Font Awesome Did Everything Right on Email. Gmail Said No Anyway.

What Happened

Font Awesome — the icon toolkit used on hundreds of millions of websites — published a detailed account of their struggle to land emails in Gmail inboxes. The numbers should have been on their side: a 99% sender reputation across industry monitoring tools, fully configured SPF, DKIM, and DMARC authentication, clean subscriber lists with low bounce rates, and sending practices that follow every best-practice guide Google has ever published.

None of it mattered. Gmail routed their messages to spam anyway.

This isn't a story about a company that got sloppy with email hygiene. Font Awesome sends emails that millions of developers actually want — license confirmations, icon kit updates, security notices. They did everything the deliverability playbooks prescribe, and Gmail's filters still treated them like a penny-stock newsletter.

The blog post struck a nerve on Hacker News, pulling over 200 upvotes, because every SaaS operator reading it recognized the same experience. The comment threads filled with war stories from companies of all sizes fighting the same invisible opponent.

Why It Matters

### The Black Box Problem

Gmail processes over 1.8 billion accounts' worth of email. Its spam filtering uses machine learning models that Google does not document, does not explain, and does not offer meaningful appeals against. When your emails get filtered, Google's official guidance amounts to 'follow best practices and wait' — which is indistinguishable from 'we won't help you.'

Google Postmaster Tools is supposed to be the diagnostic layer. It shows domain reputation, IP reputation, authentication results, and spam complaint rates. Font Awesome's Postmaster dashboard reportedly showed clean metrics across the board. The disconnect between what Postmaster Tools reports and what actually happens to your email is the central frustration. You can have a green dashboard and a spam folder full of your own messages.

### The 2024 Rules Changed Nothing Fundamental

In February 2024, Google rolled out stricter requirements for bulk senders: mandatory SPF and DKIM authentication, DMARC policies, one-click unsubscribe headers, and spam complaint rates below 0.3%. The industry spent months scrambling to comply. Companies that were already doing email correctly — like Font Awesome — already met every requirement.

The uncomfortable truth is that Google's published rules are a floor, not a ceiling. Meeting them gets you past the first gate. The second gate — the ML-driven content and behavioral filtering — has no documented rules at all. Your email's fate depends on factors like: how many recipients in a given cohort opened similar messages, how many hit "Report Spam" on emails that look structurally similar to yours, and what Google's models infer about your sending pattern relative to millions of other senders.

This creates a Kafka-esque situation where compliance is necessary but provably insufficient.

### It's Not Just Font Awesome

The HN discussion surfaced a pattern that email infrastructure engineers have known for years: Gmail deliverability problems disproportionately hit mid-size SaaS senders. Tiny senders fly under the radar. Massive senders (banks, airlines, Fortune 500) have dedicated deliverability teams and, in some cases, direct relationships with Google's postmaster team. The companies in the middle — sending tens of thousands to low millions of emails monthly, with legitimate opt-in lists — are the ones getting crushed.

Developers in the thread reported identical experiences across different ESPs (SendGrid, Postmark, Amazon SES, Mailgun). Switching providers sometimes helps temporarily — a new IP pool can buy you a few weeks of inbox placement — but the filtering eventually catches up. The problem is structural, not provider-specific.

What This Means for Your Stack

### Subdomain Isolation Is No Longer Optional

If you're sending both transactional and marketing email from the same domain (or worse, the same subdomain), stop. Gmail's filtering doesn't just evaluate individual messages — it evaluates your domain's aggregate sending behavior. One marketing campaign with slightly elevated complaints can tank deliverability for your password reset emails.

The minimum viable setup in 2026: - `mail.yourdomain.com` for transactional (password resets, receipts, security alerts) - `news.yourdomain.com` for marketing and product updates - Separate IP pools for each, ideally on different ESPs - DMARC set to `p=reject` (not just `p=none`) on all sending domains

### Monitor What Actually Happens, Not What Dashboards Say

Postmaster Tools reputation scores are a lagging indicator at best and misleading at worst. The only reliable signal is seed-list testing: maintaining Gmail test accounts and checking where your emails actually land. Services like GlockApps, Mail-Tester, and Inbox Monster exist specifically for this. If you're sending more than 10,000 emails per month and not running inbox placement tests, you're flying blind.

Also monitor your DMARC aggregate reports (`rua` reports). They'll tell you if anyone is spoofing your domain, which — if it happens at scale — will poison your reputation even if your own sends are clean.

### Warm Slowly, Complain Loudly

IP warming isn't a one-time event. If your sending volume spikes (product launch, annual sale, security incident notification), Gmail's models may interpret the spike as spam-like behavior. Ramp volume gradually, even for time-sensitive sends. This is operationally painful but mathematically necessary.

And file feedback through every channel Google offers. Postmaster Tools has a feedback form. Google has a bulk sender contact form. The Messaging Anti-Abuse Working Group (M3AAWG) has industry liaisons. None of these are fast, and none guarantee results, but collective pressure from legitimate senders is the only mechanism that has historically moved Google to adjust their models.

### Consider the BIMI Investment

Brand Indicators for Message Identification (BIMI) puts your logo next to your emails in Gmail. It requires a Verified Mark Certificate (VMC), which costs $1,000-1,500/year and requires a registered trademark. That's expensive for a logo in an inbox. But BIMI adoption signals to Gmail's filters that you're a legitimate, invested sender. Several companies in the HN thread reported improved placement after BIMI implementation — though correlation isn't causation in the deliverability world.

Looking Ahead

Font Awesome's experience exposes a fundamental asymmetry in modern email: Google controls inbox access for roughly 30% of all email addresses, and they've optimized entirely for reducing spam at the cost of legitimate mail. There's no SLA, no appeals court, and no competitive pressure (your users aren't switching away from Gmail because your emails land in spam — they're blaming you). Until regulators or market forces change that dynamic, email deliverability isn't a configuration problem you solve once. It's an ongoing ops discipline, like uptime monitoring or security patching, that requires dedicated attention and budget. The Font Awesome post is a useful artifact — not because it reveals new information, but because it proves that doing everything right still isn't enough.