Signal based outbound

Signal-based outbound

Cold outbound to a static list mostly fails because there is no reason to talk now. This kit finds the moments when a prospect is likely to care and decides, ahead of time, exactly what happens when each one fires.

When to use

• Designing outbound for a new product, segment, or team.
• The user has signal sources (job-change alerts, website visitor identification, product usage, news alerts) and no plan for them.
• Reviewing which signals actually led to pipeline last quarter.
• Not for writing the emails themselves (hand off to sequence-writer) or sizing the market (tam-builder).

Inputs

Ask in one round, then draft:

1. Product and ICP: what the product does and for whom. Reuse data/icp.json from tam-builder if present.
2. Best customers: 5 to 10 recent wins. What was happening at each company when they bought? This is the main source of real signals.
3. Data the user can see: CRM, product analytics, website analytics, job boards, news, enrichment tools, communities. Only design signals the user can actually detect.
4. Team capacity: how many signal-driven touches per rep per week is realistic.

Steps

1. List candidate signals from the won deals first, then common categories: hiring patterns, leadership changes, funding and M&A, expansion, tech adopted or dropped, product usage (free or existing customers), website intent, champion job changes, regulation or deadlines, public pain (reviews, outages, posts).
2. Score each: strength (strong, medium, weak) based on evidence from won deals, points, and a decay window in days. A signal with no supporting win is "to test", not strong.
3. Define the evidence precisely enough that a person or tool would flag the same accounts ("3+ open sales roles in a new region", not "growing").
4. Write a play for each strong or medium signal: who (roles), angle (why this signal makes the offer relevant), channels, speed (how fast after the signal), sequence name, and success measure.
5. Set routing: the score at which an account becomes active, how signals stack, and a cap of active plays per account so prospects are not hit by three plays at once.
6. Plan the test: for each "to test" signal, a 4 week test with a small account set and the result that would make it strong.
7. Write out/signals.json (shape: the kit's examples/signals.json) and out/signals-playbook.md (format below). Show the playbook.

Output format

# Signal-based outbound playbook: <product>

Act when an account scores 40+ within the decay window. One active play per account.

| Signal | Strength | Points | Decay | Play | Speed |
| --- | --- | ---: | ---: | --- | --- |
| Past champion joined a new company | Strong | 50 | 120 days | Champion reunion | 14 days |
| Opening a new sales region | Strong | 40 | 60 days | New-team routing | 5 business days |
| Repeat pricing page visits | Medium | 20 | 14 days | Intent check-in | 2 business days |

## Play: New-team routing
- Who: VP Revenue Operations, CRO
- Angle: a second team splits the forecast and breaks routing within a quarter
- Channels: email, LinkedIn. Sequence: routing-for-new-teams (write with sequence-writer)
- Success: meeting booked or referral to the owner

## To test
- Compliance deadline in the next 6 months: 4 week test on 50 accounts; strong if 3+ meetings.

Guardrails

• Vendor-neutral. Describe where to find each signal by type of source; name a tool only if the user already uses it.
• Do not design plays that reveal tracking in a creepy way ("I saw you on our pricing page"). Use the signal to choose timing and angle, not as the message.
• Respect privacy rules for personal data and the terms of each data source. Flag signals that rely on personal data (job changes, website visitor identity) so the user checks consent and opt-out handling.
• Do not contact anyone or set up automations. The output is a plan and a spec.

Signal-based outbound

Works out which buying signals matter for your product and turns each one into a ready play: who to contact, what angle to use, which channel, and how fast. You get a readable playbook and a signals.json spec you can wire into any tool.

Who it is for

GTM engineers, SDR leaders, and founders replacing "spray the list" outbound with outreach that has a reason to happen now. Vendor-neutral.

Example

Input: "We sell forecasting and lead routing to RevOps teams at 200 to 2,000 person companies. Our last 8 wins: 5 had just opened a second sales team, 2 had a new RevOps leader, 1 came from a former champion."

Output (out/signals-playbook.md):

Plus one section per play and a test plan for unproven signals. See examples/signals.json for the spec format.

What it can access

• Reads data/ (plus the kit's own examples/), writes out/. No scripts, no network.

Setup

None. Pairs well with tam-builder (who to watch) and sequence-writer (what to send).