Subject line advice has a measurement problem that almost nobody discloses. Since September 2021, Apple Mail Privacy Protection has loaded tracking pixels automatically for protected users, which means the open rate you are optimizing toward counts messages no human read. Most published subject line "data" is built on that metric. Here is what still holds up, and how to test without fooling yourself.
Why open rate stopped being a scoreboard
Apple’s documentation states plainly what the feature does: Mail Privacy Protection prevents senders from seeing whether you opened a message, hides your IP address, and downloads remote content privately in the background when the message arrives rather than when you view it. Open tracking works by loading a remote image, so for every protected recipient the pixel fires on delivery.
The consequence is not that open rate is zero information. It is that the number is diluted by an unknown, client-dependent proportion of automatic loads, and that proportion differs between segments and drifts over time. You can still compare two variants sent to the same list on the same day, since the dilution is roughly equal on both sides. You cannot compare to a benchmark, to last year, or between segments with different device mixes.
Three fields decide whether the email opens
The subject line is one of three things visible in the inbox list view. The sender name answers "do I know who this is," and it is usually the strongest of the three: an unrecognized sender gets deleted regardless of what the subject says. The subject line answers "is this relevant." The preview text, pulled from the first text in the body unless you set it explicitly, either extends the promise or wastes the space on "View this email in your browser."
Set the preview text deliberately on every send. It is the most commonly wasted real estate in email, it takes ten seconds, and it should read as the second half of a sentence the subject line started, not a repetition of it.
Principles that survive the measurement problem
- Specific beats clever. "Three ways to cut your shipping costs" tells the reader what they get; "You won’t believe this" tells them nothing and costs them trust when the email underdelivers.
- Front-load the meaning. Mobile clients truncate, and the truncation point varies by client and orientation, so put the load-bearing words first rather than optimizing for a character count you cannot control.
- Match the body. A subject line that outperforms and then disappoints trains the recipient to stop opening, which is a slower and more expensive failure than a weak subject line.
- Use urgency only when it is real. A countdown that resets is noticed, and under CAN-SPAM a subject line that misrepresents the contents of the message is a legal problem, not a stylistic one.
- Personalize on something you know. The signup source or the last product viewed carries more weight than a first name, which every sender uses and every recipient discounts.
Formulas worth adapting
| Pattern | Example shape | Best used for |
|---|---|---|
| Outcome without the pain | How to [outcome] without [obstacle] | Educational content, top of funnel |
| Enumerated | [Number] [things] that [benefit] | Newsletters and roundups |
| Direct question | Still paying too much for [thing]? | Objection-led sequences |
| Plain offer | 24 hours: 20% off everything | Promotions, where clarity beats craft |
| Curiosity gap | The mistake costing you customers | Use sparingly. It works until the body fails to deliver, then it stops working permanently |
| Reply prompt | Quick question about your order | Welcome and post-purchase, where a reply is a strong engagement signal |
Spam triggers: the real ones and the folklore
The idea that specific banned words send you to spam is mostly folklore. Modern filtering weighs authentication, sending history, complaint rate, and engagement far more heavily than vocabulary. Since February 2024 Google and Yahoo have required bulk senders to authenticate with SPF, DKIM, and DMARC, to support one-click unsubscribe, and to keep spam complaints below 0.3%. Failing any of those does more damage than every exclamation mark you have ever typed.
That said, ALL CAPS, stacked punctuation, and hyperbolic offers correlate with recipients marking mail as spam, and complaint rate is exactly what the 0.3% cap measures. So the advice survives, with a different explanation: avoid them because they make people report you, not because a filter has a word list.
Key takeaways
- Apple Mail Privacy Protection loads tracking pixels on delivery, so open rate is only comparable between two variants sent to the same list at the same time.
- Declare A/B test winners on clicks. A subject that wins opens and loses clicks was over-promising.
- Sender name, subject line, and preview text work as one unit. Set preview text explicitly on every send.
- Front-load meaning rather than chasing a character count, since truncation varies by client.
- Spam word lists are folklore. Authentication, complaint rate under 0.3%, and one-click unsubscribe are what mailbox providers actually enforce.
Related reading
Sources
- Use Mail Privacy Protection on iPhone, Apple Support
- Email sender guidelines (bulk sender requirements), Google Workspace Admin Help (2024)
- CAN-SPAM Act: A Compliance Guide for Business, US Federal Trade Commission
- Email marketing benchmarks, Omnisend (2025)
- Copywriting
- Email Marketing
- Conversion
Written by
Valter Brandt
Chief Marketing Officer
Valter Brandt is the Chief Marketing Officer of ThisCom, working with clients across the United States and internationally. He has led marketing strategy through the major shifts in social advertising, mobile, content marketing, programmatic media, and marketing automation.
All articles by Valter Brandt →
