Skip to content

Crowd9/giveaway-skills / skills/gleam-campaign-setup

gleam-campaign-setup

Verified

Map a giveaway plan to documented Gleam Competitions settings. Use for 'set this up in Gleam', 'fraud setting', 'Gleam draw', 'Gleam impressions', 'Gleam terms', 'mandatory action', 'daily entries', 'free entry alternative', 'export entries', 'Gleam on Shopify', 'Quick Draws', 'repeat Winners', or 'admin entries'. Cover setup, operation and reporting.

FileKindSize
SKILL.mdskill md13.3 KB
evals/cases.mdmarkdown5.4 KB
evals/evals.jsonconfig21.7 KB
references/campaign-setup.mdmarkdown8.2 KB
references/drawing-winners.mdmarkdown3.6 KB
references/evidence-detail.mdmarkdown3.2 KB
references/house-style.mdmarkdown4.7 KB
references/reporting-and-fraud.mdmarkdown5.9 KB
references/settings-evidence.mdmarkdown33.8 KB
references/shopify.mdmarkdown4.2 KB
references/tips-from-gleam.mdmarkdown2.1 KB
scripts/style_check.pyexecutablescript40.3 KB

references/settings-evidence.md

View on GitHub
# What the data says about Gleam settings

Extracted from the campaigns we could compare fairly in the dataset: campaigns that reached 100 or more unique Entrants, crypto and purchase-only campaigns removed. (116,499 campaigns.)

Action-completion figures below count completion events per 100 Entrants for each offered Action, then take the median across those Actions. A campaign offering several matching Actions contributes several observations. The counts show offering campaigns, while the number of Action observations is not supplied. For comparisons after summing matching completions per campaign, use the asset and per-action percentile tables in the results-review skill. A figure of 51 means 51 completion events per 100 Entrants, with the underlying ratio in brackets. Repeatable actions can exceed 100. Percentage labels in the tables also describe these event ratios. They do not identify distinct people, unique subscribers or mailing permission. Use individual consent records to determine which addresses may be mailed. Missing completion-record and consent-record coverage is not reported in these aggregates.

Action names are the dataset's generic names, which are history. Check the current name on the [How to Enter page](https://gleam.io/docs/competitions/setup/how-to-enter) before quoting one. Every figure describes what businesses chose and what their Entrants did. No comparison group of failed campaigns exists, so none of it shows cause. ()

## Starting point (extracted)

From the campaign analysis's campaign-history match, covering campaigns with 100 or more Entrants, finance and crypto organizers excluded, an organizer's first campaign in the dataset more often starts blank than from a template.

<!-- generated:se2_start_source -->
| Starting point (organizer's first campaign) | Share |
|---|---|
| Gleam library template | 31% |
| Blank | 53% |
| Copied from outside the dataset | 16% |

(17,633 first campaigns, holding out 229 tagged as a copy of the organizer's own earlier campaign, a data anomaly.)
<!-- /generated -->

Templates in use are 68% Pro-tier across every tier measured. (9,955 campaigns, 6,467 businesses.)

Template use has fallen since 2020 while own-copy use has risen sharply, over the same years:

<!-- generated:se2_start_year -->
| Year | Template use | Own-copy use | Campaigns | Businesses |
|---|---|---|---|---|
| 2020 | 16% | 23% | 1,159 | 718 |
| 2026 | 6% | 66% | 7,992 | 2,233 |
<!-- /generated -->

By organizer tenure, a sixth-or-later campaign built from a template converts better than a sixth-or-later own copy:

<!-- generated:se2_start_tenure -->
| Sixth-or-later campaign, source | Conversion Rate | Campaigns | Businesses |
|---|---|---|---|
| Built from a template | 33.3% | 2,060 | 625 |
| Own copy | 26.6% | 56,179 | 2,701 |
<!-- /generated -->

Source: `analysis/output/templates.json` (`source_mix_first_campaign`, `template_tier_mix`, `template_share_by_year`, `template_by_organizer_tenure`).

## Completion events per 100 Entrants for each Gleam action

Actions offered by at least 300 campaigns, in order of how many campaigns offered them.

