Plain text vs HTML cold email is one of the few outbound debates with a clean verdict: cold email should be plain text, or minimally formatted HTML that renders as plain text. Designed templates, images, buttons, and heavy markup cost you inbox placement before a human ever sees the message, and the data behind that claim is unusually one-sided. This article covers what the numbers actually show, the Gmail Promotions tab mechanic that does most of the damage, what the format choice means for signatures, links, and tracking, and the narrow cases where HTML genuinely earns its place.
The verdict is not theoretical for us. Our parent agency, Referral Program Pros, has run more than 4,000 outbound campaigns and booked over 7,000 meetings, and GTM Bud is built on that agency’s playbook. Every one of those email campaigns went out as plain text, because the guarantee we put in writing, 5 percent positive replies on LinkedIn or 1.5 percent on email or a full refund, is only survivable when messages reach the primary inbox. Format is the cheapest deliverability lever there is: it costs nothing to pull and it is entirely under your control.
Why is cold email plain text?
Cold email is plain text because every mailbox provider is trying to answer one question about your message: is this a person writing to a person, or a campaign writing to a list? Plain text is what person-to-person email looks like. HTML templates, images, buttons, and custom styling are what campaigns look like, and filters have had two decades of training data to learn the difference. The cost of looking like a campaign is paid twice: spam filters score HTML weight and remote images against you at delivery time, and Gmail’s tab classifier routes campaign-shaped mail into Promotions, where B2B prospects rarely look. A cold prospect gave you no consent and no engagement history to offset those signals, so the message must earn primary-inbox treatment on its shape alone. The fix costs nothing: write short plain text that would look normal sent from one colleague to another.
There is a second, quieter reason: rendering. A designed template renders differently in Gmail, Outlook, and Apple Mail, and Outlook’s HTML engine in particular is notorious for breaking layouts. Plain text renders identically everywhere, on every client and every screen width. The message you wrote is the message every prospect sees.
Almost nothing you send is actually plain text
Here is the technical subtlety most format debates skip: nearly all “plain text” cold email is HTML. When you hit send in Gmail, Outlook, or any sending tool, the message goes out as multipart MIME, specifically a multipart/alternative container holding two versions of the same content: a text/plain part and a text/html part. The receiving client picks one to display, and under the multipart rules defined in RFC 2046, parts are ordered by increasing preference, so the HTML part listed last is what almost every modern client renders. The “plain text email” your prospect reads is usually a thin HTML document: your sentences wrapped in a couple of <div> or <p> tags, default font, no styling. Send yourself a test and view the raw source, and you will see both parts sitting side by side in the same payload.
This matters because it reframes the real question. The choice is not between the text/plain format and HTML as technologies. It is between minimal HTML that renders as one-to-one correspondence and heavy HTML that renders as a marketing asset. Filters make the same distinction: they do not punish the existence of an HTML part, which describes practically all email on earth, they score what the HTML contains. Tables of nested layout markup, remote images, tracked links, and web fonts push a message toward the campaign class. A few paragraph tags do not.
So when this article says plain text, read it precisely: a message whose HTML part contains text and almost nothing else. That is the standard output of a well-configured sending tool, and it is the shape every claim below is measuring.
What does the data show about plain text vs HTML?
The strongest published number comes from Hunter.io, which analyzed 2.2 million emails sent by its users and reported that HTML-heavy cold emails received 674 percent more bounces than plain text ones. Hunter’s write-up attributes the gap to filtering: HTML elements like graphic signatures, images, formatting, and multiple links make a message look like a bulk blast, which triggers stricter treatment up to and including rejection at the door. A bounce multiple that large is not a rounding artifact. It is the difference between a list you can keep mailing and a domain you have to retire.
You will also find vendor blogs circulating specific reply-rate multipliers for plain text, including claims of double or triple the replies of designed templates. We could not trace any of those figures to a published dataset with methodology, so we will not lean on them. The honest statement is narrower: no large public study we could find isolates format as a reply-rate variable, while the bounce and placement evidence all points the same direction, and reply rate is downstream of placement. An email in spam or Promotions has a reply rate of approximately zero regardless of how well it was written.
