Your subject line, your sender name, your domain reputation — all of it can be perfect, and your email can still land in spam. That gap between effort and outcome is the quiet frustration behind a much louder question now being asked in Washington: is Gmail's spam filter applied evenly?
On August 28, 2025, Federal Trade Commission Chairman Andrew Ferguson sent a letter to Alphabet CEO Sundar Pichai warning that Gmail's spam filtering may violate the FTC Act if it treats comparable senders differently based on politics. Google denies it. Courts have dismissed similar claims. Researchers disagree about what the data shows.
This article does not try to settle who is right. It lays out what the regulator, Google, researchers and courts actually say — and then focuses on what increased scrutiny of filtering transparency could mean for commercial senders.
What you will learn:
- What the FTC letter actually says, and what it does not do
- Google's technical defence, in its own words
- The evidence on both sides, with the numbers
- Why proving political intent is so difficult
- What this could mean for commercial deliverability
The Short Version
If you only read one section, read this.
- The FTC chair warned Alphabet that alleged partisan Gmail filtering may violate the FTC Act. It is a warning, not a filed case.
- The allegation rests on reporting and tests, including a WinRed-versus-ActBlue comparison and a 2022 academic study showing a large disparity.
- Google says its filters use objective signals — user spam complaints and sender behaviour — applied equally to all senders.
- Courts and the FEC have dismissed similar claims, on standing and legal grounds rather than a finding that filtering is neutral.
- A 2026 working paper finds large real-world fundraising impacts but does not claim the filtering was intentional.
- The lasting issue is filtering transparency, which could eventually affect how commercial senders see and contest inbox placement.
Key Terms, in Plain English
The debate uses a lot of jargon. Here is what the terms actually mean.
- Section 5 of the FTC Act — the provision banning "unfair or deceptive acts or practices." It is the FTC's main consumer-protection tool.
- Unfair practice — conduct that causes or is likely to cause substantial injury consumers cannot reasonably avoid, and that is not outweighed by benefits to consumers or competition.
- Deceptive practice — a material representation or omission likely to mislead a reasonable consumer.
- Algorithmic filtering — automated systems that decide which messages reach an inbox, usually by scoring signals.
- Inbox placement — where a message lands: inbox, spam, or nowhere. It is the outcome senders care about.
- Spam complaint rate — the share of delivered messages that recipients mark as spam. It is the single most influential filtering signal.
- Link reputation — a provider's learned trust in the domains and URLs a message contains. A donation platform or shortener can accumulate a reputation of its own.
- Dormant Commerce Clause — a legal doctrine limiting how far one state's rules can reach into interstate commerce. It has appeared in related filtering disputes.
What Actually Happened
The dispute did not begin with the FTC. It began with reporting, and it built on years of complaints.
| Date |
Development |
| Mar 2022 |
North Carolina State University researchers publish a study on political bias in spam filters. |
| Oct 2022 |
The Republican National Committee sues Google over alleged spam filtering of its fundraising mail. |
| Jan 2023 |
The Federal Election Commission dismisses a related Republican complaint. |
| Aug 2023–Jul 2024 |
A federal district court dismisses the RNC lawsuit, later with prejudice. |
| Aug 13, 2025 |
The New York Post reports tests by consulting firm Targeted Victory comparing WinRed and ActBlue links. |
| Aug 28, 2025 |
FTC Chairman Ferguson sends the warning letter to Alphabet CEO Sundar Pichai. |
| Jan 16, 2026 |
The Ninth Circuit affirms the dismissal of the RNC's lawsuit. |
| Jul 1, 2026 |
The FTC issues a proposed policy statement on ideological bias in AI systems under executive order 14365. |
Two things are worth noting about this timeline.
- The FTC letter is a warning, not an enforcement action — reporting at the time said the agency was not announcing a new investigation.
- The underlying allegation has been litigated and dismissed before, which is why the letter's framing matters so much.
What the FTC Letter Does — and Doesn't Do
It is worth being precise about the letter, because it has been described in ways that overstate and understate it.
What it does:
- States the FTC's understanding of the reporting on Gmail's filtering.
