GTM Skills
⏱ 0 uses★ 0 stars
GTM OS — Signal Library
Builds and maintains signal-library.md — buying signals, outreach hooks, and a performance log. Turns pattern-matching guesses into tested, improving triggers tied to your specific ICP.
The prompt
# GTM OS — Signal Library
This skill owns signal-library.md — the buying signal and hook file. Depends on icp-definition.md existing first: signals without a defined buyer are noise. This file is what separates reactive selling from proactive pipeline generation.
## What signal-library.md captures
### 1. Signal inventory
A categorized list of signals that indicate buying intent or readiness for your specific product. For each signal:
- Signal name (brief, memorable)
- Category: Trigger event / Behavioral / Technographic / Organizational
- What it indicates (why this signal matters for your ICP)
- How to detect it (source: LinkedIn, job postings, news, G2, etc.)
- Confidence level: High / Medium / Low (how reliably this signal leads to a real conversation)
### 2. Outreach hooks (per signal)
For each high-confidence signal, one outreach hook:
- The opening line that references the signal without being creepy
- The bridge from the signal to your value (why this signal makes now a good time)
- The ask (specific, low-friction, relevant to where they are)
Good hooks are specific and timely. Bad hooks are "I noticed you might be interested in..." Push for hooks that would make the rep say "I'd reply to that."
### 3. Signal-to-ICP mapping
Each signal should be mapped to at least one ICP segment. If a signal applies to everyone, it applies to no one.
- Signal → Segment(s) it fires for
- Why it fires for that segment (the logical connection)
- Segments it does NOT apply to (anti-mapping)
### 4. Performance log
The asset that makes this file compound over time. After each campaign or outreach sequence using a signal:
- Signal used
- Volume (how many prospects were targeted)
- Reply rate
- Meeting rate
- Key observations (what worked, what flopped, what surprised you)
A signal library without a performance log is just a list of guesses. The log is what turns it into knowledge.
## How to build or update this file
When building from scratch, start with what's already working:
"Tell me about the last three outbound meetings you booked. What triggered you to reach out to that specific person at that specific time?"
Then expand:
"What events or changes at a company make you think 'we should reach out right now'?"
For hooks:
"Write me the opening line you'd use if a prospect just posted that they hired a new VP of Sales. Now write the one you'd send if they just announced a Series B."
## Anti-patterns to flag
- Signals without detection sources — if you can't reliably find it, it's not a signal, it's a hope.
- No performance log — the most common failure mode. Flag it every session if it hasn't been updated.
- Hooks that don't reference the signal — if the hook could be sent to anyone, it's not a hook.
- Signal-to-ICP mapping that says "all segments" — force at least one exclusion.
- Signals that are just "they visited our website" — push for external, detectable signals that don't require the prospect to already know you.
Suggest an improvement
Propose a change. The owner will be notified and can accept or decline. If accepted, you'll be credited as a contributor.
Log in to suggest improvements.