the configurations collect something else.

Where a featured giveaway is also in the population, its completion can be read.

Table 7d · Completion of an optional Answer a Question Action by what it asks, featured giveaways in the population

Optional, fixed-award Actions only. The median open question collected about 450 answers. Set mandatory, the open question completed at 89% on 42 Actions, which is the gate of section 5 again.

On the same featured campaigns, with six or more Actions on the widget, the gap holds. About 15% of entrants answered an optional open question and about 7% left a comment. Only 0.3% uploaded.

The open-question cells sit below the 200-campaign floor this series uses for a benchmark, so they are the direction and not the number. The direction is the one the comment benchmark already gave. Typing completes, uploading doesn't, and a question that invites a sentence sits with the comment.

Table 8 · Completion of an optional strict UGC Action by its position, widgets of six or more Actions

Moving an upload up the widget did nothing. Moving a follow up doubled it. Median share of entrants completing an optional upload and an optional X Follow by tile position, widgets of six or more Actions. 6,069 upload Actions from 1,369 businesses and 68,266 X Follow Actions, May 2020 to August 2026.

An X Follow on the same kind of widget shows what position can do. In the first tile, about 45% of entrants followed. By the seventh tile or later, that fell to about 20%. So position is a strong lever for a one-click Action and no lever for an upload. The upload moved less than half a point across the whole widget, while the follow moved about 25 points. Inside each widget-length band the upload figures move within a point or two, with no pattern that survives.

Organisers do bury the Action. About half (52%) of UGC Actions sit in the bottom half of their widget. Moving them up isn't the fix the data supports. An entrant who won't take a photo won't take one because the request came first.

They pay more. On UGC giveaways with two or more Actions, the typical UGC Action was worth 5 entries, where the typical other Action on the same widget was worth 2. It paid more than the others in about six campaigns in ten.

Table 9 · Completion of an optional strict UGC Action by its award relative to the other Actions on the widget, widgets of four or more

Read as a cause, this table says that paying for uploads suppresses them, which is almost certainly wrong. Read as a choice, it says what organisers already know. The award is set before the campaign runs, by someone who can see how hard the task is.

A video of yourself using the product gets 10 entries. A selfie gets 2. The flat widgets where every Action pays one entry are the ones where the upload was easy enough not to need a premium. The award is a thermometer for difficulty, not a lever on it, and this data can't separate the two.

The comparison is held inside plan and inside widget length, because UGC giveaways run on higher plans with longer widgets. These are different businesses, and the table describes what they achieved.

Table 10 · Median entrants, competitions with and without a strict UGC Action, paid plans, inside widget-length bands

Every row clears 200 UGC campaigns and 50 businesses. The no-UGC side excludes campaigns with any UGC-family Action, comments and posts included. Differences under 5% are shown as none.

The fewer the other Actions, the more reach a UGC giveaway gives up. Median entrants with and without a strict UGC Action, Business plan, inside widget-length bands. 2,683 UGC campaigns and 20,181 without, May 2020 to August 2026.

Where the upload is the campaign, the campaign reached about a third of the entrants. On the Business plan the gap closes at nine to twelve Actions, and with thirteen or more the UGC giveaways were slightly larger. Entries per entrant were lower with UGC in every band, which is section 6 seen from the other side: Entrants skip the upload and move on.

The exchange rate is the planning number. On the Business plan, a UGC giveaway with two or three Actions reached 380 entrants at the median, where a comparable giveaway without UGC reached about 1,100. Almost a quarter of those 380 uploaded (Table 6). That's roughly 90 pieces of content for about 720 entrants given up. A giveaway with thirteen or more Actions gave up no entrants, and its upload collected a median of 6 submissions. Everything in between is your call.

Two other differences travel with UGC. First, fewer of the people who saw the widget entered. In the one conversion cell that clears every threshold, about 20% of widget views turned into entrants with UGC, against about 27% without. That's a gap of 6.5 points.

Source: Gleam campaign data, May 2020 to August 2026. Business and Pro plans, 9 to 12 Actions, 14 days or fewer, no repeatable Action. 245 UGC campaigns from 93 businesses.

Second, UGC giveaways finished less often. About 74% drew winners, against about 81% of the rest. That fits a contest that needs judging, which takes longer to close and sometimes never does.

Industry is a site-level label from the 21 verticals on gleam.io/industries. It covers most UGC campaigns (about 84%) but only about 61% of all campaigns, so this section reports where UGC giveaways come from. A rate per industry waits on the next labelling pass (How We Did This).