<!-- generated:se2_uptake_actions -->
| Action | Offering campaigns | Median across offered Actions, completion events per 100 Entrants | Middle half of offered Actions, events per 100 Entrants |
|---|---|---|---|
| X Follows | 65,714 | 54 (0.54) | 39 to 70 (0.39 to 0.70) |
| Visit a Page | 62,783 | 78 (0.78) | 56 to 100 (0.56 to 1.00) |
| Instagram Profile Visits | 59,033 | 82 (0.82) | 68 to 92 (0.68 to 0.92) |
| Facebook visits | 46,607 | 82 (0.82) | 65 to 92 (0.65 to 0.92) |
| YouTube Channel Visits | 44,616 | 87 (0.87) | 71 to 97 (0.71 to 0.97) |
| Viral Shares | 42,536 | 11 (0.11) | 6 to 23 (0.06 to 0.23) |
| Bonus | 39,177 | 103 (1.03) | 61 to 201 (0.61 to 2.01) |
| Email Subscriptions | 38,390 | 85 (0.85) | 72 to 101 (0.72 to 1.01) |
| X Reposts | 27,411 | 43 (0.43) | 31 to 59 (0.31 to 0.59) |
| Custom Actions | 22,743 | 50 (0.50) | 26 to 80 (0.26 to 0.80) |
| Twitch Follows | 22,490 | 68 (0.68) | 48 to 88 (0.48 to 0.88) |
| Chat Members | 18,890 | 45 (0.45) | 29 to 64 (0.29 to 0.64) |
| X Posts | 15,413 | 34 (0.34) | 25 to 43 (0.25 to 0.43) |
| TikTok Follows | 14,295 | 35 (0.35) | 25 to 49 (0.25 to 0.49) |
| Answer a Question | 13,847 | 73 (0.73) | 40 to 100 (0.40 to 1.00) |
| Secret Code | 8,935 | 27 (0.27) | 9 to 57 (0.09 to 0.57) |
| Instagram Follows | 8,208 | 51 (0.51) | 40 to 67 (0.40 to 0.67) |
| Pinterest Visits | 7,859 | 66 (0.66) | 51 to 76 (0.51 to 0.76) |
| Facebook Entries | 7,567 | 54 (0.54) | 41 to 68 (0.41 to 0.68) |
| Twitch Subscribers | 7,366 | 11 (0.11) | 3 to 36 (0.03 to 0.36) |
| Instagram Post Views | 7,182 | 74 (0.74) | 58 to 84 (0.58 to 0.84) |
| TikTok Video Views | 6,731 | 53 (0.53) | 38 to 68 (0.38 to 0.68) |
| Loyalty Bonuses | 6,101 | 56 (0.56) | 34 to 72 (0.34 to 0.72) |
| Blog Comment | 6,097 | 22 (0.22) | 13 to 41 (0.13 to 0.41) |
| Single Choice List | 4,838 | 77 (0.77) | 57 to 101 (0.57 to 1.01) |
| Facebook Post Views | 4,056 | 71 (0.71) | 52 to 83 (0.52 to 0.83) |
| Facebook Likes | 3,960 | 39 (0.39) | 30 to 47 (0.30 to 0.47) |
| Telegram Channel Members | 3,916 | 85 (0.85) | 59 to 96 (0.59 to 0.96) |
| Instagram Entries | 3,800 | 53 (0.53) | 38 to 67 (0.38 to 0.67) |
| YouTube Entries | 3,492 | 81 (0.81) | 53 to 97 (0.53 to 0.97) |
| App Downloads | 3,463 | 41 (0.41) | 28 to 62 (0.28 to 0.62) |
| Group Members | 3,323 | 54 (0.54) | 36 to 66 (0.36 to 0.66) |
| Reddit Visits | 2,735 | 77 (0.77) | 58 to 88 (0.58 to 0.88) |
| X Entries | 2,611 | 59 (0.59) | 42 to 78 (0.42 to 0.78) |
| Podcast Subscriptions | 2,487 | 35 (0.35) | 28 to 44 (0.28 to 0.44) |
| Media Submits | 2,185 | 10 (0.10) | 3 to 65 (0.03 to 0.65) |
| Multiple Choice Checkboxes | 2,182 | 82 (0.82) | 59 to 101 (0.59 to 1.01) |
| Spotify follows | 2,182 | 32 (0.32) | 22 to 48 (0.22 to 0.48) |
| Goodreads Follows | 2,055 | 48 (0.48) | 41 to 52 (0.41 to 0.52) |
| Twitch Channel Rewards | 2,009 | 48 (0.48) | 12 to 138 (0.12 to 1.38) |
| Bluesky Follows | 1,909 | 29 (0.29) | 20 to 41 (0.20 to 0.41) |
| LinkedIn Follow | 1,720 | 28 (0.28) | 21 to 39 (0.21 to 0.39) |
| Pinterest Entries | 1,605 | 40 (0.40) | 26 to 48 (0.26 to 0.48) |
| Submit URL | 1,601 | 19 (0.19) | 8 to 43 (0.08 to 0.43) |
| Instagram Comments | 1,555 | 35 (0.35) | 27 to 48 (0.27 to 0.48) |
| Patreon Visits | 1,521 | 44 (0.44) | 8 to 73 (0.08 to 0.73) |
| Twitch Entries | 1,490 | 75 (0.75) | 54 to 96 (0.54 to 0.96) |
| X Hashtag Posts | 1,471 | 36 (0.36) | 23 to 56 (0.23 to 0.56) |
| X Post Views | 1,444 | 66 (0.66) | 51 to 83 (0.51 to 0.83) |
| Threads Follows | 1,216 | 24 (0.24) | 19 to 31 (0.19 to 0.31) |
| File Uploads | 1,213 | 17 (0.17) | 4 to 70 (0.04 to 0.70) |
| Promote Entries | 1,183 | 49 (0.49) | 37 to 57 (0.37 to 0.57) |
| RSS Feed | 1,171 | 50 (0.50) | 25 to 61 (0.25 to 0.61) |
| Gleam Subscriber | 1,160 | 62 (0.62) | 52 to 71 (0.52 to 0.71) |
| Snapchat | 981 | 21 (0.21) | 15 to 34 (0.15 to 0.34) |
| Spotify listen | 923 | 58 (0.58) | 34 to 123 (0.34 to 1.23) |
| Facebook Join Group | 879 | 35 (0.35) | 31 to 41 (0.31 to 0.41) |
| Kick Follow | 748 | 45 (0.45) | 32 to 61 (0.32 to 0.61) |
| csv_import | 723 | 5 (0.05) | 0 to 79 (0.00 to 0.79) |
| Choose Image | 675 | 97 (0.97) | 82 to 102 (0.82 to 1.02) |
| Discord Entries | 669 | 61 (0.61) | 42 to 81 (0.42 to 0.81) |
| Bluesky Reposts | 665 | 25 (0.25) | 19 to 33 (0.19 to 0.33) |
| SoundCloud Follows | 665 | 21 (0.21) | 12 to 32 (0.12 to 0.32) |
| Steam Entries | 603 | 70 (0.70) | 46 to 91 (0.46 to 0.91) |
| Instagram Uploads | 595 | 3 (0.03) | 1 to 6 (0.01 to 0.06) |
| Game Players | 573 | 11 (0.11) | 3 to 29 (0.03 to 0.29) |
| Crypto Wallet Address | 532 | 110 (1.10) | 88 to 145 (0.88 to 1.45) |
| Watch a Video | 496 | 47 (0.47) | 35 to 70 (0.35 to 0.70) |
| Facebook Uploads | 493 | 3 (0.03) | 1 to 6 (0.01 to 0.06) |
| Spotify saves | 485 | 41 (0.41) | 25 to 64 (0.25 to 0.64) |
| Kickstarter visits | 433 | 80 (0.80) | 59 to 96 (0.59 to 0.96) |
| Blog Post | 397 | 1 (0.01) | 0 to 2 (0.00 to 0.02) |
| Tumblr Follows | 374 | 16 (0.16) | 8 to 23 (0.08 to 0.23) |
| LinkedIn Share | 306 | 4 (0.04) | 2 to 8 (0.02 to 0.08) |
| Facebook check-ins | 300 | 29 (0.29) | 13 to 42 (0.13 to 0.42) |
<!-- /generated -->

