Ten years of earlier-signal practice, in twelve modules.
This is for the First Mover: a dealmaker who evaluates companies but doesn’t read code, corp-dev, PE, an angel, a scout, a seed fund. You never read a line of code, the read is done for you. What you get here is the practice behind it: how to notice startup momentum earlier, and act on it before the round, in plain business English.
Twelve modules · 340 minutes total · designed to be read in one sitting on a Saturday morning, or in chunks across a week. Skip the modules you already know. Each module ends with links to the live page where its idea is wired into the product, so you can see it in action, not just read about it.
Module 1 · 25 minutes
Identity, name the buyer you actually are
Question: Why does identity precede method?
If you name the problem clearly, the right tools, prices, and rhythm start making sense much faster. If you name the problem badly, you buy the wrong stack and then blame the price. That first naming move shapes everything that follows.
Question: Why commit-velocity acceleration, not commit count?
Counts are noise. Acceleration relative to a company's own baseline is a regime change. We use a 14-day rolling window, two-period confirmation, contributor-quality filter. The SSRN paper (n=219) shows the signal preceded fundraises by 21-47 days IQR. The methodology is the only durable moat, every other element on the site rests on it.
A dealmaker who evaluates companies but doesn't read code, solo angel, scout, seed fund, corp-dev, or PE operator writing €5k-€50k checks (or building an acquisition shortlist), 5-40 calls a year, who wants the engineering read translated, not raw. Three thesis axes: AI infrastructure, dev tools, technical SaaS. The disqualifier is as important as the qualifier, Series-B+ partners with six-figure data budgets are not us.
Funnel architecture, six rungs, one ladder, one cadence
Question: Why six tiers and not three? And why a monthly drop on top?
Free → €7 → €49 → €197 → €497 → €1,997. Every rung exists because the rung below it doesn't fit one specific check size or one specific cadence. Free is for cadence-builders. €7 is for thesis-testers. €49 is for the daily-rhythm buyer. €197 is for the syndicate. €497 is for the small fund. €1,997 is for the deep-dive on a single sector. Never let a buyer leave at zero commitment when there's a free rung below. Above the rungs sits the continuity layer, a net-new monthly drop (sector deep-dive, methodology release, founder essay, or tool) that turns the paid tier from a tool subscription into an anticipation engine.
Reverse-engineering, what the leaders are optimising for
Question: How do you read a competitor's funnel?
If you can explain the difference between timing and verification, the rest of the comparison work gets easier. Harmonic optimises for partner budgets. Tracxn optimises for sector breadth. Affinity optimises for warm-intro routing. The useful teardown is the one that shows which job they solve well, and which one they leave open.
Question: What's the single belief that, if held, makes everything else inevitable?
If commit-velocity acceleration is the most leading public signal in venture capital, every other deal-flow source, pitch decks, AngelList, Crunchbase, warm intros, is a lagging indicator. The whole investing thesis falls or stands on whether that belief is true. Once you accept it, everything we sell follows automatically. If you don't accept it, no amount of stack will move you. Find this belief for your own product before you write copy.
The Conversion Story, five steps from belief to action
Question: Old way → new vehicle → external → internal → frameworks. Why this order?
The canonical conversion script. (1) Name the old way the buyer was sold, for us, 'best deals come from your network.' (2) Reveal the new vehicle, engineering acceleration. (3) Remove the external struggle, you don't need fund-grade tooling. (4) Remove the internal struggle, you don't need to become a different person to source. (5) Show the frameworks, Sunday digest, Wednesday filter, end-of-quarter sweep. Each step is a beat the reader has to walk through; skip one and they bounce.
Question: Why itemise standalone value before the price?
Eight objects, each with a standalone value, totalling €1,728/yr. The buyer's brain anchors on the total. The price (€49/mo = €588/yr) becomes a 14× discount, which is the actual story. Stack order matters: the most desirable thing first, the bonus last. Never bury the dashboard inside the methodology vault. Lead with the thing the buyer wants most.
The Closes, five named patterns, one buyer profile each
Question: Money / Identity / Pricing / Urgency / Encore, when does each fire?
Money close fires for the spreadsheet buyer. Identity close fires for the engineer who's tired of pretending to be a partner. Pricing close fires for the rate-anchor, €1,728 retail vs €588 founding. Urgency close fires for the calendar, every Monday skipped is one 21-47-day window closed. Encore is the safety net for everyone, eight lines, the whole offer in one block, before the FAQ. One of the five always lands; you don't pick, you stack.
Traffic, earned, owned, and the one we don't pay for
Question: Why earned and owned, not paid?
Owned: email list (free Acceleration Watch), RSS, MCP server in the buyer's IDE. Earned: Reddit AEO, dev.to long-form, Substack mirror, federated social (Bluesky / Mastodon / Farcaster), academic SSRN citation. Paid: deferred under HOLD until earned-only proof. The reason isn't ideology, it's compounding, earned channels keep paying after you stop. Paid channels stop the day the budget does.
Question: Why publish to /md, agents.json, llms.txt, and OpenAPI?
Half the readers in 2026 are agents, not browsers. We publish six redundant agent surfaces: /md/<path> markdown mirror, agents.json discovery, llms.txt + llms-full.txt corpus, OpenAPI 21 endpoints, knowledge-graph.json, model.json. The agent-side reader doesn't see your hero CTA, they see the schema. The site is 1,060+ pages on the human side and 1,400+ surfaces on the agent side. Both sides matter.
Anonymity rule: no podcasts, no founder-face content, no real-name signatures. The product is a dataset, not a personality. Methodology rule: every claim is reproducible against the public Zenodo dataset; we don't keep a private edge. Pricing rule: founding-member rate locks forever before the public hike, never renegotiated retroactively. Free-tier rule: the 5 core MCP tools are free forever; new paid tools are added alongside, never gated. Scale stops where the buyer's trust would have to.
Don’t pick one module. Run the literal 30-day plan.
The curriculum compresses a decade. The quickstart compresses the curriculum into a calendar. Day 1 is the day you read this; Day 30 is the day you either lock the cadence or hit refund. Both are wins, the only failure mode is opening this page and doing nothing.
Week 1 · Foundation
Name the buyer you are. Read the methodology. Wire the free tools.
You owed it to yourself to test, not to commit. The guarantee exists for exactly this beat.
DAY30
If the cadence did earn its place, schedule Month 2's Sunday read in your calendar.
15 min
Months 2-12: same beat, lower friction. The decade compresses because the cadence is locked.
Or, if 30 days feels heavy
Pick Module 6, the core claim, and live in it for a week.
The 30-day quickstart above is the disciplined path. The seven-day “one module” path is the soft path. Both beat zero. Once you have your core claim re-written for whatever you’re building this quarter, every other module re-reads in five minutes.