Table 11 · Strict UGC campaigns by industry, against each industry's share of all labelled campaigns

4,957 labelled UGC campaigns. Music (207 campaigns, 65% from one business), Cryptocurrency (139, excluded as a segment), Pets, Automotive, Events, SaaS, Wine and Spirits, Apps, Baby and Podcasting fall below the floor or the dominance rule and aren't shown. B2B and Board Games each carry one business above 25% of the industry's UGC campaigns and are shown with that warning. Every row below Smart Wear and Gadgets sits under the 75-business floor for an industry cut and is indicative. Other, the catch-all label (297 campaigns from 85 businesses), isn't a segment and is left out.

Creators run 30% of labelled giveaways but only 15% of UGC giveaways, so the largest segment on Gleam uses UGC at half its weight. B2B, Board Games, Hobbies and Gadgets brands use it above their weight. The clearer pattern is in what each asks for.

Table 12 · What the UGC Action asks for, by industry, strict UGC Action rows on labelled businesses

Sports asks for the photo four times in five, apparel three times in four. Creators and B2B split their UGC Actions between content and proof that the entrant shared something, which for a creator is the audience and for a B2B brand is the reach. Brick and Mortar is the one industry where one upload in five is a receipt. E-commerce does a little of everything. So the industry answer is less "who uses UGC" than "what UGC means where you are", and a benchmark for an apparel brand's photo contest hasn'thing to say to a retailer's receipt upload.

The same reading of Answer a Question texts, on the featured giveaways where the text exists, gives the industry view of the third tier.

Table 12b · What Answer a Question asks, by industry, featured public giveaways, question Action rows on labelled businesses

1,128 of 1,187 question Actions sit on labelled businesses, 95%. Cells of 30 Actions or more. RPG and Board Games rests on 11 businesses and Brick and Mortar on 8, below the 75-business floor, and both are indicative. The order-check column is omitted at 2% or less in every row.

Creators ask for an opinion one time in three, the highest share, and a quiz answer as often. Gadget brands ask for a profile first, which is the product-interest question ("which features would level up your chair"). E-commerce and retail use the box as a quiz, almost always a fact from a product page, which is a visit with a check on it. Board game brands use it as a membership check. B2B asks for a profile or a contact field two times in three, which is lead capture with a prize attached.

So the industry question for the third tier is the same as for uploads. It is what each industry uses the Action for, and creators are the one industry that uses it mainly for words.

From the Action Types table on gleam.io/pricing: File Upload is available on Business and Premium. Submit Media is available from Pro, with Facebook, Instagram and TikTok as sources, and gains the File Upload source on Business. Submit URL is on every plan. Instagram hashtag and mention imports are Business and above, and the X hashtag import is Pro and above.

Table 13 · Share of competitions with a strict UGC Action, by plan

Free is the export's own plan label for campaigns run on a no-cost tier. The current pricing page doesn't list it as a plan, and its campaigns appear in every year of the window. Its only strict UGC Action is Submit URL, on 240 campaigns from 141 businesses, which is the one Action every plan has. Premium sits below the 50-business floor for a benchmark and is shown for completeness.

Business campaigns asked for UGC two and a half times as often as Pro campaigns, and part of that is access: the File Upload Action and the upload source don't exist below Business. Submit URL is the one strict UGC Action every plan has, so it is the only one on which a cross-plan comparison is fair. The Pro rows in Tables 10 and 13 come from the UGC Actions Pro can run, such as Submit Media and Submit URL, so only the File Upload figures are Business-and-above benchmarks.

There's no UGC season. By start quarter, from 2021 to 2025, the share of giveaways asking for UGC stayed between about 4.9% and 5.5%. By month, averaging each year's own index, July ran highest, at about 1.2 times the yearly share. December ran lowest, at about three quarters of it. Every other month sat within about 12% of the yearly share. December is the one dip, and it explains itself.

Table 14 · December starts against the other eleven months, 2021 to 2025

December giveaways are short, follow-me giveaways: one-click follows on the platforms where gift-season audiences are, fewer list-building and referral Actions, and fewer uploads. A UGC giveaway runs three weeks at the median and asks for effort, and December is the month neither is on offer. The two facts are one fact: effort Actions lose share when campaigns get short.

The businesses behind UGC giveaways aren't alike. About two in three ran exactly one UGC giveaway. At the other end, a small group of heavy users (under 5% of UGC businesses) ran over 40% of all UGC giveaways. The typical business that has used UGC at all puts it on 60% of its giveaways.