Source: `analysis/output/gleam_settings.json` key `uptake_by_gleam_action`.

Reading it: visits and email actions have relatively high completion-event ratios. Viral Shares record referral completion events, without multiplying by Entry worth. These totals cannot identify the number of distinct referrers. Uploads, blog posts and paid Twitch subscriptions sit among the lowest ratios. Bonus exceeds one completion per Entrant on average. None of these ratios identifies the number of people who completed an action.

## Position in the action list

Typical completion events per 100 Entrants, by the action's list position, split by family. The table's percentage labels describe event ratios, not the percentage of distinct people who completed an action. Businesses put the action they care about first, and Gleam shows actions in list order, so this mixes ordering with choice. Carry that into the sentence that uses the table. "Position does most of the work here" is the claim this cut cannot make, because the action a business put first is the one it cared about most, and a caveat added at the end of the answer does not undo a causal sentence in the body. Measured answers made this exact claim twice.

<!-- generated:se_position -->
| Family | 1st, % of Entrants | 2nd to 4th, % of Entrants | 5th and later, % of Entrants |
|---|---|---|---|
| Email signup | 101 (1.01, 20,986 Actions) | 78 (0.78, 10,182 Actions) | 73 (0.73, 16,859 Actions) |
| Visit or view | 98 (0.98, 14,449 Actions) | 86 (0.86, 84,512 Actions) | 75 (0.75, 132,797 Actions) |
| Follow, subscribe, join | 82 (0.82, 20,468 Actions) | 56 (0.56, 75,063 Actions) | 42 (0.42, 116,619 Actions) |
| Viral Share, refer | 16 (0.16, 993 Actions) | 12 (0.12, 11,344 Actions) | 11 (0.11, 30,470 Actions) |
| Content upload | 81 (0.81, 2,573 Actions) | 27 (0.27, 4,657 Actions) | 13 (0.13, 6,599 Actions) |
<!-- /generated -->

Every family has a lower completion-event ratio in later positions in this comparison, Viral Share included. Use the family row for the proposed position. These ratios do not identify how many distinct Entrants completed an action, and the association does not establish the effect of moving it. Source: `analysis/output/field_cuts.json` `uptake_by_family_and_position`.

