Email Subject Lines: Write and Test Them Yourself for Free

Good email subject lines for business email do one job: they tell the right reader, in the first few words, what is inside and why it is worth opening now. You can write them and test them yourself for free. Draft two versions with a simple formula, pair each with preview text, split a small list in half at random in Google Sheets, send one version to each half, and keep the one that earns more clicks or replies. Opens are a weak score, for reasons we explain below.

How to write email subject lines people open

A subject line is a promise. The reader scans the sender name, the subject and the preview text. Then they decide. It takes a second. So start from the email, not the subject: write the one sentence the reader should remember, then compress it. If you cannot say what the email gives them, no wording will save it.

Most email marketing subject lines that work fall into a few patterns. The examples below are our own illustrations for a small bakery and a bookkeeping firm, not tested winners.

1. The plain benefit. Say what they get. Your holiday order form is open.
2. The specific number. Concrete beats vague. 3 receipts your accountant needs before April.
3. The question they already ask. Can you deduct your home office this year?
4. The real deadline. Only when it is true. Pickup slots close Friday at noon.
5. The person, not the brand. Write like one human to another. Quick note from Maria about Saturday.
6. The curiosity with a payoff. Hint, then deliver in the first line of the email. The mistake we see on every first tax return.

Pick two of these for the same email. That pair becomes your test. Keep the topic identical and change only the angle, so the result tells you something about how you said it, not what you sent.

Personal touches help when they are real. A first name in the subject is easy with mail merge, but it is no magic word: readers have seen it a thousand times. A detail that matches the reader's situation (their town, the product they bought, the event they signed up for) usually carries more weight than their name.

How long should email subject lines be?

There is no single right number. What is certain is mechanical: each inbox shows a different number of characters, and a phone shows fewer than a desktop. So put the meaning in the first few words, where every screen shows it, and treat the rest as a bonus that some readers will never see.

The practical check costs nothing. Use your own inbox. Send the email to yourself, then look at it on your phone and on your computer before it goes out. If the cut-off version still makes sense, the length is fine. If the important word falls off the edge, move it forward. This is the simplest email subject lines tester there is, and it tests the inbox your readers actually use.

Write the preview text on purpose

The preview text is the gray line that most inboxes show after the subject. If you leave it empty, the inbox fills it with whatever text comes first in your email, which is often a view in browser link or an address line. Treat it as the second half of the subject. The subject makes the promise; the preview adds the detail. Example, again an illustration: subject Pickup slots close Friday, preview Pick a time for your pie order in under a minute.

What to avoid in email marketing subject lines

The hard line comes from the inboxes themselves. Google's sender guidelines say the subject, headers and display names should accurately represent the sender and the content, and they give a clear example: no subject starting with Re: or Fwd: unless the message really is a reply or a forward ^[1]. The same page says not to put subject or message content into the display name. Fake familiarity buys one open and then costs trust.

The rest is judgment, and it is about respect for the reader:

• No bait. If the email does not deliver what the subject promises, the next one gets ignored.
• No shouting. All caps and a row of exclamation marks read as a flyer, not a message.
• No false urgency. A deadline that keeps moving teaches people to ignore deadlines.
• No mystery without a payoff. A curious subject must be answered in the first lines.
• No changing two things at once when you test. Otherwise you will not know which change worked.

Run your own subject line test in Google Sheets

You do not need a platform to run a basic A/B test. You need a list of people who agreed to hear from you, a spreadsheet and a way to tell the two versions apart. The split must be random, because sorting by signup date or by name can put your most loyal readers in one half.

1. Put your list in Google Sheets. One row per subscriber, the email address in column A.
2. Add a random number to every row. In the first empty column, type =RAND() and fill it down. Google's help page says RAND returns a random number between 0 inclusive and 1 exclusive ^[2].
3. Freeze the numbers. RAND recalculates when the sheet changes ^[3], so copy the column and paste it back as values only. Your split now stays put.
4. Sort by that column and split the list in the middle. Label the top half A and the bottom half B in a new column.
5. Write two versions that differ only in the subject line (and its preview text, if you are testing the pair). Same body, same send time.
6. Tag the link. Give the main link in version A a marker such as ?v=a and in version B ?v=b, so your site analytics can tell the visits apart. If you want replies instead, ask a one-line question both versions share.
7. Send both halves at the same time, from the tool you already use, and wait a few days before you count.
8. Record the result in the sheet: sends, clicks (or replies) and the rate for each half.

Reading the result: clicks and replies, not opens

Why not compare open rates? An open is counted when the email loads a tiny remote image, and Apple Mail can load that content for the reader. Apple's own help page says that with Protect Mail Activity turned on, the reader's IP address is hidden and remote content is downloaded privately in the background when the message arrives, instead of when it is viewed ^[4]. So an email can register as opened when nobody read it. Clicks and replies need a real person. Our guide to what a good open rate looks like goes deeper into why opens became unreliable.

Here is a worked illustration with round, invented numbers. A shop emails a list of 400 people, 200 per half. Version A gets 12 clicks, a 6% click rate. Version B gets 18 clicks, a 9% click rate. B looks better. But the gap is six people. On a list this small, chance alone can move a handful of clicks, so one test is a hint, not a law.

The honest way to use small tests is to repeat them. If the same kind of subject (say, the specific number) wins three sends in a row, adopt it as your default. If the winner flips each time, the two angles are about equal for your readers, and your time is better spent on the email itself.

When manual splits stop being practical

The spreadsheet method works for one test a month on a list of a few hundred people. It starts to break in three specific places.

First, the list outgrows a mailbox. Sending to thousands from a personal account runs into the sending limits your mailbox provider sets, and it gives you no unsubscribe or bounce handling.

Second, you want the winner to go to everyone. The real payoff of a test is sending a small share both versions, then the better one to the rest of the list. By hand, that means waiting, counting and sending again. Third, you want different subjects for different readers, such as new subscribers versus past buyers. Keeping those segments current in a sheet becomes a weekly chore, and that is the point where segment sends belong in a platform, as our guide to email segmentation tools explains.

Three tools do this testing step for you. Kit pairs its subject-line generator with built-in tests. Brevo adds AI subject lines on Starter, and its A/B tests and send-time optimization sit on Standard. Mailchimp tests from Essentials and adds Send Time Optimization on Standard and Premium.

The bottom line

Writing better email subject lines is a skill you can practice for free: say the benefit in the first few words, write preview text on purpose, never fake a reply or a deadline, and judge every test by clicks or replies rather than opens. A random split in Google Sheets is enough to learn what your readers respond to. Once you test every send, want the winner delivered to the rest automatically, or write different subjects for different groups, a platform's A/B testing and segment sends take over the counting.