Gleam Giveaway Benchmarks 2026, Edition 2: Viral Share

We looked at 4,219 competitions run by 1,028 businesses that offered the Viral Share entry Action, out of 10,249 competitions on paid Gleam plans started between 1 September 2025 and 13 August 2026. We wanted to show giveaway organizers what a refer-a-friend Action typically brings in, who switches it on, and which settings travel with a higher referral share. Viral Share is available on Pro, Business and Premium plans, so every comparison on this page is between paid-plan campaigns. Every figure below is a median, and every table names its sample and its window.

1. What we found
2. Who offers the share Action
3. The referral funnel, from Entrant to click to referred Entrant
4. What the extra entries are worth
5. Where the Action sits in the widget
6. Campaigns with the Action against campaigns without
7. Referral share by industry
8. Duration, and when the clicks arrive
9. Promotion and sharing move together
10. Actions that travel with Viral Share
11. Inside the top decile
12. Invalid entries and referrals
13. What to change on your next campaign
14. Benchmark tables
15. Frequently asked questions
16. How we did this
17. Cite this research

Viral Share is a paid-plan feature, so the first thing to measure is how many of the businesses that can switch it on actually do, and whether that changes by plan or by industry.

Among the 10,249 paid-plan competitions in the window, 4,219 offered Viral Share, 41.2 percent, run by 1,028 of the 2,226 businesses. Adoption falls slightly as the plan gets larger.

Table 1 · Viral Share adoption by plan, 1 Sep 2025 to 13 Aug 2026

Premium sits below the 50-business minimum and is shown for completeness. Free and Hobby plans do not include the Action and are excluded from the whole page. 3,848 campaigns in the window ran on those plans. The three rows are counted before the exclusion of four campaigns with faulty click records described in How We Did This, so they sum to 10,253 against the 10,249 used elsewhere.

Adoption varies more by industry than by plan.

Table 2 · Viral Share adoption by industry, paid plans, 1 Sep 2025 to 13 Aug 2026

Industry rows use the 300-campaign, 75-business floor. Where one business held more than 10 percent of an industry's campaigns, that business was removed and the row recounted, which is why the campaign counts are lower than the raw industry totals. Twelve industries fall below the floor and are not shown, Brick and Mortar among them at 261 campaigns. Hobbies clears the business floor at 93 and is shown at 361 campaigns, above the 300 floor.

The spread runs from one campaign in four to two in three. Hobbies, crafts and sports brands switch the Action on far more often than B2B or general-purpose organizers. One plausible reason, which this data cannot test, is that their audiences already gather in communities. Section 7 asks whether those brands then see a higher referral share.

Every referral passes through three gates. An Entrant has to share the link, somebody has to click it, and the person who clicked has to enter. The table below is the whole funnel for the typical campaign, reported one median at a time.

Table 3 · The referral funnel on the typical Viral Share campaign, paid plans, 1 Sep 2025 to 13 Aug 2026

n = 4,208 campaigns across 1,026 businesses with at least one share-link click day. 11 campaigns carried no click record and are left out of this table. Each row is its own median, so the rows do not multiply into one another. The 95 percent confidence interval on the 7.5 percent median, from a business-clustered bootstrap, runs from 5.6 to 9.4 percent.

For every 100 Entrants, the typical share campaign had 64 share-link clicks and 8 referred Entrants. Each row is its own median across campaigns, so the rows do not multiply into one another. n = 4,208 campaigns across 1,026 businesses on paid plans, 1 September 2025 to 13 August 2026.

A referred Entrant is also an Entrant, so the share is bounded, and the campaign-level median of 7.5 percent means one Entrant in thirteen arrived through somebody else's link. The distribution is wide. One campaign in ten drew under 1.5 percent of its Entrants through referrals, one in four drew over 15 percent, and one in ten drew over 26 percent.

Only 2.5 percent of campaigns that offered the Action had no referral at all, so if you switch it on, you will almost certainly see some.

The typical campaign had 0.64 share-link clicks per Entrant. This data cannot say how many Entrants shared at all, because a click is recorded on the receiving end. The figure describes what the links drew. Of the clicks that happened, fewer than one in seven ended in an entry. On the typical campaign a click on a share link converted to an Entrant at 12.8 percent, against 25.2 percent for every other widget view. One likely reason is that a referred visitor arrives colder than someone who found the giveaway on their own.