## Description length

Words in the campaign description, among the campaigns we can compare fairly: no repeatable action, 14 days or less.

<!-- generated:gs_desc_length -->
| Words | Campaigns | Businesses | Entrants | Conversion Rate | Entries per Entrant |
|---|---|---|---|---|---|
| 0-25 | 14,787 | 2,954 | 420 | 39% | 3.46 |
| 26-75 | 14,941 | 3,882 | 421 | 32% | 3.81 |
| 76-150 | 7,151 | 2,027 | 589 | 33% | 3.94 |
| 151-300 | 2,355 | 633 | 623 | 33% | 3.78 |
| 301+ | 268 | 78 | 330 | 29% | 4.15 |
<!-- /generated -->

No description and short descriptions sit with the simplest campaigns, which also see the highest share of viewers enter, so the 39% says as much about the campaign as the copy. Longer descriptions run with more actions and more Entries per Entrant. Source: `analysis/output/gleam_settings.json` `description_length_clean`.

## Custom Actions templates

<!-- generated:se2_custom_templates -->
| Template | Actions | Typical completion events per 100 Entrants |
|---|---|---|
| visit | 158,511 | 78 (0.78) |
| none | 52,187 | 50 (0.50) |
| bonus | 44,260 | 103 (1.03) |
| question | 19,233 | 73 (0.73) |
| choose_option | 7,924 | 77 (0.77) |
| multiple_choice | 2,756 | 82 (0.82) |
| choose_image | 972 | 97 (0.97) |
<!-- /generated -->

Custom Actions with no template (a free-form instruction) recorded about 50 completion events per 100 Entrants, the lowest ratio of any template. Question, single-choice and visit templates sit in the seventies, multiple choice and choose image higher, and bonus above 100 because a bonus can repeat (see table above).

## Throwaway account restriction

Actions with the throwaway account restriction enabled recorded a lower completion-event ratio:

<!-- generated:se2_throwaway -->
| Throwaway account restriction | Actions | Typical completion, % of Entrants |
|---|---|---|
| On | 7,637 | 38% (0.38) |
| Off | 138,267 | 49% (0.49) |
<!-- /generated -->

The groups used different campaign configurations. The comparison does not establish the effect of enabling the restriction or identify which rejected accounts belonged to real people.

## Visit action options

Visit actions can complete on click, after a delay, or after a question about the page. Actions completing on click recorded the highest completion-event ratio, and actions with a post-visit question recorded the lowest.

<!-- generated:se2_visit_mode -->
| Completion option | Typical completion, % of Entrants | Actions |
|---|---|---|
| Complete on click | 84% (0.84) | 175,741 |
| After a delay | 77% (0.77) | 9,332 |
| After a question | 71% (0.71) | 17,184 |
| Option unset | 78% (0.78) | 164,426 |
<!-- /generated -->

Among Visit a Page link options, plain links recorded a higher completion-event ratio than pages embedded as HTML inside the widget. Open Graph previews recorded the highest ratio.

<!-- generated:se2_visit_link -->
| Link option | Typical completion, % of Entrants | Actions |
|---|---|---|
| Plain link | 79% (0.79) | 138,017 |
| Page content as HTML in the widget | 57% (0.57) | 15,254 |
| Open Graph preview | 86% (0.86) | 5,240 |
<!-- /generated -->

Nearly every Viral Share action carried custom share text, not the default, so the dataset cannot compare the two. (41,140 Actions with custom text against 1,667 with the default, from `config_cuts` `share_text_length`.)

## Completion-event ratios by setting

Read from the action's own configuration. The count is Actions, one row per Action offered, so it runs above the campaign population and is never a campaign count. Configurations with fewer choices or shorter waits generally recorded higher completion-event ratios. X throwaway account restriction matches the direction of the per-action figure in Throwaway account restriction above. These comparisons describe different campaigns and do not predict the result of changing a setting.

<!-- generated:se_config -->
| Setting | Option | Completion, % of Entrants (decimal) | Actions | Businesses |
|---|---|---|---|---|
| Email opt-in checkbox | Off | 92% (0.92) | 29,510 | 4,506 |
| Email opt-in checkbox | On | 85% (0.85) | 9,764 | 1,618 |
| Email opt-in checkbox | Auto | 78% (0.78) | 7,584 | 326 |
| Newsletter description | None | 95% (0.95) | 20,856 | 3,554 |
| Newsletter description | Long | 81% (0.81) | 23,680 | 2,744 |
| Newsletter description | Short | 81% (0.81) | 2,322 | 539 |
| Instagram visit delay | 5 seconds or less | 83% (0.83) | 192,023 | 12,912 |
| Instagram visit delay | 6 to 15 seconds | 76% (0.77) | 1,672 | 299 |
| Instagram visit delay | 16 seconds or more | 60% (0.60) | 612 | 87 |
| X throwaway account restriction | Open | 49% (0.49) | 134,484 | 10,202 |
| X throwaway account restriction | Restricted | 38% (0.38) | 7,355 | 839 |
| Share text length | 60 to 140 characters | 12% (0.12) | 30,789 | 5,447 |
| Share text length | Over 140 characters | 9% (0.09) | 10,075 | 1,867 |
| Share text length | Default text | 11% (0.11) | 1,667 | 451 |
| Secret Code length | 4 to 13 characters (open-ended) | 19 to 34% | - | - |
| Secret Code length | 8 characters, fixed-length pattern | 9% (0.09) | 449 | 92 |
| Actions required | 1 to 7 (range) | 44 to 102% | - | - |
| Actions required | 7 (exact) | 49% (0.49) | 2,356 | 912 |
| Actions required | 8 | 27% (0.27) | 1,691 | 440 |
<!-- /generated -->

