The personalisation tokens that actually move reply rates

{{first_name}} and {{company_name}} are not personalisation. They’re mail merge. Everyone receiving your sequence knows their own name. Seeing it in a subject line signals automation, not attention.

The tokens that move reply rates are the ones a recipient can tell required actual knowledge of their situation.

Tech stack tokens. Knowing that a prospect uses Salesforce + Outreach + Gong lets you frame your message around the specific integration context they’re already in. “You’re running Outreach on top of Salesforce. Here’s where that creates routing friction” is not a generic claim. It’s specific to their setup.

Hiring velocity tokens. “You’ve added 8 SDRs in the last 60 days” is a number they know is correct and tells them you’re paying attention to their growth, not sending a blast. Pair it with something relevant to that growth signal.

Recent company event tokens. Funding round, product launch, new market entry, executive hire. The specificity creates credibility. The timing creates relevance. Together they create a natural reason for outreach that doesn’t feel manufactured.

Mutual connection or shared context tokens. If you know which conference they spoke at, which community they’re active in, or which tool they’ve publicly endorsed, that’s a connection point that signals you exist in the same world.

The infrastructure to make this work: Clay pulls and formats these data points at the account and contact level. A column per token, clean output, no truncation issues. The sequence template has a slot for each one. Claude API handles the cases where the raw data needs to be shaped into a sentence rather than dropped in verbatim.

The trap to avoid: over-templating. If every email in your sequence uses all six tokens in a visible formula, it reads as a checklist. Pick the two or three most relevant for each account and let the rest inform the message angle without showing up explicitly.