One genuine counterpoint deserves attribution rather than silence. Migma, an AI email design tool, published a case study reporting that professionally built HTML beat plain text on inbox placement by roughly 20 percentage points. Read the context before updating your beliefs: that test measured designed marketing email of the kind sent to opted-in audiences, where engagement history and expectations are completely different. For cold outreach to strangers, the disagreement dissolves, and even HTML-friendly sources concede plain text is the safe choice there. Where sources conflict, the conflict itself is informative: format norms are context-dependent, and the cold context is the harshest one.
How does the Gmail Promotions tab classify cold email?
Gmail sorts incoming mail into five categories automatically, and its own documentation defines Promotions as deals, offers, and other promotional emails. Google has described the mechanism in a Workspace blog post: machine learning models classify each message using a variety of signals, including who sent it, what kind of content it contains, and how users have interacted with similar mail. Google does not publish the model’s features, but the content signal is the one you control on every send, and deliverability practitioners consistently observe the same triggers: image-heavy layout, CTA buttons, marketing language, tracked links, and bulk-pattern sending all push mail toward Promotions, while short, plain, personally addressed text stays in Primary. For a B2B cold email, Promotions is a quiet failure mode: the message was technically delivered, no spam alarm fired, and the prospect never saw it.
Two properties of this classifier are worth internalizing. First, it is personalized: the same message can land in Primary for one recipient and Promotions for another, depending on how each has treated your domain before, which is why single seed tests mislead. Second, it sits on top of spam filtering, not instead of it, so a designed template takes two rolls of the dice against you: one at the spam gate, one at the tab sort. Our cold email deliverability guide covers the full stack beneath both, from authentication to complaint rates. Format is the layer above that stack, and it is the layer where HTML quietly loses campaigns that did everything else right.
Plain text vs HTML cold email, side by side
| Dimension | Plain text (minimal HTML) | Designed HTML |
|---|---|---|
| Deliverability | The low-bounce baseline in Hunter.io’s 2.2 million email analysis | 674 percent more bounces than plain text in that same Hunter.io dataset |
| Gmail tab placement | Reads as correspondence; defaults toward Primary | Campaign signals push toward Promotions, per the classifier behavior described above |
| Rendering | Identical in every client and screen width | Varies by client; Outlook in particular breaks layouts |
| Tracking | None by default, which is the correct cold email setting | Usually ships with open pixels and rewritten links, both placement risks |
| Reply behavior | Reads as a person; replying feels like answering a colleague | Reads as a campaign; the expected action is a click or a delete, not a reply |
| Weight and load time | A few kilobytes of text | Heavy payload; remote assets load slowly or get blocked |
| Effort per send | Writing only | Design, testing across clients, and maintenance |
The reply behavior row is the one that surprises people. Formatting sets the social contract of a message: a designed email announces itself as broadcast, and nobody replies to a broadcast. A plain email asks a question the way a person does, and the natural response to a person is an answer.
What the format choice means for signatures, images, links, and tracking
Choosing plain text is not one decision. It cascades into four smaller ones, and cold senders lose the benefit by getting the small ones wrong.
- Signatures. Text only: name, title, company, and at most one plain link. Logo images, social icon rows, and promotional banners convert your carefully minimal email back into an HTML asset. Hunter.io’s bounce data specifically calls out graphic signatures as one of the elements associated with the 674 percent gap.
- Images. None. Every image is a remote asset that adds weight, triggers image filtering, and marks the message as designed. If a visual genuinely sells the offer, send a link to it once the prospect replies, or skip email and put the visual where it belongs, in the deck.
- Links. One at most, and often zero in the first touch. Multiple links, mismatched link text, and shorteners are classic filter triggers. With only 50 to 125 words to spend, per the benchmarks in our guide to how long a cold email should be, a first touch usually works better closing with a question than with a URL.
- Tracking. Off. Open tracking injects a remote pixel, the exact HTML element this whole article argues against, and it buys you a metric that Apple’s proxies corrupted years ago. Our verdict piece on open tracking in cold email covers why the pixel costs placement and, for European recipients, now raises a consent problem too. Click tracking rewrites your one link through a shared tracking domain, which imports other senders’ reputations into your campaign. Measure replies instead; they are the metric the campaign exists for anyway.
