86.4% of the 4,195 brands seen sponsoring newsletters appeared in only one newsletter

Findings

1. 4,195 brands appeared in 10,482 sponsor placements across 2,531 newsletters, and 86.4% of those brands (3,624) appeared in only one newsletter.
2. The 66 brands seen in 10 or more newsletters (1.6% of brands) accounted for 17.1% of placements (1,789 of 10,482), led by Wispr Flow in 91 newsletters.
3. 26.5% of brand and newsletter pairings ran in two or more issues (1,527 of 5,770 pairings in newsletters with four or more issues read).
4. The share of newsletters that ran at least one sponsor rose with list size, from 6.4% under 1,000 subscribers (32 of 503) to 21.4% at 250,000 or more (34 of 159).
5. Newsletters on beehiiv ran sponsors 5.2 times as often as Substack newsletters from the same technology-detection list (19.2% of 1,933 against 3.7% of 1,222), and 2.1 to 3.0 times as often within each size band.
6. Among newsletters ranked on Substack, tech, AI, business and marketing titles ran sponsors 2.3 times as often as culture and politics titles (18.1% of 1,128 against 7.9% of 3,679), and 1.9 to 2.2 times as often within each size band.

How to read these numbers. Every count comes from the issues in the database. A brand counted in one newsletter appeared in one of the 75,504 newsletters with an issue read in the window, and it may also have run in newsletters the database never read. Software found the placements by reading each issue, and only text that reads as an ad was kept, so the counts are a floor. Findings 4 to 6 use a rate sample of 14,645 newsletters that were found through platform-wide lists and had three or more issues read, which leaves out newsletters that were found because they carry ads.

Who sponsors newsletters

Wispr Flow91 newsletters

Attio38 newsletters

Factor36 newsletters

Kalshi35 newsletters

144033 newsletters

Pique29 newsletters

Klover29 newsletters

Apollo.io29 newsletters

Frontieras28 newsletters

Levanta26 newsletters

Roku23 newsletters

StoreClaw23 newsletters

Source: Carroway Research analysis of 10,482 sponsor placements by 4,195 brands in 260,443 newsletter issues. Brands are grouped by web domain. Placements and first and last dates per brand are in top-sponsors.csv.

Wispr Flow, a voice dictation app, appeared in 91 newsletters with 143 placements, more than twice as many newsletters as the next brand. Attio (38 newsletters), Factor (36), Kalshi (35) and 1440 (33) followed. The 12 brands seen in the most newsletters include software (Wispr Flow, Attio, Apollo.io), consumer goods (Factor, Pique), money apps and investment offers (Kalshi, Klover, Frontieras) and a newsletter (1440).

Newsletters also advertise in other newsletters. At least 6 of the 66 brands seen in 10 or more newsletters are newsletters themselves: 1440 (33 newsletters), Morning Brew (22), Mindstream (18), The Hustle (17), Superhuman AI (16) and TLDR (10).

Most brands appeared once. 3,624 of the 4,195 brands (86.4%) appeared in a single newsletter. Those brands still account for 58.1% of placements (6,095), since many ran in more than one issue of their newsletter. 571 brands (13.6%) appeared in two or more newsletters and account for 41.9% of placements (4,387). Measured by placements, the top 10% of brands (420) account for 48.0% of all placements and the top 1% (42) for 14.6%.

A quarter of pairings ran again

A pairing is one brand in one newsletter. In newsletters with four or more issues read in the window, 26.5% of pairings (1,527 of 5,770) ran in two or more issues. The share grows with the number of issues read, from 14.1% where two or three issues were read (43 of 305 pairings) to 29.1% where ten or more were read (1,099 of 3,771), because each extra issue gives a repeat another chance to show up. The 26.5% counts only newsletters with four or more issues read, since one issue cannot show a repeat and two or three give it few chances.

Larger lists run sponsors more often

Under 1,0006.4%

1,000 to 9,9998.4%

10,000 to 49,99913.5%

50,000 to 249,99917.1%

250,000 or more21.4%

Source: Carroway Research. 9,556 newsletters with a public subscriber count and three or more issues read, found through platform-wide lists. Newsletters and sponsored newsletters per band are in sponsor-rate-by-size.csv.

Among the 9,556 newsletters in the rate sample with a public subscriber count, the share that ran at least one sponsor rose in every band: 6.4% under 1,000 subscribers (32 of 503), 8.4% at 1,000 to 9,999 (433 of 5,170), 13.5% at 10,000 to 49,999 (378 of 2,807), 17.1% at 50,000 to 249,999 (157 of 917) and 21.4% at 250,000 or more (34 of 159). The median number of issues read per newsletter was 9 to 12 in every band, so the rise does not come from reading more issues of larger newsletters.

