Signal-based outbound for enterprise comms teams

If you are selling to corporate communications teams, your whole market might fit in a spreadsheet. A few hundred companies have an in-house comms team big enough to buy, each with two or three people who matter: the chief communications officer and a VP or two. Burn them with a generic sequence and there is no next list.

Signal-based outbound is the alternative. You write only when something happened that changes the buyer’s work, and the message opens with that event. This guide covers which signals matter, how to score them, what to write, and how to pace a market you can’t exhaust.

Why comms buyers need a different playbook

The market is small and doesn’t refill. A contact who marks you as spam, or a sending domain that lands in junk, stays that way for the accounts you care about most.

The buyers read for a living. Comms leaders judge messaging all day. A template with their company name merged into line one tells them how you’d treat their own stakeholders.

Their work is event-driven. Earnings, leadership changes, deals, rebrands and crises set the comms calendar. Trigger-based outreach that follows the same events arrives when the topic is already on their desk.

The signals that matter, and why

Each of these changes what the comms team has to say or who says it.

The first two are exec change signals and usually the strongest. A new comms leader is the buyer. A new CEO or CFO is why the buyer’s year just changed.

SEC 8-K signals, in plain terms

Public companies file a Form 8-K to report material events, generally within four business days. Three items cover most of what comms buyers care about:

• Item 2.01, completion of acquisition or disposition of assets. The deal has closed.
• Item 2.02, results of operations and financial condition. Usually filed alongside the earnings release.
• Item 5.02, departure, election or appointment of directors or certain officers. It also covers board seats and officer compensation, so confirm the role before you treat it as a new CEO or CFO.

Filings are free, structured and dated. Their limit is coverage: a VP of communications is rarely one of the officers item 5.02 covers, so comms-leader hires mostly show up in news and newsrooms.

Crisis is volume, not a headline

One negative story is noise. Count distinct outlets running crisis terms in their headlines over 48 hours, and set a minimum before the signal fires. Then let a person decide whether to act at all. Writing to a CCO mid-crisis can read as opportunism.

Detection: rules with the company as subject

Use rules, not a model, so every signal says why it fired:

1. The company must be the subject. “Halvard names…” or “…joins Halvard as…” counts. A passing mention does not.
2. Skip market commentary. “3 reasons to buy…”, “COMMENTARY:” and “X vs. Y” are about the stock.
3. A hire only counts when a persona title appears. A new head of procurement is not a comms signal.

Rules have a known failure mode: name collisions. A headline about “Halvard Robotics”, a different firm, matches a rule written for Halvard Industries. The fix is a cheap second check on each candidate (“is this headline about Halvard Industries?”), plus a human glance before anything sends.

One event, one signal

An earnings release shows up as an 8-K item 2.02, a newsroom press release and a dozen news stories. That is one event. If each source makes its own signal, you’ll draft to the same person three times.

The rule: while a signal is fresh, more coverage of the same type for the same account joins it instead of making a new one. That also handles a merger rumor reported over two weeks.

Scoring: weight, fit and freshness

A score you can’t explain is a score you can’t tune. Multiply four factors and store each one:

score = weight × account fit × persona fit × freshness

Weight is how much the signal type matters to your product. Account fit is whether the company is in your target tier. Persona fit is whether the signal involves one of your personas, and at what priority. Freshness falls from 1 to 0 across the type’s window, and past the window the signal drops out.

Here is a scored signal a day after the headline ran. The company and person are made up.

signal   exec_hire · Halvard Industries (fictional)
title    Halvard names Dana Okafor VP of Corporate Communications
score    0.84
reasons  weight       1.0    exec_hire weight in the brain
         account_fit  1.0    tier 'target'
         persona_fit  0.85   persona vp_comms (priority 2)
         freshness    0.983  60 days window

When a score looks wrong, the reasons tell you which setting to change.

The hook: their news, your read

The first message opens with their news and what your product’s analysis of it shows. Never with you. Cite the source of every claim about their coverage. Keep it under about 90 words, with one ask and no attachments.

An example from Northlight, a made-up media-analysis product, to the fictional VP above:

Subject: How Halvard's CFO news is landing

Dana, congratulations on the new role. Halvard's CFO transition
(8-K, 6 October) has run in 14 outlets so far. Most repeat the
release, but 3 trade titles frame it as a cost-cutting move,
which the release never says (links below). We ran Northlight
on the coverage and wrote a one-page read on where that framing
started and who picked it up. Want the full read?

The hook is the product working on their news. Delete the company name and the message falls apart. That is the test.

Pacing a small market

1. Approve every draft. See approval-first outreach.
2. Cap new contacts per day at a number you can review properly.
3. Limit contacts per account over a rolling window.
4. Send in the recipient’s time zone, on weekday mornings.
5. Stop on reply.
6. Send from a domain used only for this product.

In config, with illustrative values:

pace:
  mode: approve
  new_contacts_per_day: 8
  per_account_per_30_days: 1
  send_window: { days: [tue, wed, thu], hours: "09:00-12:00", timezone: recipient }

A worked example

Everything here is fictional. Halvard Industries is a made-up 9,000-person industrial company in your target tier.

• Day 0. Halvard files an 8-K with item 5.02: the CFO is leaving. Six news stories the same day join one leadership_change signal. No persona is involved, so no play starts.
• Day 3. “Halvard names Dana Okafor VP of Corporate Communications.” Halvard is the subject and a persona title appears.
• Day 4. The signal scores 0.84 and a play drafts to Dana, with your product’s read of the CFO news as the hook.
• Day 4, later. You trim a sentence and approve. It sends Tuesday at 09:20 her time.
• Day 11. The per-account limit keeps the play away from Halvard’s CCO this month. If Dana replies, the play ends for her.

Where Optic fits

Optic GTM runs this method. The signal layer is built and runs hourly in production: SEC EDGAR 8-K items 2.01, 2.02 and 5.02, news search through GDELT, newsroom RSS and Atom feeds, and IR pages. Detection is rules, one event makes one signal, and every score stores its factors. You review signals from Claude with list_signals (see MCP for GTM). Name collisions still get through today.

Drafting with the product hook, approve and reject, and sending are next on the build plan. We run Optic on our own products, and our enterprise comms product’s brain looks nothing like our developer tool’s: far fewer accounts, a slower pace, different signals. More on the product page.

If you sell to comms teams and want to try this on your own accounts, request early access.