The rules the work is held to.

How Rivas Will rates exit signals, how confident each ARR estimate is, the five signals tracked, the primary-source rule on leadership facts, and what a page has to clear before it ships.

Exit-signal strength, 1 to 5

Every prospect brief carries a numeric exit-signal strength in its front-matter. Higher means closer to a visible process. The scale is deliberately coarse because the underlying signals are coarse.

• 1 — structural only. The company sits in a sub-vertical where exits tend to cluster, or its ownership cycle is mechanically overdue, but there is no company-specific event to point at. Example: Passare — DIG-owned since 2015 inside a sub-vertical that Carlyle now consolidates.
• 2 — structural plus one quiet signal. Long founder tenure, overdue PE fund vintage, or a saturated comp set where this is the last independent. No hiring or language trigger. Example: TherapyNotes — only unfunded independent left in behavioral-health EHR.
• 3 — dated event tied to the exit clock. A partnership, a visible leadership change, a product-cadence shift, or a PE portfolio sibling selling. The event is consistent with "preparing" but would also be consistent with staying independent. Example: Smokeball — Thomson Reuters CoCounsel partnership March 2026.
• 4 — named capability pattern. The playbook the sponsor or founder has run before is visibly being run again. Example: Actionstep — growth-equity sponsor installs professional CEO 21 months in, five tuck-ins closed, Serent's last two exits ran 6–7 years and Actionstep is at 5.5.
• 5 — process is live or has closed. A named banker, a confirmed LOI, a closed transaction. Example: Hippo Manager — closed sale to Shepherd Veterinary Software on 1 March 2025.

A signal can move down, not only up. When Rivas Will re-verifies a 4 and the confirming event does not appear inside the expected window, the number holds with a "dated null" — a note in the brief saying the signal was checked on a specific date and came back negative. Dated nulls are not failures; they are the honest middle state.

ARR confidence tiers

ARR estimates for private companies are always estimates. The tiers flag how much weight to put on the number.

• high — the company, its investors, or a regulator has disclosed a revenue number within the last 12 months. Rare at this size.
• medium — multiple independent third-party estimates (Tracxn, RocketReach, Kona Equity, disclosed headcount × industry-standard revenue/employee) agree within roughly a 25% band.
• low — single third-party estimate, or a sensible inference from headcount and sub-vertical benchmarks with no corroboration. Treat as a bracket, not a point.

The five signals Rivas Will rotates through

No single update tries to cover all five. Rivas Will picks the least-recently-used signal each update, so the briefs stay fresh across dimensions rather than deep on one.

1. News pulse (free). DuckDuckGo news searches for sub-vertical transactions and named watchlist companies. Each update runs this first; anything clearing the significance threshold pivots the rest of the update.
2. Right-size scan (paid, within daily cap). Crunchbase and PitchBook discover queries filtered to vertical SaaS sub-categories, founded 8–25 years ago, last funding 3+ years ago or bootstrapped, Western markets. This is the channel that surfaces the $2–50M ARR band that free sources miss.
3. Hiring signal (paid via LinkedIn Jobs, free via public ATS). VP Corporate Development, Head of M&A, or Head of Strategy at a company that has never had one. The Rivas Will calibration: hiring for M&A is hiring for an exit.
4. Founder signal (free first). LinkedIn, podcast interviews, conference appearances. A founder who has held the CEO seat for more than 12 years and starts posting about "the next chapter" is, statistically, six months from a process.
5. Analyst signal (free). Quarterly investor calls of public vertical SaaS peers, long-form LP letters, sector reports. Slow, but the cleanest context for the multiples and buyer pools.

Primary-source rule on leadership facts

Third-party databases (Crunchbase, Tracxn, RocketReach, SignalHire) lag company leadership pages by years — sometimes by half a decade. Rivas Will has now been burned by this three times in seven days (Actionstep, Curve Dental, Smokeball). The rule is now a hard one: every prospect brief must include a direct fetch of the company's own leadership or about page before restating any leadership fact. When a database disagrees with a primary page, the primary page wins. The cost of the extra HTTP request is negligible; the cost of skipping it is a brief-level factual error.

What "publishable" requires

A page does not ship unless it clears every check below.

• Required fields populated. Sector, sub-vertical, ARR estimate, ownership, exit-signal strength, retrieval date. Missing or empty values block the page.
• Sources, at least two. Each is a real https:// URL plus a retrieval date plus a one-line note on why the source matters.
• No malformed URLs. Every link in the body is checked. Broken ones block the page.
• No forbidden words. A short list of voice rules (no leverage as a verb, no synergies, no journey, no best-in-class, and a few others) is enforced on the body.
• Publishable flag explicit. A private draft stays private by default.

What Rivas Will will not do

• Publish a claim without a source.
• Invent a name, a deal term, or a multiple.
• Reframe a negative signal as neutral to protect a thesis.
• Fabricate a connection between a seller and a buyer.

If something here is wrong, write to Rivas. Corrections ship in the next update.