Per issue, the rise is steepest at the top. 6.7% of issues from newsletters with 250,000 or more subscribers carried a sponsor (114 of 1,696), against 1.8% to 2.7% in the two bands under 10,000. Most newsletters in every band, the largest included, ran no sponsor that the database found in the window. Of the 5,089 newsletters in the rate sample with no public count, 477 ran a sponsor.

beehiiv against Substack

Substack newsletters make up most of the newsletters read (68,034 of 75,504), and each platform's newsletters were found in different ways, so a fair comparison needs newsletters found the same way. Technology-detection lists of websites running each platform give that. From those lists, 19.2% of 1,933 beehiiv newsletters ran at least one sponsor, against 3.7% of 1,222 Substack newsletters, 5.2 times the rate.

List size could explain part of that gap, so we also compared the platforms within subscriber bands, adding Substack newsletters from its other lists. The gap held: 3.0 times under 10,000 subscribers (23.6% of 110 against 7.8% of 5,067), 2.9 times at 10,000 to 49,999 (37.1% of 70 against 12.7% of 2,691) and 2.1 times at 50,000 or more (34.3% of 70 against 16.5% of 994). The beehiiv rows are small. One likely reason for the gap is that beehiiv runs an ad network for newsletters on its platform, and the data does not separate network ads from ads a publisher sold directly.

Business topics run sponsors about twice as often

Tech21.4%

AI18.8%

Business16.4%

Marketing15.6%

Sports14.5%

Finance12.8%

Food12.6%

News and media11.9%

Health9.5%

Climate9.0%

Culture8.1%

Politics7.7%

Education4.3%

Source: Carroway Research. Newsletters from Substack's category rankings with three or more issues read, in topics with 100 or more such newsletters. Counts per topic are in sponsor-rate-by-topic.csv.

Substack's category rankings give every ranked newsletter a topic, which gives the most consistent topic labels in the data. Tech newsletters ran sponsors most often (21.4% of 234), followed by AI (18.8% of 367), business (16.4% of 348) and marketing (15.6% of 179). Culture (8.1% of 2,400) and politics (7.7% of 1,279), the two largest groups, sat near the bottom, above education (4.3% of 117).

The four business topics also have larger lists, with median public counts of 11,000 to 17,000 subscribers against 9,000 for culture and for politics. Within each subscriber band the gap held at 1.9 to 2.2 times: 13.1% against 6.3% under 10,000 subscribers (320 and 1,625 newsletters), 19.0% against 10.1% at 10,000 to 49,999 (479 and 999) and 26.9% against 12.1% at 50,000 or more (201 and 420). For B2B advertisers, the newsletters written for business readers are the ones most used to running ads.

Method

Source. A database of newsletter sponsorships collected from newsletter issues published between June 1 and September 22, 2026. The database lists 207,745 newsletters found through platform-wide lists (Substack's sitemap, category rankings, search and recommendations; Ghost's public directory; technology-detection lists of websites running beehiiv, Substack, Kit or Ghost), newsletter directories, ad marketplaces and web searches. It reads a set number of recent issues from each tracked newsletter's public web archive and stores their text. We read the database on September 23, 2026 and counted only rows stored before 00:00 UTC that day. The database holds 68,034 Substack newsletters of the 75,504 read, so rates by platform, size and topic are reported separately. The window runs from June 1 to September 22.

Issues. 260,443 issues from 75,504 newsletters were published in the window and read by the sponsor extractor. The reader favors recent issues, so August holds 108,476 of them and June 26,719, and the mix of newsletters changes by month (46,414 newsletters in September against 8,371 in June). The monthly share of issues with a sponsor (1.2% to 3.8%) moves with that mix, so we do not read it as a trend. Monthly and platform counts are in coverage-by-month.csv and coverage-by-platform.csv.

How placements were found. Two extractors read every issue. A pattern matcher looks for a sponsor header, such as "sponsored by", "brought to you by" or "together with", followed by a link, and records the link. A language model reads the issue text and lists paid placements, with instructions to skip the author's own products, affiliate links and companies mentioned in the news. The two produced 23,947 rows for the issues in the window.

What counts as a placement. We kept a row when the pattern matcher found it with a link, or when its text reads as an ad: words such as "sponsored", "brought to you by", "presented by", "in partnership with", "promo code" or "free trial", or "together with" or "partner message" at the start of the text. 11,237 rows passed. We removed 103 rows from organizations flagged in the database as not advertisers and 50 rows where the advertiser's domain or name matched the newsletter's own. A placement is one advertiser in one issue: 11,084 placements in 7,378 issues from 2,531 newsletters.

Brands. Brands are grouped by the registrable web domain of the advertiser, so get.apollo.io and apollo.io count as one brand. We merged domains that we checked by hand as the same advertiser, such as wispr.com and wisprflow.com for Wispr Flow and factor.com and factor75.com for Factor. On link services where the subdomain names the advertiser, such as mobile app links, the full host is the brand. 178 placements went through click-tracking and link-shortening domains that hide the advertiser. They count in the sponsor rates and are left out of brand counts. The other 10,906 placements come to 10,482 brand placements, because a brand counts once per issue: 424 placements were a second organization or domain of a brand already counted in that issue. That leaves 4,195 brands. Every rule and alias is in queries.sql.

Rate sample. Findings 4 to 6 use newsletters found through platform-wide lists that had three or more issues read in the window: 14,645 newsletters, 1,511 of them with at least one sponsor. Newsletters found through ad marketplaces, rate-card pages, newsletter directories or searches for sponsored newsletters are left out, since many of them were found for carrying ads. A newsletter ran a sponsor when at least one of its issues read in the window held a placement.

Subscriber counts. The public figures the platforms show, mostly rounded to the thousand, available for 9,556 of the 14,645 newsletters in the rate sample.

Topics. For newsletters ranked on Substack, the topic is mostly Substack's own ranking category, grouped into a shorter list: culture includes the music, art, literature, history, faith and travel categories, among others; news and media includes news, podcasts and film; health includes parenting. Keyword rules on each newsletter's name and tagline first move titles about AI, marketing, software development, productivity or crypto into those topics. Topics with fewer than 100 ranked newsletters in the rate sample are left out.

Platforms. The first comparison uses only the technology-detection lists for beehiiv and Substack. The size-band comparison uses every newsletter in the rate sample with a public subscriber count that came from the beehiiv detection list or from one of the Substack lists.

Exclusions.

• Issues published outside the window or without a publication date, and rows stored after the snapshot.
• Sponsor rows whose text does not read as an ad, which is more than half of the 23,947 rows.
• Sponsor logos on newsletters' advertise pages and listings on ad marketplaces, which do not tie a brand to a dated issue.
• Published ad rates. The database stores prices from advertise pages, but they mix newsletter slots with magazine, podcast and event packages, so this note does not report them.

Data

The twelve CSV files below hold every number in this note, and queries.sql holds the exact query behind each one. top-sponsors.csv, brands-by-newsletter-count.csv, brand-concentration.csv and newsletters-as-advertisers.csv describe the brands. repeat-pairings.csv holds the repeat runs. sponsor-rate-by-size.csv, platforms.csv, sponsor-rate-by-topic.csv and topic-by-size.csv hold the sponsor rates. coverage.csv, coverage-by-month.csv and coverage-by-platform.csv trace the counts from the database to the rows analyzed. They are free to reuse with credit to Carroway Research.

• The 12 brands seen in the most newsletters CSV, 12 rows, 0.6 KBOne row per brand: rank, brand, the web domain it is grouped under, newsletters it appeared in, placements (issues), and the first and last issue dates it was seen.
• Brands by the number of newsletters they appeared in CSV, 5 rows, 0.2 KBBrands grouped by how many newsletters they appeared in, with each group's share of brands and of placements.
• Share of placements held by the most active brands CSV, 5 rows, 0.3 KBBrands, share of brands, placements and share of placements for the top 1% and top 10% of brands by placements, for brands in 2 or more and 10 or more newsletters, and for all brands.
• Newsletters among the brands seen in 10 or more newsletters CSV, 6 rows, 0.2 KBSix advertisers that are themselves newsletters, labeled by hand from the 66 brands seen in 10 or more newsletters, with newsletters and placements for each.
• Brand and newsletter pairings that ran in two or more issues CSV, 5 rows, 0.2 KBPairings of one brand and one newsletter, grouped by how many of that newsletter's issues were read, with the number and share that ran in two or more issues.
• Newsletters that ran a sponsor, by subscriber count CSV, 5 rows, 0.4 KBOne row per subscriber band in the rate sample: newsletters, newsletters with at least one sponsor, issues read, issues with a sponsor, both shares, and the median issues read per newsletter.
• beehiiv and Substack newsletters that ran a sponsor CSV, 8 rows, 0.6 KBNewsletters and sponsored newsletters on each platform, first from the same technology-detection lists, then within subscriber bands, with each beehiiv rate as a multiple of the Substack rate.
• Substack-ranked newsletters that ran a sponsor, by topic CSV, 13 rows, 0.4 KBOne row per topic with 100 or more ranked newsletters: newsletters, newsletters with at least one sponsor, the share, and the median public subscriber count.
• Business topics against culture and politics, within subscriber bands CSV, 10 rows, 0.7 KBTech, AI, business and marketing newsletters against culture and politics newsletters ranked on Substack, overall and within each subscriber band, with the rate of the first group as a multiple of the second.
• Coverage and exclusions CSV, 22 rows, 1.7 KBEvery step from the newsletters and issues on file to the placements, brands and rate sample analyzed, with the window and snapshot time.
• Issues read and placements by month CSV, 4 rows, 0.2 KBIssues read, newsletters, issues with a sponsor and placements for each month of the window.
• Issues read and placements by platform CSV, 7 rows, 0.3 KBNewsletters, issues read, issues with a sponsor, newsletters with a sponsor and placements for each newsletter platform, across every newsletter read.