Count only the businesses that ran another giveaway afterwards, and most came back to UGC (62%). They typically used it on about one later giveaway in six. UGC is also often why a business arrived. Over three in five UGC businesses used it on their very first giveaway in the data, so it's more often the reason a business came than a tactic it grew into.

Source: Gleam campaign data, May 2020 to August 2026. 5,922 UGC giveaways from 1,960 businesses, of which 1,256 ran UGC once and 86 ran 2,508 UGC giveaways. Return rate from 956 businesses whose first UGC giveaway ran from 2021 to 2024 and who launched again afterwards.

Start here

Half of the Actions Gleam files under uploads collect content and a quarter collect proof. If you want proof of a purchase or a share, the Action works as a gate and nothing in this report about volume applies. If you want words, ask a question or a comment. About 17% of entrants answered an optional open question, and a comment got 5% to 33% depending on widget length. On the same widget lengths an upload got anywhere from under 1% to about 23%. If you want photos, read on.

Biggest measured setting

Make it mandatory and a typical 76% of entrants send something in. Leave it optional with two other Actions beside it and almost a quarter do. Put eight others beside it and fewer than one in a hundred do. The widget length is the one setting that moved completion by more than a few points, and it moved it by a factor of thirty.

The trade

A UGC-first giveaway on the Business plan reached about a third of the entrants of a comparable giveaway without one (380 against about 1,100 at the median). Almost a quarter of them uploaded. If 90 pieces of content are worth 720 Entrants to you, run it that way. If they aren't, add the Action as a bonus, expect a handful of submissions, and size the prize for the entries.

Smallest measured settings

Position made no measured difference to upload completion. The award tracked how hard the task was rather than moving it. Both are an hour of fiddling with no return.

Timing

Not because uploads fail in December, which this data can't show, but because the December giveaway is a 10-day follow-me campaign and a UGC giveaway is a 21-day effort campaign. July ran highest, at 1.2 times the yearly share.

One thing this study can't tell you is what arrived. It counts submissions and can't see them, so nothing here speaks to whether the photos were usable, on brand or cleared for reuse. The seven widgets in section 2 show what each kind of ask looks like to an entrant. What arrived in response isn't in this data.

These are the canonical tables for this edition, restated next edition on the same definitions.

Table 15 · Submissions per strict UGC Action, planning ranges

Plan against the P25 figure if you need a number you are confident of clearing, and against the median if you are setting an expectation. The 9-or-more row pools the 9-to-12 and 13-plus bands.

Table 16 · Strict UGC adoption, canonical

Everything above rests on the decisions in this section. It sits at the end so the findings come first, and it is written to be checked.

Every figure covers competitions started between 24 May 2020 and 12 August 2026 that had ended by 13 August 2026, from the export Gleam engineering supplied in September 2026. The window is multi-year, unlike Editions 1 and 2, because the trailing twelve months hold 692 UGC campaigns from 306 businesses, enough for the headline rate and not enough for any per-Action or per-setup cell. The trailing-twelve-month adoption rate is given in Table 16. Every table carries its own window.

A second export supplies what the first lacks. On 23 October 2026 Gleam engineering exported the 3,588 competitions from 1,027 businesses accepted into the public giveaway directory at gleam.io/giveaways, with every Action's configuration: the title, the instructions, the file types and the question text. 2,970 of them are also in the population, which is how their completion figures are read. Nearly all started in 2024 or later. This sample is selected twice, once when the organiser submitted the giveaway and once when Gleam accepted it, so nothing from it is a rate for Gleam campaigns in general.

It is used for three things: what Answer a Question asks and how those classes completed (Tables 7c and 12b, and the same-widget comparison beside them), a check of the title-based purpose classification against the instructions, and the file types organisers allowed (section 2). Every figure drawn from it says "featured public giveaways". The texts were sorted by a reader, not a keyword list, into the classes shown, and the labelled files sit beside the scripts in the research folder.