- Cites Section 5 of the FTC Act and explains the unfairness and deception standards.
- Frames the issue as consumer harm, not only political fairness.
- Warns that inconsistent conduct "could lead to an FTC investigation and potential enforcement action."
- Puts Alphabet on notice and invites a response.
What it does not do:
- It does not open a formal investigation by itself.
- It does not accuse Google of a specific, charged violation.
- It does not impose any penalty or deadline.
- It does not resolve whether the alleged disparity is intentional.
The Allegation, in the FTC's Words
The letter is short and direct. Ferguson wrote:
"My understanding from recent reporting is that Gmail's spam filters routinely block messages from reaching consumers when those messages come from Republican senders but fail to block similar messages sent by Democrats."
He tied the concern to consumer harm, not just politics:
"If Gmail's filters keep Americans from receiving speech they expect, or donating as they see fit, the filters may harm American consumers and may violate the FTC Act's prohibition of unfair or deceptive trade practices."
The reporting the letter cites is a New York Post story from August 13, 2025. It described tests by Targeted Victory, a consulting firm that has worked with Republican groups, in which otherwise identical emails were sent through Gmail — the only difference being whether they contained a WinRed or an ActBlue donation link. According to the report, the WinRed versions were flagged as spam and the ActBlue versions were not.
Ferguson also cited comments submitted to the FTC's earlier request for information on Technology Platform Censorship.
What Google Says
Google's response was immediate and categorical. A company spokesperson said:
"Gmail's spam filters look at a variety of objective signals — like whether people mark a particular email as spam, or if a particular ad agency is sending a high volume of emails that are often marked by people as spam. This applies equally to all senders, regardless of political ideology."
Google added that it would review the letter and "look forward to engaging constructively."
That explanation is not a dodge. It describes how modern spam filtering actually works — and it is the core of the disagreement, because those same "objective signals" can produce very different outcomes for senders with different list-building practices.
Who Says What
A quick map of the positions.
| Actor |
Position |
| FTC Chairman Ferguson |
Alleges possible partisan filtering; warns of potential FTC Act violations. |
| Google |
Denies bias; filters use objective user and sender signals applied equally. |
| RNC and Republican committees |
Allege discrimination and fundraising harm; litigation so far unsuccessful. |
| Federal courts |
Dismissed the RNC claims on standing and legal elements. |
| FEC |
Found Google's commercial-purpose explanation credible; dismissed complaint. |
| NC State researchers |
Found a disparity but no evidence of deliberate attempts to influence voters. |
| Ellis and Powell |
Find large donation impacts but cannot separate technical from partisan causes. |
The Evidence on Both Sides
This is where the story gets genuinely contested. There is real evidence pointing toward a disparity, and real evidence and reasoning pointing away from intentional bias.
Evidence cited by critics of the filtering
- The 2022 NC State study. Researchers subscribed over 100 email accounts to hundreds of campaigns during the 2020 election.
- Gmail marked 67.6% of right-leaning campaign emails as spam.
- It marked 8.2% of left-leaning emails as spam.
- Outlook and Yahoo showed the opposite lean, marking more left-leaning mail as spam.
- The Targeted Victory tests. Identical emails differing only by WinRed or ActBlue link reportedly landed differently.
- Spam-trap reporting. Security researcher Brian Krebs reported that WinRed-linked mail hit spam traps at roughly four times the rate of ActBlue mail in late July 2025.
- The Ellis and Powell working paper. "Real Impacts of Political Spam Filtering," by Cameron Ellis (University of Iowa) and Lars Powell (University of Alabama), reports:
- Inbox placement for Republican committee mail swings by about 30 percentage points across providers in a normal week.
- During a 78-day episode in 2025 when WinRed-linked messages were routed to spam, first-time Republican donations fell nearly 75%, with no matching drop for ActBlue.
- A separate Outlook episode reportedly cut Republican inbox placement from about 88% to 16%, producing an 83% first-time donor decline.
- Together, the authors estimate roughly 142,000 missing first-time donations, checked against 125 placebo dates.
Evidence and reasoning that cut the other way
- The NC State authors' own caveats.
- The study covered a limited period.
