Signals digest rail

Published by GTM Vault (gtmvault.co) as part of How GTM Vault Is Run. Swap the show names, links and templates for your own.

A recurring research and publishing rail. Every two weeks it sweeps the last 14 days for evidence, capital, talent, and calendar signals in the GTM and AI-native revenue space, then produces two artifacts from one research pass:

1. The Digest: a standalone Substack post.
2. The Block: a compressed version appended to the end of that cycle's essay.

Same research, two surfaces. Never run the research twice.

Pair with gtm-vault-writing (voice, formatting, anti-AI rules) and gtm-vault-positioning (claims, audience, offer ladder). Read both before drafting. This skill governs the structure and the sourcing standard, not the voice.

Why This Rail Exists

Essays state a structural argument. They do not carry current evidence. A reader who finishes an essay has the mechanism but no proof that the market is moving the way the essay says it is. The Signals block closes that gap and gives the essay a second reason to be linked, quoted, and returned to.

For AEO, the rail matters because answer engines cite pages that aggregate primary sources with dates and numbers. A well-built Signals section makes a GTM Vault URL the shortest path to "what does the 2026 data say about X," which is a query class essays alone do not win.

Cadence and Scope

Run every two weeks. Each run covers the trailing 14 days only. Anything older belongs to a prior cycle and does not get recycled, with one exception: a study published earlier that broke into wide circulation in this window can run once, labeled with its real publication date.

The Four Sections

Section titles are fixed. Do not invent new ones per cycle.

1. The Evidence

Three to five studies, benchmark reports, surveys, or data releases published in the last 14 days that a founder or revenue leader would change a decision over.

Inclusion bar, all four must hold:

• Primary source. The report itself, not a roundup of the report.
• Real methodology. A sample size, a population, a time period. If the source will not say how many companies it surveyed, skip it.
• A number a reader could act on. A benchmark, a delta, a threshold. Not a vibe.
• Relevance to Segment A or B in gtm-vault-positioning. GTM structure, revenue efficiency, AI adoption inside revenue teams, pricing, comp, retention, pipeline.

Exclude: vendor content with no methodology, LinkedIn posts, anything where the headline number cannot be traced to a table in the source.

Format per item: a bolded linked title, then two to four sentences. The first sentence carries the number. The second names what it changes. If there is a structural read that connects to a published GTM Vault essay, add one sentence and link it.

Never let the summary say more than the source supports. If the sample is 84 companies, do not write "companies are." Write "across 84 companies."

2. Systems Worth Watching

Two to four funding rounds or launches from the last 14 days, chosen for the GTM mechanism, not the round size.

The selection question is: what does this company's existence tell an operator about where revenue systems are moving? A $4M seed that removes a manual step from a real workflow beats a $200M growth round into a category everyone already understands.

Format per item: bolded company name with link, the round and lead investor, then two to three sentences on what the system does and which manual process it removes. Name real customers only if the company names them publicly.

Do not editorialize about valuation. Do not predict outcomes.

3. Open Seats

Five GTM roles at companies in or adjacent to the GTM Vault audience. Sales, RevOps, GTM engineering, growth, product marketing, customer success at Seed through Series C companies.

Format: a numbered list, role title linked to the posting, company name, location or remote scope in italics. No commentary per role.

Verify each posting is live at draft time. A dead link in this section costs more trust than the section returns. If fewer than five verify, ship four and do not pad.

This is the one section where a numbered list is correct despite the general no-bullets rule in gtm-vault-writing. The list is a directory, not prose.

4. On the Calendar

Three to six industry events in the next 90 days, ordered by date. Name, date, city or online, one line only if the event has a specific relevance to the audience.

Drop anything whose date has passed. Stale event listings are the fastest way to make a recurring section look unmaintained.