Actions required is the exception worth flagging on its own: completion falls away as the number required climbs, and drops by roughly half again between 7 and 8.

Email opt-in checkbox on or set to auto, by organizer country. Poland has the highest published share in this comparison. The Malaysia share is withheld because its supporting subset is too small.

<!-- generated:se2_optin_country -->
| Organizer country | Opt-in checkbox on or set to auto | Campaigns | Businesses |
|---|---|---|---|
| Poland | 71% | 824 | 38 |
| United States | 34% | 22,522 | 3,253 |
| Australia | 17% | 2,919 | 721 |
| Malaysia | - | 168 | 14 |
<!-- /generated -->

Source: `analysis/output/field_cuts.json` key `config_cuts`, `analysis/output/indicators.json` key `email_optin_checkbox_by_country`.

## Terms by country

Custom terms text share by country: campaigns whose organizer wrote their own terms wording. English-speaking markets write their own far more often than India or Brazil.

<!-- generated:se2_terms_country -->
| Organizer country | Wrote custom terms | Campaigns | Businesses |
|---|---|---|---|
| Australia | 54% | 6,066 | 1,334 |
| United Kingdom | 45% | 15,188 | 1,710 |
| United States | 38% | 57,289 | 8,720 |
| India | 6% | 4,062 | 691 |
| Brazil | 4% | 5,299 | 497 |
<!-- /generated -->

Wording families inside that text turn up at different rates by country. Void where prohibited and no purchase necessary are classic US sweepstakes phrases. Platform disclaimer wording is an Australian habit, and permit-or-licence wording turns up in only two countries, Australia and Malaysia, at similar rates. Age wording differs by phrasing more than by presence: Canada is the one country in the table where "age of majority" outruns a stated "18+", the reverse of the United States, United Kingdom and Australia, where "18+" is the more common phrase.

<!-- generated:se2_terms_wording -->
| Wording family | Country | Share |
|---|---|---|
| Void where prohibited | United States | 28% |
| No purchase necessary | United States | 34% |
| Platform disclaimer | Australia | 39% |
| Permit or licence | Australia | 5.9% |
| Permit or licence | Malaysia | 6.7% |
| Skill or judged | Australia | 8.6% |
| "Age of majority" phrasing | Canada | 26% |
| "18+" phrasing | Canada | 22% |
<!-- /generated -->

Governing law named in the generated terms points to a country other than the organizer's own far more in some markets than others. Organizer country here comes from IP, so a genuinely foreign-registered entity reads as a mismatch too.

<!-- generated:se2_law_mismatch -->
| Organizer country | Governing law names a different country | Campaigns | Businesses |
|---|---|---|---|
| Japan | 76% | 4,784 | 235 |
| United States | 8% | 56,583 | 8,613 |
<!-- /generated -->

The generated terms set the draw window and the Winner response window at 7 days on the typical campaign in every country in the table.

Source: `analysis/output/country_cuts.json` key `terms_by_country`, `analysis/output/indicators.json` keys `governing_law_mismatch_by_country` and `age_wording_in_custom_terms_by_country`.

## Australian games of skill

Australian campaigns that carry a skill or judged signal in their name, Prize copy or terms run fewer entry methods, ask for fewer actions, get less to enter, run longer, and carry a bigger stated Prize pool than Australia's other campaigns.

<!-- generated:se2_au_skill -->
|  | Skill/judged signal | No skill/judged signal |
|---|---|---|
| Campaigns | 629 | 5,437 |
| Businesses | 124 | 1,277 |
| Entry methods (typical) | 2 | 6 |
| Entries per Entrant | 1.8 | 3.4 |
| Entry rate | 18% | 30% |
| Duration (typical) | 18 days | 14 days |
| Prize pool (typical) | $2,135 | $699 |
<!-- /generated -->

By industry, media and entertainment publishers account for about half of the skill or judged campaigns, ahead of electronics and tech, gaming, and nonprofit and community.