- It could not control for how many emails each campaign sent.
- It could not control for how long each campaign had been active.
- It could not control for how many users outside the test marked a given email as spam.
- The authors stated they had "no reason to believe there were deliberate attempts" to influence voters.
- The FEC dismissal. In January 2023, the FEC found Google "credibly supported its claim that its spam filter is in place for commercial reasons" and found no reason to believe it made prohibited in-kind contributions.
- The courts.
- A federal district court found the RNC's bad-faith allegation did "not rise above the speculative level."
- The Ninth Circuit affirmed on January 16, 2026.
- It warned that "imposing a duty of care would risk deterring beneficial spam filtering activity."
- The technical explanation. Gmail's filters weight user behaviour — opens, replies, deletions and, most heavily, spam complaints. A sender whose recipients frequently mark its mail as spam will see worse placement regardless of politics.
- The volume factor. The RNC sends very high volumes; the district court noted it "sends out a significant number of emails to individuals on its list."
- The design argument. As one prominent internet-law commentator put it, all spam filters are "biased" by definition — that is what filtering means. The open question is whether a specific bias is political.
Putting the two sides side by side
| Claim |
Counterpoint |
| Gmail filters more Republican mail (67.6% vs 8.2%) |
Study could not control for volume, list age, or user spam markings |
| Identical emails with WinRed vs ActBlue links behaved differently |
Small controlled tests may not generalise; link reputation can reflect past user reports |
| Republican donations fell ~75% during the 2025 episode |
Working paper authors do not claim intent and cannot isolate the cause |
| A pattern persists across providers and time |
Different providers lean in different directions, which argues against a single partisan design |
The honest summary: there is credible evidence of a disparity, and credible reason to doubt that the disparity is intentional. Those two things can both be true, and the gap between them is exactly what regulators and courts are arguing about.
How Gmail's Filter Actually Decides
Understanding the mechanics explains why the dispute is so hard to resolve. Filters weigh many inputs, and most are behavioural.
- Recipient engagement. Opens, replies, stars and time spent reading push a sender up.
- Spam complaints. The strongest negative signal. Recipients marking mail as spam is the clearest possible "don't show me this."
- Deletions without reading. A weaker negative signal than a complaint, but it accumulates.
- Sender reputation. Domain and IP history, authentication, and bounce behaviour.
- Link and domain reputation. Whether the URLs in the message — including donation platforms and shorteners — have been associated with spam.
- Content signals. Structure, image-to-text ratio, and patterns common to bulk mail.
- List quality. Spam traps, invalid addresses and sudden volume spikes.
Notice what is missing from that list: a political affiliation field. The argument is about whether the signals above correlate with political affiliation strongly enough to produce a partisan-looking outcome.
Why This Is Hard to Prove
If the question is "is Gmail biased," the technical answer is unsatisfying: it depends what you mean.
- Filters are supposed to discriminate. Their job is to separate wanted from unwanted mail. A filter that treated all senders identically would be a filter that did not work.
- The strongest signals are behavioural. Those behaviours reflect how a list was built, how often it is mailed, and how relevant it is — not the sender's politics.
- Correlation is not causation. A disparity between two groups of senders can exist even when the filter contains no political variable.
- Intent is the legal crux. Proving a provider acted because of politics, rather than because of signals that happen to correlate with politics, is a high bar.
- Real harms and real benefits coexist. The same filtering that diverts a small number of wanted emails also blocks a large volume of unwanted ones.
The district court captured the tension well: filtering that diverts a small number of wanted emails may not be "substantially injurious" to consumers, even when it clearly harms a sender.
Common Misconceptions
A few ideas circulate that are worth correcting.
- "The FTC has charged Google." No. The agency sent a warning letter and flagged possible violations.
- "Courts cleared Google of bias." Not quite. The cases were dismissed on standing and legal elements, not after a full factual finding on bias.
- "The study proves intentional bias." The NC State authors found a disparity and explicitly did not allege deliberate attempts to influence voters.
- "Only Republicans are affected." Outlook and Yahoo leaned the other way in the same study, and spam filters misclassify legitimate mail from across the spectrum.