Share-link clicks also made up part of total reach. On the typical share campaign, 14.3 percent of all widget views came through a share link, and on a quarter of campaigns more than 25.7 percent did. A share campaign's reach therefore already contains the share Action's own traffic, which is why section 6 subtracts it before comparing.

Table 4 · Referral funnel by plan, 1 Sep 2025 to 13 Aug 2026. Every figure is a median.

The Business row is shown with a warning: one business holds 18 percent of its campaigns, above our 10 percent dominance rule, so its figures are indicative and do not appear in the benchmark tables. Premium is suppressed at 18 businesses. The Pro row clears every threshold and is the row to plan against.

A referral is counted when the referred person enters, so a click that led to a follow, a purchase or a later entry on another campaign is invisible here. The data holds no entrant-level record either, so it cannot say whether referred Entrants complete more Actions, win more often or stay on a list longer than direct Entrants. Referred Entrants are Entrants, and we do not claim they are qualified leads.

Every organizer who switches on the Action has to decide how many entries a referral is worth, so the natural test is whether campaigns that award more see a higher referral share.

The typical campaign awards 5 entries per referral. Just over a quarter award 1, four in ten award 10 or more, and one in ten awards 50 or more. A small group, 4.3 percent, make the share Action mandatory.

Table 5 · Referral share by entries awarded per referral, paid plans, 1 Sep 2025 to 13 Aug 2026. Every figure is a median.

Differences are against the worth-1 row. In four of the six rows one business held more than 10 percent of the campaigns (27.9 percent of the worth-1 row and 33 percent of the 25-or-more row). Those businesses were removed and the rows recounted, following the dominance rule in How We Did This. Every row shown clears 200 campaigns and 50 businesses after trimming.

Referral share varied by three points at most across the award bands. Median referral share of Entrants by entries awarded per referral, after removing any business above 10 percent of a band. n = 3,381 campaigns, 1 September 2025 to 13 August 2026.

Before trimming, the worth-1 row read 5.3 percent, and the gap to the next band read 3.3 points. Most of that gap belonged to one business running 330 single-entry campaigns with a low referral share. Once it is set aside, the whole table runs from 7.8 to 11.0 percent, and the two lowest awards sit within a point of each other.

Collapsed to a single split, campaigns awarding 2 or more entries had an 8.4 percent referral share against 8.0 percent at a single entry, a difference of 0.4 points. The same split on one campaign per business gave 9.5 against 7.6 percent, and inside Pro alone 8.2 against 6.8 percent. The direction was the same each time, and the size was within two points each time.

The rows also show that campaigns awarding 10 or more entries were larger campaigns, at 1,028 and 1,389 Entrants against 781 at a single entry. The organizers of bigger campaigns chose bigger awards. Whether the award grew the campaign or the organizer of a big campaign chose a big award, this data cannot separate.

The award was not assigned at random. Organizers who set 25 entries per referral also ran longer widgets, larger campaigns and more of them, so the 3-point gap between the top and bottom rows carries all of that with it. Nothing here shows that raising the award raises the referral share.

Organizers set the order of their entry Actions, which raises the question of whether a share Action near the top comes with more referrals. Position and widget length travel together, since a share Action in ninth place can only exist on a widget with nine Actions. The raw position table comes first, and the comparison with widget length held in a band comes after it.

Table 6 · Referral share by share Action position, paid plans, 1 Sep 2025 to 13 Aug 2026

The 1st-position row is suppressed at 97 campaigns and 49 businesses. Only 2.3 percent of campaigns put the share Action first. Differences are against the 2nd-to-4th row.

Holding the widget length in a band gives the fair comparison.

Table 7 · Top-four placement against later placement, inside two widget-length bands

Every cell clears 200 campaigns and 50 businesses.

On a short widget, position made under a point of difference. On a long one, a share Action in the top four slots came with 10.4 percent of Entrants referred against 6.4 percent when it sat fifth or later, and the clicks per Entrant moved in the same direction, 0.77 against 0.63. On a long widget the share Action competes with eight or more other ways to earn entries, and one reading is that Entrants reach the top slots first.

The number of Actions on its own showed a smaller gap than position did. Campaigns with 1 to 3 Actions had a 9.3 percent referral share, 9 to 12 Actions 6.6 percent, and 13 or more 7.7 percent, so the relationship is not a straight line, and the 1-to-3 band carries 146 campaigns from 79 businesses, under the reporting floor.

