lemlist Personalization Variables With Worked Examples

Sort your CSV columns into three tiers before you write any copy:

• Identity variables: first name, company name and role. They're table stakes, and every one needs a fallback.
• Context variables: industry, city, the tool they use, a team-size band. They let one email fit a whole segment.
• Reason-to-write variables: a custom column such as a recent hire, a product launch, or one sentence you wrote after reading their site. These are the ones that change replies.

The third tier is expensive because a person or an enrichment step has to fill it. Don't spread it thin. Save hand-written or AI-drafted reason-to-write lines for your best-fit accounts and let context variables carry the rest.

A wrong first line is worse than no first line. If the variable says "congrats on the funding round" and there was no round, the sequence is finished with that reader.

Scraped and enriched lists always have gaps, so assume every variable will arrive empty for some rows. Work through these steps for each one:

1. List every variable the email uses, including any inside the subject line.
2. Pick a plain word that reads fine when the field is missing: "there" for a first name, "your team" for a company name, "your region" for a city.
3. Add the fallback using lemlist's own syntax. The exact format depends on your workspace, so check lemlist's help docs instead of copying syntax from a blog post, this one included.
4. Read the sentence with the fallback in it. "Saw that your team is expanding in your region" is vague enough to look automated, so where a fallback leaves a sentence hollow, move it into a conditional block or drop the line for those contacts.
5. Add a test row to your CSV with every custom column blank and send yourself the preview.

Step 4 is the one people skip. A fallback stops the visible error, but it doesn't make a weak sentence good.

Conditional logic lets one sequence say different things to different readers. A typical branch checks the job title and swaps one sentence, not the whole email.

For example, if the title contains "Marketing", the opener mentions campaign reporting. If it contains "Sales", it mentions pipeline coverage. The else branch uses a neutral line about how the team runs outbound. Make the else branch as good as the others, because titles in real lists are messy ("Head of Growth", "Founder", blank) and plenty of contacts will land in it.

Three rules keep branching manageable:

• Branch on one attribute at a time. Nested conditions multiply the cases you have to test.
• Change one sentence per branch, so the rest of the email stays consistent and reply rates stay comparable.
• Preview one contact per branch, plus one that matches none of them.

Dynamic images, such as the prospect's homepage with your note overlaid, stand out in a crowded inbox. They also cost you in three ways: setup time, a heavier email, and a higher chance of reading like marketing instead of a person writing.

Use them selectively. They fit a short list of high-value accounts where the image shows something real you noticed, like a specific page or a gap in a checkout flow. They fit poorly on a big list where the image is the same idea with a logo swapped in.

Keep image files small and hosted on a reliable URL. Send a test batch to inboxes at more than one mailbox provider and confirm the email lands in the primary view before you scale. If the sending domain is new, sort out warm-up and authentication first, since images won't rescue a poor reputation. For how tools compare on inbox placement, see the cold email tool comparison.

Personalization only pays for its setup time if it survives contact with a real list. Run this pass before you send:

1. Sort the CSV and scan 20 random rows for typos in names, all-caps company names and stray spaces.
2. Preview one row from each branch, plus one row with blank custom fields.
3. Check the subject line separately, because variable errors there are the most visible.
4. Send to a few test inboxes and read the result on a phone.
5. Start with a small first batch and widen it only if replies are relevant and bounces stay low.

Personalization earns the reply, not the deal. The average B2B new-logo win rate is 19 percent¹, so the meeting you book still needs a qualified buyer behind it. Pair your variables with a tight target list, and read how personalization differs from spam before you automate first lines.