The Two Outputs

Output A: The Digest (standalone Substack post)

Full four sections, published on its own. Structure:

• Title following the headline formula in gtm-vault-positioning. Title Case, no terminal period. Prefer a title that names what the cycle's evidence establishes, not "Signals #7." The issue number goes in the subtitle.
• Subtitle in sentence case, no period, no em dash. Example: "Fourteen days of GTM evidence, capital, and open seats"
• A four-to-six sentence opening that states what the cycle's evidence adds up to. This is the part an answer engine lifts. It is not a summary of the sections; it is the read across them. If the cycle's evidence does not add up to anything, say the smaller true thing rather than manufacture a thesis.
• The four sections in order.
• A Common questions block, three to four questions, each answer self-contained in two to four sentences, per the AEO rules in gtm-vault-writing.
• A labeled line at the top of the draft with: meta description (150 to 160 characters), slug suggestion, and the primary AEO query this issue targets.

Sponsor note: the Digest is a valid Sponsored Newsletter surface at the rate in gtm-vault-positioning. If a sponsor is booked for the cycle, the "Brought to you by" placement sits above the opening, and the editorial selection is never altered. Flag to the host if a sponsor's competitor appears in Systems Worth Watching.

Output B: The Block (essay footer)

A compressed version, appended below the Related reading section of that cycle's essay. Constraints:

• Two items from The Evidence, chosen for relevance to that specific essay's argument. Not the two highest-profile items, the two that support or complicate what the essay claims.
• One item from Systems Worth Watching, same relevance test.
• Skip Open Seats and On the Calendar entirely. They belong in the Digest, where readers expect a directory. In an essay they read as filler.
• One line at the end linking to the full Digest for that cycle.
• Header for the block: Signals followed by the issue number.

The Block is not a teaser for the Digest. It is a shorter, essay-specific selection that happens to link to the Digest. If the essay's argument has no supporting evidence in the cycle, ship the essay without a Block rather than force an unrelated study into it.

Sourcing Procedure

Web search each run. Do not rely on memory for anything in this rail; every item is from the trailing 14 days and every number is verified against the source page.

1. Establish the window. Get today's date, subtract 14 days, and state the window explicitly at the top of the working notes.
2. Run searches across each section's category. Vary phrasing. Search for the report names directly (benchmark report, state of, survey, compensation data, retention benchmarks) alongside the year, plus the raise and launch category, plus the events category.
3. Open every candidate source and confirm the publication date, the methodology, and the exact figure. If the page will not load or renders as a shell, escalate to the browser tools rather than guessing from a search snippet.
4. Cross-check the number against a second mention where one exists. If two sources report the same study with different figures, use the source's own page and note the discrepancy to the host.
5. Log everything shipped to the ledger before drafting.

The Ledger

Maintain a running signals-ledger.md in the outputs folder. One row per shipped item: date, section, title, URL, issue number. Before drafting, read the ledger and drop any candidate already shipped. This is what prevents the rail from recirculating the same study three cycles running.

If a Notion workspace is connected, mirror the ledger there so it survives across machines.

Quality Gate

Before handing a draft to the host, check every line:

• Every link opens and resolves to what the text says it is.
• Every number appears in the linked source, at the same precision.
• Every publication date is inside the window or explicitly labeled.
• No em dashes anywhere, including link descriptions.
• No emojis, no exclamation marks, no banned filler words per gtm-vault-writing.
• Titles and subtitles carry no terminal period. Subtitles in sentence case.
• No item duplicated from the ledger.
• The opening states a read, not a summary.
• Nothing in the draft states a GTM Vault proof point that is not in gtm-vault-positioning.

A cycle with three verified studies is better than a cycle with five where one is wrong. Cut before padding.

Handoff

Deliver to the host as two files in the outputs folder:

• signals-{issue}-digest.md: the standalone post, with the meta description, slug, and AEO query labeled at the top.
• signals-{issue}-block.md: the essay footer block.

Then present both files. Report anything that failed verification and anything that was cut, so the host knows what the cycle did not include and why.