For a widget of nine or more Actions, the placement to try is Viral Share in the first four, which came with four more Entrants in every hundred arriving through referrals. On a widget of eight or fewer Actions the order made under a point of difference, and there is no measured reason to move it.

The headline comparison flatters the share Action for a reason that has nothing to do with referrals.

Table 8 · Median campaign profile with and without Viral Share, paid plans, 1 Sep 2025 to 13 Aug 2026

Share campaigns drew 38 percent more Entrants and 73 percent more Impressions. They were also run by different businesses, with longer widgets and longer runs, so the table compares organizers as much as it compares the Action.

The cleanest comparison this data allows is the same business running both kinds of campaign in the same window. 159 paid-plan businesses did, 980 campaigns with the Action and 970 without. For each business we took the median of its share campaigns and the median of its non-share campaigns, then the median across businesses, so every business counts once and no large account can tilt the result.

Table 9 · Within the same business, campaigns with and without Viral Share, 1 Sep 2025 to 13 Aug 2026

n = 159 businesses, 1,950 campaigns. Each cell is the median across businesses of that business's own median. The ratio column is the median of per-business ratios, so it does not equal the ratio of the two medians. The middle half of businesses ran from 0.81x to 3.11x. 63.5 percent of businesses ended higher with the Action.

The same business drew 41 percent more Entrants on its campaigns with Viral Share. Bars are the median across businesses of each business's own median. The 1.41x is the median of each business's own ratio, so it does not equal 1,094 divided by 621. n = 159 businesses, 1,950 campaigns, 1 September 2025 to 13 August 2026.

The same business, on its campaigns with the Action, drew 41 percent more Entrants at the median, on views that were only 10 percent higher before any share-link click is counted. Total views were 49 percent higher, and the gap between those two figures is the traffic the share links brought. The referral share inside these 159 businesses was 9.6 percent, so the links account for part of the 41 percent and the rest travels with the campaign itself. The share campaigns also carried eight Actions against five, which suggests a business tends to switch the Action on for its bigger effort.

Narrowing to businesses whose two kinds of campaign started from similar reach, within 25 percent of each other, leaves 25 businesses and a ratio of 1.21x. Narrowing instead to similar widget lengths, within two Actions, leaves 66 businesses and 1.16x. Both are below the 50-business floor for a public cell and are reported here as direction only.

The second test holds still the reach a campaign would have had anyway. Base reach is Impressions minus share-link clicks, and both groups are compared inside bands of base reach and widget length.

Table 10 · Median Entrants with and without Viral Share, at the same base reach and widget length

Base reach is Impressions minus share-link clicks, so a campaign without the Action has base reach equal to its Impressions. Differences under 5 percent are shown as none. The 10,000-and-over, 1-to-5 cell is suppressed at 73 share campaigns from 47 businesses. Every other cell clears 200 campaigns and 50 businesses on the without side, and the with side clears 130 campaigns and 77 businesses at its smallest. Cells on the with side under 200 campaigns are Under 1,000 with 1 to 5 and 6 to 8 Actions, 1,000 to 2,999 with 1 to 5, 3,000 to 9,999 with 1 to 5, and 10,000 and over with 6 to 8, and those five are shown with that warning.

In eight of the eleven reportable cells the campaign with the Action ended with more Entrants, by 7 to 26 percent. In two the groups were within 5 percent of each other. In one, the largest campaigns with the longest widgets, the share campaigns ended 15 percent lower. A referral share of 7.5 percent at the median would, on its own, account for a difference of about 8 percent, which is where most of the cells sit. This test is deliberately conservative. It subtracts the share clicks before comparing, so the gap that remains is about the size the referred Entrants alone would add, and the Entrants who arrived some other way look much the same with or without the Action.

The same Action could bring in a different share of Entrants depending on what the organizer sells. Three industries clear the 300-campaign, 75-business floor for a share-campaign benchmark after the dominance trim.

Table 11 · Referral funnel by industry, share campaigns on paid plans, 1 Sep 2025 to 13 Aug 2026. Every figure is a median.

Differences are against the Gadgets row. In each of the three rows one business held more than 10 percent of the campaigns (19.3, 27.3 and 33.4 percent) and was removed before recounting. Sports (216 campaigns, 66 businesses, 18.1 percent) and Hobbies, Crafts and DIY (233 campaigns, 65 businesses, 6.9 percent) sit below the floor and are named here so the next edition can carry them. They are not benchmarks.