• The population holds 114,405 competitions from 17,217 businesses with 100 or more entrants, a duration of 0 to 365 days and an end date on or before 13 August 2026.
• Campaigns from cryptocurrency and Web3 organisers are excluded, on any of: a wallet-address Action, a cryptocurrency prize name, a cryptocurrency campaign name, 20% or more of the business's campaigns flagged, or 50% or more carrying a Telegram Action. The industry label catches a further 139 UGC campaigns from 27 crypto businesses the behavioural filter misses, and they are excluded from Table 11 and named here.
• Campaigns where every Action required a purchase are excluded.
• Instagram Comments, Blog Comment, Comment Article and the comment imports form the second tier and are benchmarked in section 6 on their own sample of 7,876 campaigns. They are never pooled with uploads in a single figure.
• Answer a Question is the third tier. Its texts aren't in the population export, so its split and completion are read on the featured-giveaway sample below and labelled as such wherever they appear.
• Completion and submission figures use the seven strict Action types with a fixed entry award, 8,278 Actions. The export records no count for an Action that earned no entries, so those Actions are counted at zero wherever the campaign has a count on any other Action. Seven Actions on campaigns with no counts at all are left out.
• Any cell below 200 campaigns or 50 businesses is suppressed or shown with a warning. Industry cuts use 300 campaigns and 75 businesses. Action-type rows use 300 Action rows.
• Where one business held more than 10% of a cell's campaigns, the cell is shown with that warning or recounted without the business and the caption says so. Across all 5,922 UGC campaigns the largest business holds 3.3% and the top three 7.6%.
• Conversion (Entrants per widget view) is reported for one cell only, restricted to campaigns of 14 days or fewer with no repeatable Action, because Gleam counts widget views once per visitor per day and a longer campaign accumulates views from the same people.

Submissions per Action run from zero to the tens of thousands. Half of Actions collected fewer than 10 and one in ten collected more than 389. An average built on that distribution describes an Action nobody ran. Every headline figure is a median, the benchmark tables carry the 25th and 75th percentiles, and every cell carries its campaign and business count.

The with-against-without question in section 9 is held inside plan and widget length. The resistance question in section 6 is held inside the campaign, so that every campaign is its own control. The position question in section 7 is run inside widget-length bands and against a one-click control Action. The award question in section 8 is held inside widgets of four or more Actions and reported as a choice rather than a cause. The seasonality question in section 11 averages each year's own monthly index so no single year dominates.

• It covers Gleam competitions only, run by businesses that chose Gleam. It doesn't cover Gleam Galleries, the product built for collecting and displaying UGC, which is a separate campaign type outside this export.
• It carries no causal design. Nothing here shows that adding, moving, paying for or mandating an upload causes an outcome.
• It doesn't hold the submissions. It can't say whether a photo was usable, on brand, cleared for reuse or ever reused.
• The population export holds the Action title and not the instructions, the file types allowed or the question text, and one title in four is too short to classify. Those fields exist only for the featured-giveaway sample, which is why the question benchmark and the instruction check are drawn from it and labelled.
• It doesn't hold the domains of submitted URLs, so nothing here says what entrants linked to.
• Its industry labels are unaudited. 2,849 businesses were classified from their domain for Edition 1, 556 from their public website for Edition 2, and 400 UGC businesses from site name, domain and prize names for this edition, of which 279 were labelled with high confidence, 112 medium and 9 low. Labels cover 61% of all campaigns, so this edition reports industry composition and not an industry rate.
• It doesn't measure revenue, sales or content value. Submissions are submissions, and this study doesn't claim they are assets.

Seven campaigns are shown in section 2 as examples of each kind of ask: Beginning Boutique, Minecraft, Career ACTivate, NQR, AORUS, Kakele and Scan Computers.

Edition 3, v1.1, 28 October 2026: corrected every completion and submission figure after the data-owner check found that an Action with no entries was being read as fully completed rather than as zero, and that two Action types outside the strict definition had been counted in the per-Action cells. The headline median moved from 11 to 9 submissions, the two-or-three-Action completion from 31.5 to 23.1% and the mandatory completion from 79 to 76%, and Table 12 was rebuilt on the current industry labels. v1.0, 28 October 2026: added the Answer a Question benchmark (Tables 7c and 12b) and the instruction check in section 2 from the featured-giveaway export of 23 October 2026, replaced the Instagram Action dates with the first and last campaign starts in the data, and described the Free plan row.

Draft v0.1, 19 October 2026. Corrections to this page will be published as a new version with the change named. Numbers are never altered silently.

According to the Gleam Giveaway Benchmarks 2026, Edition 3, half of the upload, media and submit-URL Actions on Gleam giveaways collected fewer than 10 submissions, based on 8,278 Actions across 5,689 competitions run by 1,946 businesses between May 2020 and August 2026.

We didn't benchmark these figures against a published industry average for UGC campaigns. The UGC statistics we found measure consumer trust in UGC or engagement on branded hashtags, and none state a denominator that matches submissions on a giveaway.