- "This is only about politics." The same transparency questions apply to any sender class that believes placement is unfair.
- "Nothing will change for commercial senders." Possibly. But disclosure rules written for one class of senders rarely stay confined to it.
What It Means for Commercial Senders
Here is the practical reason this story matters beyond politics. The principle at stake is filtering transparency.
The immediate dispute is about political mail. But if regulators establish that providers must explain or disclose placement decisions for one class of senders, the logic can extend to commercial senders who have long complained about inconsistent inbox placement — the same complaint, without the political charge.
Three threads are worth watching.
- The FTC's broader interest in algorithmic filtering. The August 2025 letter sits alongside a similar FTC warning to Apple News and the agency's request for information on platform censorship. In July 2026 the FTC issued a proposed policy statement on ideological bias in AI systems under executive order 14365, drawing First Amendment objections.
- Proposals for disclosure, not code review. The Ellis and Powell paper proposes a deliberately modest remedy: when a provider changes how it treats a named payment processor or a defined class of senders, it should disclose the change to affected senders and allow a neutral auditor to review it. No proprietary code, no new agency — just notice.
- The commercial parallel. Commercial senders already have some of this data: Postmaster Tools, seed-list testing and placement analytics. If transparency rules arrive, that data becomes more standardized and more consequential.
For deliverability teams, the lesson is not "wait for regulation." It is that placement is measurable, and measurement is the defence.
What to Watch Next
Keep an eye on these developments.
- Whether the FTC escalates from a warning letter to a formal investigation.
- How Alphabet responds to the letter, and whether it publishes any filtering data.
- Whether Congress advances email-filtering transparency or political-email legislation.
- The outcome of the FTC's AI-bias policy statement process under executive order 14365.
- Whether any commercial sender group picks up the transparency argument.
- Any follow-up research that can separate link reputation from political intent.
How to Monitor Your Own Inbox Placement
Whatever happens with the FTC, this is the discipline that protects commercial senders.
- Track placement by provider, not in aggregate. Gmail, Outlook and Yahoo behave differently. Use the inbox placement rate calculator to quantify it.
- Watch your spam complaint rate like a hawk. Keep it well under 0.1% and never near 0.3%. The spam complaint rate calculator shows where you stand.
- Authenticate fully. SPF, DKIM and DMARC alignment are table stakes and increasingly mandatory. See email deliverability in 2026.
- Clean your list continuously. Disengaged addresses are the fastest route to complaints and filtering. The list decay calculator helps you size the problem.
- Score your content before sending. A spam score checker catches technical triggers; the sender reputation checker shows how you look to providers.
- Document your practices. If placement is ever questioned, evidence of consent, hygiene and complaint management is your strongest asset. The compliance audit framework is a good starting point.
For the fundamentals, see what triggers email client spam filters and why your emails are going to spam.
Key Takeaways
- On August 28, 2025, FTC Chairman Andrew Ferguson warned Alphabet that alleged partisan Gmail filtering may violate the FTC Act — a warning, not a filed enforcement action.
- The allegation rests on reporting and tests, including Targeted Victory's WinRed-versus-ActBlue comparison and a 2022 NC State study showing a large disparity.
- Google denies bias, saying filters use objective signals such as user spam complaints and sender behaviour, applied equally regardless of ideology.
- Courts and the FEC have dismissed similar claims. The Ninth Circuit affirmed the RNC's dismissal in January 2026, citing standing and the value of beneficial spam filtering.
- Researchers disagree on causation. A 2026 working paper finds large real-world donation impacts but does not claim the filtering was intentional.
- The lasting issue is filtering transparency, which could extend to commercial senders.
- The practical response is unchanged: measure placement, manage complaints, authenticate, and document your practices.
Sources and Further Reading
Related Articles
Related tools: Measure placement with the inbox placement rate calculator, track complaints with the spam complaint rate calculator, check your standing with the sender reputation checker, and test content with the spam score checker.
This article presents allegations, responses and research for informational purposes and does not take a position on whether any provider has acted unlawfully or improperly. It is not legal advice. Reported allegations are contested; verify current developments before relying on them.