Gadgets brands drew 16.5 percent of Entrants through referrals, creators 6.1 percent. Median referral share among the three industries clearing 300 campaigns and 75 businesses after the dominance trim. n = 1,730 campaigns across 448 businesses, 1 September 2025 to 13 August 2026.

The gap between Gadgets and Creators is 10.4 points, and it opens at the first gate. A Gadgets Entrant's link drew 1.08 clicks, a Creator Entrant's 0.54. The conversion of clicks into Entrants differed by far less, 0.17 against 0.13. Gadgets Entrants' links drew more clicks, and the people who clicked entered at a similar rate.

Read alongside section 2, the industries that offered the Action most often and the industries with the highest referral share were different ones. Hobbies brands offer it on 64.5 percent of campaigns and, on the sub-floor figure, draw 6.9 percent of Entrants through it. Gadgets brands offer it on 32.8 percent and draw 16.5 percent.

Reference only, not a benchmark: 1,368 campaigns in the window were classified as cryptocurrency or Web3 (a wallet-address Action or a cryptocurrency industry label) and are excluded from every table on this page. Among them, 29.2 percent offered Viral Share and the median referral share was 35.3 percent, more than four times the mainstream figure. They are excluded because their entry mechanics, audience and fraud profile differ enough to distort a general benchmark.

Duration is the other setting every organizer picks, and the daily click record shows when inside the run people share. The typical share campaign ran 20 days, with the middle half between 10 and 31, against 13 days (7 to 29) for paid campaigns without the Action. Longer runs did not come with a higher referral share.

Table 12 · Referral share by campaign duration, share campaigns on paid plans, 1 Sep 2025 to 13 Aug 2026

Differences are against the 1-to-7-day row. Every row clears 200 campaigns and 50 businesses.

The one-week campaign had the highest referral share at 9.5 percent and the month-plus campaign the lowest at 6.0 percent, a 3.5-point range. The daily click record offers part of the explanation. Sharing is front-loaded, and a longer campaign adds days in which few people share.

Table 13 · When share-link clicks arrive, by campaign duration

Each figure is the median, across campaigns, of the share of that campaign's own clicks that fell in the period. Campaigns with no click record are excluded, which is why the counts are lower than in Table 12.

Days one and two carried 21 percent of a two-week campaign's share-link clicks. Each bar is the median, across 14-day share campaigns, of that campaign's own share of clicks on that day, so the bars do not sum to 100. Up to 528 campaigns per day, 1 September 2025 to 13 August 2026.

On a 14-day campaign, day one carried 11.0 percent of all clicks and day two 10.1 percent. By day four the daily share had settled to between 5 and 6 percent and stayed there, with a small rise to 5.5 percent on the final two days. A two-week campaign collected 63 percent of its clicks in the first week, so the second week added roughly half as many again.

Clicks cluster early in the run. If the referral share is the reason you are running the Action, the pattern points to front-loading the promotion and keeping the run short, or planning a second push mid-campaign, since the closing days carried 6 to 14 percent of clicks on runs longer than a week.

Every widget view carries the domain it came from, when the browser sends one. Half do not. Across share campaigns in the window, 47.5 percent of views arrived with no referrer, which covers apps, messaging, email clicks that lose their source and direct visits. Social sites sent 19.8 percent, the organizer's own site and other sites 16.4 percent, giveaway directories 8.5 percent, Gleam-hosted pages 4.6 percent, search 1.9 percent and email clients 1.3 percent. The typical share campaign drew traffic from six referring domains with at least 1 percent of its views.

Promotion breadth is the setting that travels most closely with the referral share.

Table 14 · Referral share by number of referring domains, share campaigns on paid plans, 1 Sep 2025 to 13 Aug 2026

Differences are against the 1-to-2 row. Referrer domains are counted from Gleam's daily traffic record and classified by domain pattern. The 5-to-7 row clears the dominance rule at 9.5 percent. The other three rows carry one business above 10 percent (35, 18 and 17 percent) and have not yet been recounted without it, so they are indicative.

Campaigns reached from eight or more sources had 10.5 percent of Entrants referred, against 3.4 percent from one or two. Median referral share by number of referring domains sending 1 percent or more of views, with clicks per Entrant beneath each bar. Three of the four bands are indicative until the dominance recount. n = 4,208 campaigns across 1,026 businesses on paid plans, 1 September 2025 to 13 August 2026.