<!-- generated:se2_au_skill_industries -->
| Industry | Campaigns (of the 629 skill/judged) |
|---|---|
| Media and entertainment | 320 |
| Electronics and tech | 51 |
| Gaming and esports | 48 |
| Nonprofit and community | 33 |
<!-- /generated -->

Source: `analysis/output/country_cuts.json` keys `skill_against_other` and `australia_skill_industries`.

## Settings the campaign analysis added

The September 2026 export carries entry worth, the mandatory flag, actions required, the paid flag, loyalty tiers, integrations, language, allowed and excluded countries, the generated terms settings and the organizer's country. Aggregates live in `analysis/output/field_cuts.json`. Headlines across the ordinary population:

- Entry worth: `entry_count` is the number of completions of the action, not a worth-weighted total, so no division by worth is needed. Making an action mandatory raises how many Entrants complete it, and typically lowers the cost per completion too, scoped to campaigns with every Prize valued (about a third of the dataset). Email was mandatory on 33% of the campaigns offering it.

<!-- generated:se2_email_mandatory -->
  | Email action | Completion, % of Entrants (decimal) | Actions |
  |---|---|---|
  | Typical campaign (any) | 85% (0.851) | 46,858 |
  | Mandatory | 102% (1.024) | - |
  | Optional | 78% (0.777) | - |
<!-- /generated -->

<!-- generated:se2_mandatory_cost -->
  | Action | Campaign size (Entrants) | Cost optional | Cost mandatory | Save | Campaigns (optional / mandatory) | Businesses (optional / mandatory) |
  |---|---|---|---|---|---|---|
  | Email | 1,000 to 2,500 | $0.44 | $0.39 | 12% | 5,880 / 3,657 | 1,446 / 1,111 |
  | Email | 2,500 to 10,000 | $0.51 | $0.25 | 51% | 3,722 / 2,852 | 833 / 803 |
  | X Follow | 1,000 to 2,500 | $0.64 | $0.64 | close, no save | 14,411 / 1,711 | 2,376 / 571 |
  | X Follow | 2,500 to 10,000 | $0.65 | $0.55 | 15% | 10,232 / 1,148 | 1,392 / 305 |
  | X Follow | 10,000 or more | $0.44 | $0.26 | 42% | 2,851 / 411 | 361 / 83 |
  | Discord_join | 1,000 to 2,500 | $0.76 | $0.94 | optional cheaper | 3,710 / 298 | 818 / 162 |
  | Discord_join | 2,500 to 10,000 | $0.77 | $0.72 | 6% | 2,254 / 178 | 500 / 89 |
<!-- /generated -->

  Follow_x follows the email pattern from 2,500 Entrants up, not below it. Discord_join is not a rule either way, its mandatory cost is cheaper at one size and pricier at the other. (Figures from `asset_yield.json`'s `yield_by_asset_and_mandatory` cut.)
- Email provider: the selected providers in the table recorded email completion rates near or above the no-integration group. Other providers recorded lower rates, so this selection does not describe every provider. These comparisons describe campaign setups and do not establish an effect of choosing a provider.

<!-- generated:se_providers -->
  | Email provider | Completion, % of Entrants (decimal) | Actions | Businesses |
  |---|---|---|---|
  | No integration | 83% (0.832) | 30,857 | 3,274 |
  | MailChimp | 85% (0.852) | 5,890 | 1,141 |
  | Klaviyo | 96% (0.956) | 3,995 | 1,114 |
  | MailerLite | 100% (1.004) | 822 | 112 |
  | ConstantContact | 86% (0.863) | 461 | 73 |
<!-- /generated -->
- Country rules: 25% of campaigns restrict entry by country. Restricted campaigns offer email far more often, and restricting costs little on the share who enter.

<!-- generated:se2_country_rule -->
  |  | Restricted by country | Open |
  |---|---|---|
  | Campaigns | 29,521 | 86,978 |
  | Businesses | 4,919 | 14,473 |
  | Offers email | 54% | 26% |
  | Entry rate | 27% | 27% |
<!-- /generated -->
- Paid actions: campaigns that carry a paid action run more methods and more Entries per Entrant than campaigns without one, with a lower share of viewers entering.

<!-- generated:se2_paid -->
  |  | Has a paid action | No paid action |
  |---|---|---|
  | Campaigns | 3,694 | 112,805 |
  | Businesses | 1,157 | 17,116 |
  | Methods (typical) | 10 | 7 |
  | Entries per Entrant | 5.35 | 4.36 |
  | Entry rate | 23% | 27% |
<!-- /generated -->
- Language: English dominates the dataset and runs the longest campaigns. Portuguese campaigns are one-day runs that almost never ask for email. German, Spanish and French sit between the two on duration, and French asks for email about as often as English does.