Strip all of that away and what remains is the text. That is the honest trade of the plain text verdict: you give up every crutch, so the words carry everything, and personalization becomes the whole game. This is where an AI cold email writer grounded in real prospect research earns its place, generating the specific first line that makes 80 plain words land, rather than decorating generic words with design.
When does HTML formatting actually make sense?
The exceptions are real but narrow, and each one works because some ingredient of the cold context is missing.
After the reply. Once a prospect answers, you are in a conversation with engagement history on both sides. Sending the deck, a chart, or a calendar link with light formatting is normal correspondence at that point. The plain text discipline applies to earning the reply, not to everything that follows it.
Opt-in audiences. Newsletters and product marketing go to people who asked, open regularly, and expect design. That is the context where Migma’s HTML-favorable placement data was measured, and it is fair to let it govern there. None of it transfers to strangers.
When light structure serves the reader. A single bolded phrase or a three-item bullet list inside an otherwise plain email is still minimal HTML and still renders as correspondence. Use it when it makes 100 words scan faster, not to make the email prettier.
When the visual is the offer. A design studio pitching redesigns can justify one linked example. Even then, the winning shape in our agency’s campaigns has been a plain email that describes the result and links once, not an embedded portfolio.
If a case does not fit one of those four, it is not an exception. “Our brand guidelines require the template” is the most expensive sentence in cold email.
Frequently asked questions about plain text vs HTML cold email
Can you use bold text or bullet points in a cold email?
Sparingly, yes. A single bolded phrase or a short bullet list is simple HTML that filters treat as normal correspondence, because real one-to-one emails contain exactly that kind of light formatting. The risk curve bends at decoration: colors, buttons, banners, custom fonts, and multi-column layouts are the signals that reclassify a message as a campaign. If a formatting choice would look strange in an email to a colleague, leave it out of a cold email too.
Should your cold email signature include a logo or banner?
No. A logo is a remote image and a banner usually adds another link, and both are inputs that spam and tab classifiers score against you. Hunter.io lists graphic signatures among the HTML elements associated with dramatically higher bounce rates in its 2.2 million email analysis. Use a text signature: name, title, company, and at most one plain link. It reads as a person and gives filters nothing to flag.
Is a pure plain text email better for deliverability than simple HTML?
Marginally at best, and the difference rarely matters in practice. Nearly every sending tool emits multipart MIME with a text/plain part and a simple HTML part, and mailbox providers are thoroughly used to that shape. What filters punish is not the presence of an HTML part but what it contains: images, heavy markup, link density, and tracking. A minimal HTML email that renders as plain text sits in the same risk class as pure plain text for cold outreach.
Should agencies use branded HTML templates for client cold email campaigns?
No, even when the client asks for brand polish. Cold prospects have no relationship with the brand, so the template signals campaign rather than correspondence and drags placement down across every mailbox the agency sends from. The deliverable that protects a retainer is replies, and replies come from plain text that reads like the client founder wrote it personally, a discipline our cold email for agencies platform enforces by default. Save the branded design for the deck you send after the prospect answers.
Do HTML emails still work for newsletters and marketing campaigns?
Yes, because the context differs in every way that matters. Newsletter recipients opted in, engage with the sender regularly, and expect visual formatting, so designed HTML performs well there, which is exactly what Migma’s placement case study measured. The plain text verdict applies to cold outreach, where no consent or engagement history exists to offset campaign signals. Different audience, different classifier outcome, different right answer.
Win the inbox before you write a word
The plain text vs HTML question is a trade-off with a lopsided answer for cold email: HTML buys you brand polish a stranger never asked for, and it pays for that polish with bounces, spam placement, and Promotions tab burial, per Hunter.io’s data and Gmail’s own classification design. Plain text, or the minimal HTML that renders as it, costs you nothing but design vanity and hands the entire budget to the only element that earns replies, the words.
GTM Bud makes that discipline the default rather than a checklist. Our cold email automation tool sends every campaign as plain text with minimal formatting, no pixels and no templates, with deliverability watched at the infrastructure level and reporting built on replies and meetings. Pricing is a flat monthly rate per connected sending account, and every campaign carries the same written guarantee our agency work does: 1.5 percent positive replies on email, or a full refund. Send mail that looks like a person wrote it, because a person did.