Find the moment, not the persona.

The four outputs of Step 1.

Step 1 isn't a workshop, a deck, or a Notion doc. It's four concrete artifacts that the rest of the system runs on. By the end, you have:

You're not targeting a profile. You're targeting a moment.

Old-school outbound asks "Who is this person?" — first name, job title, company, industry. You drop the tokens, you hit send, your reply is buried with 200 others doing the same thing.

Signal-led outbound asks something different: "What just changed for them?"

Only 5% of your market is actively in a buying cycle at any given time. Signal data is how you find that sliver — a few steps before they enter the cycle, before your competitors even know they exist. — Cannonball GTM, "Signals: The Gateway Drug to Public Data"

That's why we don't buy lists. We build a system that watches for moments — the funding round, the new CAIO hire, the job post that hints at a stack migration, the 10-K disclosure that names a new strategic priority. Each one is a door cracking open. Our job is to be there before anyone else is.

A real ICP, not a firmographic.

Most founders show up with what they think is an ICP. It's actually a filter. There's a difference.

The first one filters the universe down to 2,000 companies. The second one names a specific scenario a specific person is in — and gives you a precise wedge for why they'd open their wallet.

Two paths to a real ICP

Depending on what you already have, we work one of two ways:

You're not done — knowing the person, not just the segment

An ICP segment is half the picture. The other half is the human inside it:

• How do they like to be communicated with? (Channel, format, tone, frequency.)
• What does their typical day actually look like? When is their attention available?
• What language do they use for the problem? Not what your engineers call it — what they call it.

If you don't already know this, we do customer research and real interviews to validate it. No assumptions sneak past this gate.

For each ICP, what's the moment?

Once we know the segment and the person, we work backward to the event or action that triggers the pain scenario they find themselves in. That's the buying window. Examples we'd brainstorm together:

• Hiring signals. A specific role gets posted — Director of AI Security, Head of Detection Engineering, "Agentic AI Engineer." That role exists because a problem just appeared.
• Earnings calls and 10-K language. Public companies disclose strategic priorities a quarter or two before they hire for them.
• Funding events with thematic intent. Not just "they raised" — but "they raised specifically for X."
• Public filings and procurement records. Permits, contracts awarded, RFP pre-stages — public data that reveals decisions before they're shopped.
• Product launches and announcements. A competitor's announcement is often your buyer's moment.

Then I back into how to obtain that information. Some signals come from licensed feeds. Most of the high-leverage ones come from public data nobody is bothering to assemble — government filings, job-board scraping, regulatory disclosures, industry-specific records. That's where the moat is. Most competitors will never do that work. That's the point.

Licensed signal data tells you when to knock. Public data tells you what's going on before the door even opens.

A message so good they'd pay to receive it — even if they never buy.

Your first outbound message isn't a pitch. It's a Permissionless Value Proposition (PVP) — a framework developed by Doug Bell at Cannonball GTM that we use as the messaging foundation for every campaign we run.

In Doug's words, a PVP is "a message so good a prospect would pay to receive it, even if they never bought your product." It is not a teaser. It is not a pitch with personalization tokens. The prospect doesn't have to click, reply, or take a meeting for the value to land. The value is the message.

A striking marker of a true PVP: the sender's brand name often isn't even mentioned. The message is entirely about the prospect's pain and opportunity. — Doug Bell, Cannonball GTM

Why this framework, and not another

Cannonball's PVP framework gets the hardest part of cold outbound right: it forces the message to earn its way into the inbox by delivering real, specific, publicly-sourced insight before asking for anything in return. It anchors the message on a pain so concrete the recipient stops scrolling. And it builds in a clear quality bar — there's a specific set of criteria a message must meet to count as a PVP at all, and a clear concept (an "existential data point") that turns ICP segmentation from generic firmographics into something actually predictive of buying behavior.

The full framework — the seven criteria, the existential data point concept, the pain-based segmentation methodology, plus working examples — lives on Cannonball GTM's Substack. We strongly recommend reading it directly. It's worth the subscription.

How we use it in our managed service

For each ICP segment we identify together, we draft a PVP candidate to Cannonball's standard, then validate it against the market. Reply rates and reply quality tell us whether the segment really is the pain-qualified segment — or whether we need to recalibrate. The framework isn't decorative; it's the gate every outbound message passes through before it leaves the queue.

Multiple ICPs. Tested in parallel.

The reason a fractional GTM advisor running this manually couldn't scale is the math: you'd test one ICP at a time, take 8–12 weeks per cycle, and lose six months proving the obvious.

With the signal monitoring system, we run 2–3 ICP campaigns simultaneously. Each gets its own signals, its own messaging, its own daily queue. By month two, the data tells you which segment is converting at what rate, which messaging variant is landing, and whether the original hypothesis needs to evolve.

That's the compression from 1–2 years to 4–6 months. Not magic. Just parallelism applied where it actually pays off. After six months, tops, it will be obvious whether you have strong PMF — or whether you need to pivot to find it.