<!-- generated:se2_language -->
  | Language | Share of campaigns | Campaigns | Businesses | Duration (typical) | Email offered |
  |---|---|---|---|---|---|
  | English | 93% | 108,473 | 17,002 | 15 days | 34% |
  | Portuguese | 2.0% | 2,331 | 116 | 1 day | 2% |
  | German | 0.8% | 889 | 154 | 9 days | 20% |
  | Spanish | 0.8% | 924 | 252 | 13 days | 25% |
  | French | 0.8% | 951 | 216 | 10 days | 35% |
<!-- /generated -->
- Terms: the generated terms set the draw and the Winner response window at 7 days on most campaigns. 39% of campaigns used fully custom terms. Governing law runs US, GB and CA as the top three, then AU and BR.

<!-- generated:se2_law_top -->
  | Governing law | Campaigns |
  |---|---|
  | US | 55,088 |
  | GB | 14,002 |
  | CA | 6,979 |
<!-- /generated -->
- Build lead: 22% of campaigns show a creation time after the start date, so the field reads as a modified-at time on those rows, not a true build lead. (25,338 campaigns, 6,644 businesses.) Among the rest, lead time barely moves the share who enter.

<!-- generated:se2_build_lead -->
  | Build lead | Entry rate |
  |---|---|
  | Same day | 27% |
  | 1 to 2 days | 26% |
  | 3 to 7 days | 27% |
  | 8 days or more | 26% |
<!-- /generated -->
- Viral Share clicks: campaigns with a Viral Share action see share clicks equal to 20% of Impressions, and the typical campaign sees about 93 share clicks per 100 Entrants (0.927 per Entrant).
- Landing: Impressions split about evenly between the hosted gleam.io page and an embed, with the directory listing a rounding error. Traffic source is dominated by direct, ahead of organizer sites, Meta and YouTube. Of the Impressions carrying a UTM tag, about a third trace to email newsletters and a fifth to Meta.

<!-- generated:se2_landing -->
  | Landing page type | Share of Impressions |
  |---|---|
  | Hosted gleam.io page | 49% |
  | Embed | 51% |
  | Directory listing | 0.4% |
<!-- /generated -->

<!-- generated:se2_traffic -->
  | Traffic source | Share of Impressions |
  |---|---|
  | Direct | 48% |
  | Organizer sites | 9.6% |
  | Meta | 9.0% |
  | YouTube | 6.1% |
  | Giveaway directories | 4.8% |
  | Search | 3.9% |
  | X | 3.7% |
  | Gaming communities | 3.1% |
  | Gleam pages | 3.1% |
<!-- /generated -->

  (15% of Impressions carry a UTM tag. Of those, 32% trace to email newsletters and 23% to Meta.)
- Hosted share: campaigns landing almost entirely on the hosted page get the most to enter, over the shortest run, with the least email offered. Campaigns that lean on an embed run longer, offer email far more, and get less to enter. This is the largest split in the dataset, with no comparison group to show what caused it.

<!-- generated:se2_hosted -->
  | Landing mix | Campaigns | Businesses | Entry rate | Duration | Email offered |
  |---|---|---|---|---|---|
  | 90% or more hosted | 70,080 | 13,712 | 31% | 12 days | 24% |
  | 50 to 90% hosted | 7,912 | 2,161 | 24% | 16 days | 47% |
  | 10 to 50% hosted | 17,319 | 2,923 | 20% | 21 days | 52% |
  | Under 10% hosted, embed mostly | 20,872 | 3,212 | 23% | 16 days | 44% |
<!-- /generated -->

Still not in the dataset: fraud level and CAPTCHA mode, require login and the login types, email and phone verification, age restriction, daily and interval settings, feature images and video, post-entry emails, redirects, pixels and webhooks. The docs describe what they do.

## Actions that top campaigns used, by vertical

The entry-method-planner skill holds the table of actions over-represented in the top fifth of campaigns by vertical, in Gleam action names (Twitch Subscribers and Secret Code in gaming, single choice questions in technology, YouTube Channel Visits and Pinterest Visits in home, and so on). Load `mix-by-objective.md` from that skill when the user asks which Gleam actions fit their industry. As a default mix in Gleam terms: one acquire action (Subscribe to an Email List, or Custom for account creation), one follow on the network the audience uses, one Single Choice Question or Answer a Question that collects something useful, one Visit a Page or secret code for exposure, and Viral Share for reach.

## Wording that showed up in the data

The entry-method-planner skill holds recorded completion events per Entrant by question type, by share copy traits, by visit destination and by the newsletter description under the email action (see table). The idea generator holds title wording. Load them when the user is writing the action titles and the description.

