Outreach personalization

Overview

Personalization isn't "insert their first name." It's making the recipient feel the message could only have been written to them. The currency is relevance, not how much you know about their weekend.

The test: could this exact message be sent to 500 other people unchanged? If yes, it isn't personalized - it's a template with a name slot.

Relevance vs. personal (stay on the right line)

• Relevant = tied to their role, company, industry, stack, or a trigger - and to the value you offer. ("Saw you're hiring 2 SREs - usually means a SOC 2 crunch.")
• Personal-but-irrelevant = creepy. ("Loved your vacation photos!") It signals surveillance, not understanding. Avoid it.

Tiers - match effort to account value

Don't deep-research everyone; spend where it pays (see ideal-customer-profile):

All three are legitimate - the mistake is spending A-effort on C-accounts or sending C-copy to A-accounts.

What to pull, and from where

• Company: recent funding, launches, hiring, news, tech stack, positioning on their own site (their words > your paraphrase).
• Person: role and tenure, recent LinkedIn posts/comments, past companies, talks/podcasts, mutual connections.
• Trigger: the single most recent relevant change - that's your opener.

Turn the best one into one specific line. More isn't better; one sharp, true observation beats three generic ones.

A snippet system (scales personalization)

Build reusable, swappable building blocks so volume stays relevant:

• {{trigger_line}} - the one researched observation (A/B tiers).
• {{pain_by_role}} - a pain mapped to their title/segment.
• {{proof_by_industry}} - a peer/result in their vertical.

Compose email = trigger_line → pain_by_role → proof_by_industry → one ask. The blocks are reusable; the trigger line is what makes each unique.

Guardrails

• Never ship a misfired token (Hi {{FirstName}}, / "at {{Company}}"). One broken merge tells the reader they're #437 on a list.
• Verify the fact before you reference it - wrong details are worse than none.
• Real LLM-generated, per-recipient copy - never a hardcoded template blasted to a list. If a batch comes out with identical bodies (minus a name), that's a template, not personalization; fix it.

Do / Don't

• Do lead with one true, relevant observation about them.
• Do tier your effort to account value.
• Don't reference personal-but-irrelevant details.
• Don't trust a merge field you haven't validated.

Common mistakes

• "Personalization" that's just a first-name token.
• Research lines that are true but irrelevant to your value.
• Over-researching low-tier accounts and under-researching whales.
• Identical bodies across a "personalized" batch.

Go live with Chronos (optional)

• geo_research / geo_enrich_contact - gather the real facts per person/company.
• geo_set_personality - lock your brand voice so every draft sounds like you.
• geo_add_personalization_rule - encode rules the copy must follow.
• geo_add_document / geo_add_database_item - give Chronos proof points, case studies, and context to draw from.
• geo_draft_message - generate per-recipient copy (real LLM output, never a template).

Connect in ~30s → CONNECT-CHRONOS.md. Free, no card.

See also: cold-email-writing, linkedin-outreach, account-research.