Campaigns reached from eight or more sources drew 10.5 percent of Entrants through referrals, campaigns reached from one or two sources 3.4 percent. The clicks per Entrant moved with it, 0.20 to 0.77, and the entry rate after the click did not. One reading is that Entrants on widely promoted campaigns had more places to share the link, though the data cannot separate that from the bigger budgets and lists those campaigns also carried.

Two channels differed from the rest. Campaigns where giveaway directories sent 10 to 50 percent of views had a referral share of 11.0 percent (1,478 campaigns, 478 businesses) against 4.7 percent where directories sent under 1 percent (1,343 campaigns, 459 businesses). A plausible reason is that directory visitors come for the entries, and the share Action pays in entries. Campaigns where social sites sent 60 percent or more of views ran the other way, at 4.9 percent (175 campaigns, 124 businesses, below the floor), against 7.1 to 8.6 percent elsewhere.

The daily record shows the same thing from inside a single campaign. Across 3,741 share campaigns with at least seven days of data, the day-by-day count of share-link clicks moved with the day-by-day count of other views at a median rank correlation of 0.46, with the middle half of campaigns between 0.25 and 0.67. In 62 percent of campaigns the peak day for share clicks fell on the same day as the peak day for other traffic, or one day either side. The median lag was zero days.

Two limits sit on these figures. A referred visitor arrives carrying the referrer of wherever the link was shared, so the social and no-referrer shares include traffic that came through the share Action itself, and part of the correlation between daily clicks and daily views is that relationship read backwards. Referrer classification is by domain pattern and unaudited. Campaigns promoted through many sources may also have bigger budgets, larger lists and more experienced organizers, and none of those are measured here.

For an organizer, the pattern makes the share Action an amplifier of promotion already running. Each push you make, an email send, a creator post, a directory listing, is also a day your Entrants share, so plan the pushes around the days you want referrals on.

Organizers rarely run Viral Share alone, and the widget they build around it differs from one without it.

Table 15 · Share of campaigns carrying each Action, with and without Viral Share, paid plans, 1 Sep 2025 to 13 Aug 2026

n = 4,219 campaigns with the Action and 6,019 without. A campaign with eleven Actions carries more of everything, so most rows rise. The rows to read are the ones that rise by more than the widget length alone would predict.

Email Subscribe is the widest pairing gap. It sits on more than half of share campaigns and under a third of the rest, a 27-point gap that is twice the next largest. Organizers who want referrals usually also collect the address, and the two Actions together make a list-building widget where an Entrant enters by email and then brings a friend.

The pairings say nothing about which combination performs. Among share campaigns, the ones in the top tenth by referral share carried Email Subscribe 42 percent of the time, the bottom four tenths 62 percent, so the most common pairing sat more often on the campaigns with the lowest referral share (section 11).

Ranking share campaigns by referral share shows what a campaign looks like when the Action works, and where it differs from one where it barely fires.

Table 16 · Share campaigns ranked by referral share, paid plans, 1 Sep 2025 to 13 Aug 2026. Every figure is a median.

Award is entries per referral. Entered is Entrants per widget view, and the top tenth's lower figure is expected, because share-link clicks are views that convert at 12.8 percent.

The top tenth's links drew 2.15 clicks per Entrant, seven times the bottom group's 0.29. Median share-link clicks per Entrant and referred Entrants per click, share campaigns ranked by referral share. n = 4,208 campaigns across 1,026 businesses on paid plans, 1 September 2025 to 13 August 2026.

The top tenth separated from the rest at both gates, and mostly at the first. Its Entrants' links drew 2.15 clicks each, seven times the bottom group's 0.29, and its clicks converted at 0.19 against 0.09. The settings organizers control differed less. The top tenth awarded 10 entries per referral against 5, ran a slightly shorter campaign, and placed the Action in the seventh slot, like the next and middle groups.

Two things did differ. The top tenth ran campaigns more than twice the size of the bottom group, 1,048 Entrants against 440, and its industry mix was different. Sports brands placed 36.6 percent of their share campaigns in the top tenth and Smart Wear and Gadgets 23.0 percent, against 4.5 percent for E-commerce and 5.4 percent for Content Creators.

Ranking campaigns by the outcome and then describing them will always find differences, and this table names those differences without explaining them. Audience, prize appeal and whether the organizer promoted the share link outside the widget are unmeasured here, and the sports and gadgets concentration suggests the audience accounts for much of the gap. On this evidence, setting the award to 10 would not move a campaign into the top tenth.

A refer-a-friend Action could come with more entries that Gleam's filters reject, so we compared the invalid share with and without it.

Table 17 · Invalid entries as a share of all entries, paid plans, 1 Sep 2025 to 13 Aug 2026

Campaigns with no invalid-entry record are excluded, which is why the counts fall short of the section totals. Every row clears the 500-campaign, 100-business floor for sensitive metrics.

The invalid share doubled across the three referral bands, from 2.1 to 4.2 percent. Every figure here is after Gleam's filtering, so the Entrant and referral counts on this page exclude those entries already. The pattern is consistent with a share link drawing a small number of people who try the campaign more than once, and with Gleam's filters catching them. This report analyzes performance, and it does not assess any campaign's compliance.

Start here

On the typical paid campaign, 7.5 percent of Entrants arrived through a share link, and the middle half of campaigns sat between 3.4 and 15 percent. On a 1,000-Entrant campaign that is 34 to 150 people. If the number you need is bigger than that, the Action alone will not get you there and the other levers in this report, and in Edition 1, matter more.

Biggest measured setting

On widgets of nine or more Actions, the top four positions came with a 10.4 percent referral share against 6.4 percent lower down. On a shorter widget the order made under a point of difference.

Timing and channels

Day one and day two of a two-week campaign carried 21 percent of all share-link clicks between them, and the first week 63 percent. Campaigns reached from eight or more referring sources had a 10.5 percent referral share against 3.4 percent from one or two, and share clicks peaked within a day of the traffic peak in 62 percent of campaigns. Treat the share Action as an amplifier of the promotion you run.

Smallest measured setting

Once one large account was set aside, single-entry campaigns had an 8.0 percent referral share and every other award band sat between 7.8 and 11.0 percent. The data gives only a small reason to set it above 1 and no measured reason to set it above 10.

One thing this study cannot tell you is whether switching the Action on would raise your Entrant count. Campaigns that offered it ended with more Entrants in most reach bands, by about the size of their referral share, and the organizers who offered it also ran longer, larger and more frequent campaigns. Treat the figures above as what campaigns like yours achieved.

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

Table 18 · Referral share of Entrants, full distribution, share campaigns on paid plans, 1 Sep 2025 to 13 Aug 2026

Table 19 · Referral funnel planning ranges, share campaigns on paid plans, 1 Sep 2025 to 13 Aug 2026

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.

Table 20 · Entries awarded per referral, distribution, 1 Sep 2025 to 13 Aug 2026

Table 21 · What to expect by widget position and length, 1 Sep 2025 to 13 Aug 2026

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.

Headline figures cover campaigns starting between 1 September 2025 and 13 August 2026, by campaign start date, the same window as Edition 1, and only campaigns that had ended by 13 August 2026, so that every referral count is final. Entrant counts come from the complete dump supplied by Gleam engineering on 10 September 2026. Share-link clicks come from Gleam's daily viral-click record, 1,232,141 campaign-days across the platform. Every table on this page carries its own window in the caption.

• The all-plan window holds 15,469 campaigns from 3,490 businesses with 100 or more Entrants, a duration of 1 to 365 days and an end date on or before 13 August 2026.
• 3,848 campaigns on Free and Hobby plans are excluded, because those plans do not include the Action. Hobby records 0.05 percent adoption, one campaign, and Free none.
• 1,368 campaigns classified as cryptocurrency or Web3 are excluded, on either a wallet-address entry Action or a cryptocurrency industry label. Their referral share is reported once, in section 7, as a reference figure and never as a benchmark.
• 4 campaigns recording more share-link clicks than Impressions are excluded as instrumentation faults.
• 16 campaigns record more referred Entrants than Entrants and are kept, because the referral count is the Action's completion count and can exceed unique Entrants where one Entrant refers more than once. They are 0.4 percent of share campaigns and do not move any median.
• 11 campaigns offering the Action carry no daily click record and are excluded from funnel tables that need clicks, leaving 4,208 of 4,219.
• Campaigns under 100 Entrants are excluded throughout, the same floor as Edition 1, and it compresses the comparison in section 6, because a campaign that reached 100 Entrants with the Action's help is in the sample and one that fell short without it is not.
• Any cell below 200 campaigns or 50 businesses is suppressed and shown as n/a. Industry cuts use 300 campaigns and 75 businesses, and the invalid-entry cut uses 500 and 100. Suppressed on this page: Premium in Tables 1 and 4, the 1st-position row in Table 6, one cell in Table 10, the narrowed cells in section 6, and every industry beyond three in Table 11.
• No reported cell is dominated by one business. Where one business held more than 10 percent of a cell's campaigns, that business was removed and the cell recounted, and the caption says so. That applied to four award bands in Table 5, all three industries in Table 11 and eight industries in Table 2. Three rows of Table 14 still await that recount. Across all 4,219 share campaigns the largest business holds 7.9 percent of campaigns and 3.7 percent of referred Entrants. The Business plan row in Table 4 fails the rule at 18 percent and is shown with a warning rather than suppressed.

Referral counts are extremely skewed. Half of share campaigns drew fewer than 63 referred Entrants and a quarter drew more than 218, with a long tail above that. An average built on that distribution describes a campaign nobody ran. Every headline figure on this page is a median, the benchmark tables carry the 25th and 75th percentiles, and every cell carries its campaign and business count. Where a confidence interval is given, it comes from a bootstrap that resamples businesses rather than campaigns, 1,000 times, because campaigns from the same business are not independent.

The with-against-without question in section 6 was tested two ways: inside the same business, with every business weighted once, and pooled across businesses inside bands of starting reach and widget length. Section 9 uses Gleam's daily referrer record, classified by domain pattern into no referrer, social, search, email client, giveaway directory, Gleam-hosted page and other site. The classification is unaudited and the caveat box says what it cannot separate.

Every comparison was run four ways before publication. Across the full sample, on one campaign per business, inside the Pro plan alone, and inside bands of widget length or campaign size. A difference that appears only in the first is reported as not holding. Where two settings move together, such as position and widget length, or the Action and Impressions, the comparison is repeated with one of them held inside a band, and the controlled figure is the one published.

• It covers Gleam competitions only, run by businesses on paid Gleam plans that chose Gleam, and no other giveaways.
• It carries no causal design. Nothing here shows that switching the Action on, moving it, or changing its award causes an outcome. Every comparison is between campaigns that already differed.
• It holds no entrant-level record. It cannot say who referred whom, how many people one Entrant referred, whether referred Entrants behave differently afterwards, or whether the same people refer across a brand's campaigns.
• It does not hold the share text. The message organizers pre-fill for sharing lives in the Action's configuration, which the research dataset does not carry.
• Its industry labels are unaudited. 2,849 sites were classified by hand from their domain for Edition 1, and a further 556 were classified from their public website for this edition, of which 376 were labeled with high confidence, 159 medium and 21 low. No blind re-classification audit has been run.
• It does not measure revenue, sales or list quality. Referred Entrants are Entrants, and this study does not claim they are qualified leads.

No customers are named on this page. No customer is named anywhere in the benchmark tables, and no single campaign in the dataset is described in a way that would identify the business behind it.

Edition 2, published 2 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 2, the typical paid-plan Gleam competition offering the Viral Share entry Action drew 7.5 percent of its Entrants through referrals, based on 4,208 campaigns run by 1,026 businesses in the window of 1 September 2025 to 13 August 2026.

• Schmitt, P., Skiera, B. and Van den Bulte, C. (2011). Referral programs and customer value. Journal of Marketing 75(1), 46-59. Bank customer records, 2006 to 2008. Directional comparison only, because the study measures referred customers' lifetime value and this page measures referred Entrants on a giveaway.
• Van den Bulte, C., Bayer, E., Skiera, B. and Schmitt, P. (2018). How customer referral programs turn social capital into economic capital. Journal of Marketing Research 55(1), 132-146. Context only.
• Aral, S. and Walker, D. (2011). Creating social contagion through viral product design. Management Science 57(9), 1623-1639. Randomized field experiment on a Facebook application. Context only, because the design tested is an application feature and the outcome is adoption rather than entry.
• Berman, B. (2016). Referral marketing: Harnessing the power of your customers. Business Horizons 59(1), 19-28. Context only.

We did not benchmark the referral share against a published industry average. The referral benchmarks we found measure share of customers or share of revenue on a referral program, and none of them state a denominator that matches Entrants on a giveaway, so a comparison would put two unlike rates side by side.