<!-- generated:se3_wording -->
| Measure | Category | Actions | Businesses | Completed it, per 100 Entrants (typical) |
|---|---|---|---|---|
| Question type | other | 6,937 | 1,509 | 75 |
| Question type | detail capture | 6,755 | 851 | 68 |
| Question type | preference | 3,475 | 898 | 70 |
| Question type | trivia | 1,640 | 325 | 62 |
| Question type | feedback or open | 839 | 374 | 86 |
| Visit destination | another site | 121,345 | 6,668 | 77 |
| Visit destination | YouTube | 49,326 | 7,552 | 85 |
| Visit destination | the business's own site | 10,195 | 2,564 | 97 |
| Visit destination | another Gleam campaign | 583 | 271 | 85 |
| Visit destination | Steam | 303 | 30 | 79 |
| Visit destination | Instagram | 194 | 39 | 95 |
| Visit destination | Facebook | 109 | 33 | 57 |
| Visit destination | Amazon | 100 | 29 | 69 |
| Newsletter description (email action) | no description | 20,856 | 3,554 | 95 |
| Newsletter description (email action) | description without those | 18,701 | 2,478 | 81 |
| Newsletter description (email action) | mentions frequency or unsubscribe | 7,301 | 863 | 81 |
<!-- /generated -->

Source: `analysis/output/text_and_context.json` keys `question_types`, `visit_destinations` and `newsletter_wording`, the same keys the planner's tables read.

## Settings the other skills already measured

- Email opt-in checkbox on: 89 completion events per 100 Entrants against 100 with it off (giveaway-entry-method-planner). The difference does not count people who skipped consent or addresses eligible for mailing. That gap is wider than the 92% against 85% in the table above because it uses the subset with no repeatable action and runs of 14 days or less (2,726 checkbox-on and 7,734 checkbox-off campaigns, `analysis/output/extra_cuts.json`, `optin_email_uptake_clean`), while the table above pools all run lengths. Quote whichever base the answer is about and say which one it is.
- Validated-answer question: a typical 3% of entries come back invalid in the fair-comparison subset (509 campaigns, 77 businesses, `analysis/output/extra_cuts.json`, `invalid_validated_question_clean`). This subset excludes long runs and repeatable actions.
- Custom terms: written by 39% of campaigns (`analysis/output/extra_cuts.json`, `custom_terms_share_all`).
- Eleven or more actions: 17% more Entrants, with 31% of viewers entering against 44% for one to three, among campaigns with no repeatable action and runs of 14 days or less (6,762 and 10,802 campaigns respectively, `analysis/output/comparisons.json`, `method_count_clean`, also in giveaway-entry-method-planner).

Source for the invalid-entry share: `analysis/output/invalid_share.json` (`distribution`, `by_share_action`).

- Invalid entries: a typical 4.2% of entries, higher with Viral Share, Discord and retweet actions (giveaway-winner-structure).

## What this suggests for the setup, as advice

- Put the action that produces the asset first and make it the single mandatory action. Completion-event ratios were lower in later positions, but this comparison also reflects which actions businesses chose to put first (see the position table above). If giveaway-entry-method-planner is installed, its mix-by-objective reference has the fuller asset list and the ratios by campaign size. Otherwise use the position table above and leave the additional comparisons unavailable.
- Choose mandatory status according to the campaign objective. Mandatory email actions recorded lower stated Prize value per completion in the entry-worth table above, with wider differences in larger size bands. Follow_x showed the same direction from 2,500 Entrants up. Discord_join showed opposite directions across the two sizes tested. These are observed comparisons, with no measured saving from changing mandatory status.
- Keep unnecessary effort out of the list. Later-position completion varies by family, so use its row in the position table to decide whether an action earns its place. Eleven or more actions came with a lower Conversion Rate but more Entrants than one to three actions in that comparison.
- Set Viral Share expectations at one referred entry per eight Entrants, and weight it accordingly. It falls away down the list like every other family: 16 completions per 100 Entrants in first position, 12 in the middle, 11 fifth or later. As practice, offer optional Viral Share when referrals support the campaign objective, including with a small Prize. Omit the Action when referrals do not serve that objective. Decide the referral cap separately: set a cap before launch when the terms should limit how many referrals each Entrant can earn Entries for, including with an expensive Prize. Leave referrals uncapped only when unlimited referral credit is intended. Prize size alone does not determine either choice. Review suspicious referrals before the draw under the published terms.
- Use a question or choice template when you want an answer. Free-form custom actions recorded about 50 completion events per 100 Entrants in the template table above.
- Use on-click completion for a simple visit objective. Add a delay or question when the objective requires time on the page or an answer. The visit-option tables above record lower completion-event ratios for questions than clicks, and for embedded HTML than plain links. They do not establish how many completions this campaign would lose by changing its configuration.
- Keep descriptions concise and lead with eligibility and Winner count. As a drafting preference, start around 150 words when useful and retain essential Prize and fulfilment details. The description-length comparison above does not establish a cutoff.
- Leave the fraud level on High, the default, and review Invalid entries on the Actions tab before drawing. Expect a few percent to fall away.
- Turn on the opt-in checkbox when collecting mailing consent. Build the mailable list from individual consent records, with duplicate addresses and subsequent unsubscribes accounted for. An email-action completion ratio alone cannot establish how many subscribers you can email.
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall