A new academic analysis estimates that filtering of emails containing Republican WinRed fundraising links resulted in 142,126 fewer expected first-time donations in 2025, representing an estimated $117 million in lost donor value, while Google disputes the researchers’ methodology and denies that its spam protections favor either political party.

By yourNEWS Media Newsroom

Republican political committees and candidates may have lost an estimated $117 million in fundraising value during two periods in 2025 when emails containing WinRed donation links were disproportionately routed away from recipients’ inboxes, according to a newly published study examining Federal Election Commission donation data and email-placement tests.

The research by Cameron Ellis of the University of Iowa and Lars Powell of the University of Alabama concluded that two distinct filtering episodes corresponded with a combined 142,126 fewer expected first-time contributions through WinRed, the Republican fundraising platform. The researchers calculated the financial impact using FEC records showing the subsequent value of newly acquired donors, according to a report on the findings published by The New York Post.

The researchers did not simply compare total Republican and Democratic fundraising. They separated first-time contributors from repeat donors in itemized FEC records for WinRed and ActBlue, its Democratic counterpart, and examined changes during periods when email-placement testing indicated that messages containing WinRed links were being filtered.

Their analysis identified one episode involving Gmail, Microsoft Outlook and Yahoo from June 30 through Sept. 15, 2025. During that period, the study found that first-time WinRed donations dropped approximately 75%.

A second episode involved Outlook alone between February and April 2025. First-time WinRed contributions fell approximately 83% during that period, according to the researchers.

Together, the study attributed 142,126 missing first-time donations to the two episodes. Ellis and Powell then used FEC contribution records and the longer-term value of newly acquired donors to arrive at the estimated $117 million fundraising impact.

“It’s a pretty big number,” Powell told the Post.

Powell said the effect could also be viewed in terms of the advantage received by Democratic recipients when Republican fundraising messages did not reach inboxes.

“You can really think of that as an in-kind gift to ActBlue recipients,” he said.

Google rejected that characterization and challenged the study itself.

A Google spokesperson said the analysis “relies on deeply flawed methodology that fundamentally misunderstands how Gmail works,” arguing that the $117 million figure represents a theoretical estimate rather than money demonstrably lost in actual transactions.

“Our spam protections apply equally to everyone, regardless of political affiliation and we encourage bulk senders to follow best practices to reach inboxes,” the spokesperson said.

Google also emphasized users’ ability to control how their email is handled.

“People are in full control of their inboxes, with the power to mark messages as spam, block senders, or unsubscribe at any time,” the spokesperson said.

Microsoft declined to comment on the findings, while Yahoo did not respond to a request for comment.

Ellis and Powell acknowledged that publicly available evidence cannot establish the internal intent of companies operating proprietary filtering systems.

“We can’t read their minds,” Powell said.

The researchers said, however, that independent email-placement testing gave them a way to identify when the treatment of WinRed-linked messages changed.

“Inbox placement data and paired email tests showed when the filtering happened and supported that WinRed links were the trigger,” Powell said.

Ellis and Powell described their broader conclusion this way: “a narrow, link-level placement decision by a small number of mailbox providers produced a measurable and economically meaningful drop in political fundraising, and that the mechanism was observable only because private paired-test evidence leaked into public reporting.”

The study traces the longer of the two filtering periods to June 30, 2025, when Republican consulting firm Targeted Victory alerted Google that fundraising emails containing WinRed links were being directed to spam.

Targeted Victory later publicly disclosed testing that it said demonstrated different treatment depending on which political fundraising link appeared in otherwise identical messages.

The firm, whose clients included Rep. Steve Scalise of Louisiana, Sen. Marsha Blackburn of Tennessee and the National Republican Senatorial Committee, tested messages containing WinRed and ActBlue donation links. According to the 2025 reporting on those tests, the WinRed-linked versions were repeatedly identified as suspicious or dangerous, while the corresponding ActBlue-linked versions reached inboxes without the same warnings.

The consulting firm said the pattern persisted during June and July 2025.

According to the account of the testing, the emails were otherwise identical, making the fundraising link the principal variable being examined.

Targeted Victory brought its findings to Google after clients complained about their messages being sent to spam. The subsequent dispute eventually focused on SURBL, a service whose data were being used in email screening.

The researchers identified Sept. 15, 2025, as the end of the broader filtering episode. Google stopped using SURBL in connection with the filtering at that point, and the study said Republican email placement then “normalized.”

The change followed weeks of scrutiny over how emails containing WinRed links were being handled.

Google also adjusted its approach to political email during the period and later introduced a “verified sender program” under which political campaigns and PACs can bypass Gmail spam filters if they remain compliant with applicable policies. The program is scheduled to take effect Tuesday.

Powell said he had not communicated with Google about the study or its conclusions.

He nevertheless offered a business explanation for why an email provider might want to address complaints about filtering.

“They make money by having people send more emails. If people start changing, that’s bad for their business model,” Powell said. “I think that’s when we find out whether they’re more blue or green.”

Republican officials cited the study as support for concerns they had raised previously about political email delivery.

Bernadette Breslin, press secretary for the National Republican Senatorial Committee, said, “Google’s biased algorithms sidelined Republican fundraising emails for years while boosting Democrats’ donor outreach.”

“The NRSC has helped lead the investigation into Google’s suppression of mail to conservative subscribers, and this long-overdue update is a chance to level the playing field for online fundraising,” Breslin added.

A Republican adviser familiar with the dispute called the research a “massive proof point in what we already knew to be true” and a “smoking gun that con[firms]” the pattern.

Those assessments represent the views of Republican officials and advisers; Google maintains that its systems apply the same spam standards without regard to political affiliation.

The estimated scale of the researchers’ finding is substantial compared with national Republican committee fundraising in 2025. The Republican National Committee, National Republican Congressional Committee and National Republican Senatorial Committee raised a combined $377 million during the year, according to the figures cited with the study.

Gmail, meanwhile, has more than 3 billion monthly users, giving changes in email placement the potential to affect large numbers of political messages.

The findings add new financial estimates to a dispute that had previously centered largely on inbox-placement tests and accusations over how political fundraising links were treated. The study’s central contention is not that every filtered email would have generated a donation, but that observed changes among first-time WinRed donors during identified filtering periods can be used to estimate contributions that otherwise would have been expected.

Google disputes that methodology and rejects the suggestion that its filtering practices are politically discriminatory.

Original article