Field Guide

Notes on distribution, written down.

Anyone can build now. Getting found is the hard part. These are the frameworks we use to give every portfolio company distribution: how we pick what to write, keep the facts straight, measure what works, and run it all with agents. Short entries, plain words, linked like a wiki.

Growth loopsgorilla

Paid ads, run by agents: Google, Meta, TikTok and LinkedIn

Paid works best as an amplifier for messages organic has already proven. Agents can launch, test and rebalance across four platforms every day, as long as budgets are capped and every dollar is judged on paying customers.

Nov 15, 2026 · 3 min read →
Discoveryelephant

llms.txt: an experiment, not a requirement

llms.txt is a proposed file that points AI models to your most useful content. It's cheap to try and unproven. Ship it as a measured experiment and keep your expectations calibrated.

Nov 14, 2026 · 3 min read →
Agent operationsrhino

Build a claim library in a week

A claim library is a spreadsheet of every fact your company states publicly, with an owner and a source. Here's how to build one in five working days and keep it honest after that.

Nov 12, 2026 · 3 min read →
Truth & trustelephant

E-E-A-T for agent-run companies

Experience, expertise, authoritativeness and trust are how engines decide whom to believe. Agents can draft, but the experience and accountability have to come from real people, visibly.

Nov 10, 2026 · 3 min read →
Measurementlion

Kill lists

The most valuable output of a quarter is often the list of things you'll stop doing. Every channel, flow and page has to keep earning its place, or it goes.

Nov 5, 2026 · 3 min read →
Growth loopsgorilla

Organic and paid are one loop

Treating organic and paid as rival budgets wastes what each one learns. Wire them together: organic finds what's true, paid scales it, and paid feedback tells organic what to write next.

Nov 3, 2026 · 3 min read →
Discoveryleopard

Own your comparison pages

Buyers in the middle of the funnel search for comparisons. If a roundup or a competitor writes that page, they shape the decision. Write it yourself, and write it honestly.

Oct 28, 2026 · 3 min read →
Measurementgorilla

Retire last-click attribution

Last-click gives all the credit to whatever happened last, which is usually a branded search or a retargeting ad. For considered purchases, replace it with first touch, holdouts and geo tests.

Oct 24, 2026 · 3 min read →
Agent operationselephant

Agent-run content ops, with humans at the gates

Agents do the research, briefs, drafts, checks and monitoring. Humans approve anything regulated, irreversible or reputational. That split is what lets a tiny team publish like a large one.

Oct 21, 2026 · 3 min read →
Growth loopsbuffalo

Lifecycle is a growth channel

The cheapest customer is the one who almost bought, and the one you already have. Recovery, value-moment and referral messages often beat new acquisition, if you measure them with holdouts.

Oct 18, 2026 · 3 min read →
Measurementjaguar

Visibility isn't success: instrument first, then grow

Rankings and AI citations are leading indicators. Paying customers are the outcome. Measure both on one page, and build the measurement before the content.

Oct 14, 2026 · 3 min read →
Growth loopsbuffalo

K-factor, measured honestly

Referral multiplies a product that already works. A k below 1 is normal and still valuable. Inventing the number is the real failure.

Oct 7, 2026 · 3 min read →
Discoveryelephant

Schema that earns citations

Structured data tells machines what a page is about without making them guess. Use it to describe what's true. Never use it to decorate.

Oct 2, 2026 · 3 min read →
Discoverytiger

YouTube is a search engine. Run it like one.

Most companies treat YouTube as a brand channel. Treat it as search: answer one real question per video, and link every video to a page you own.

Sep 30, 2026 · 3 min read →
Truth & trustleopard

Entity truth for founders

Every engine keeps a picture of who your company is and who you are. Make sure that picture is accurate, consistent and confirmed by sources you control.

Sep 26, 2026 · 3 min read →
Discoveryelephant

The question map: choose content by intent before volume

List every question a buyer asks, group them by how close they are to buying, and write the pages that win the decision first.

Sep 23, 2026 · 3 min read →
Truth & trustjaguar

Claim lock: one fact, one version, everywhere

Answer engines cite the source that's most consistent. If your price or definition differs across your own pages, you've handed the answer to someone else.

Sep 16, 2026 · 3 min read →
Discoveryelephant

Search everywhere: SEO, AEO and GEO are one system

Buyers now ask the same question in a search engine, an answer engine and a chatbot. Treat all three as one discovery system with one set of facts and one scorecard.

Sep 9, 2026 · 3 min read →
Glossary

Words we use, defined.

AEO
Answer engine optimization. Earning the direct answer in snippets, AI overviews and voice. Read →
Architecture of attention
The deliberate design of how owned content, AI answers, creators and outbound work together to earn attention.
Blended CAC
Total acquisition spend across all channels divided by new paying customers. Read →
Boundary condition
A line agents or teams never cross, regardless of what a test suggests. Read →
Brief seed
A short brief per page: the question, the two-sentence answer, entities, required claims and sources. Read →
Brittle consensus
A belief an industry holds mostly out of habit, and that breaks quickly once someone tests it.
Calibrated confidence
Being exactly as sure as the evidence allows, and saying so. Read →
Causal inference
Measuring what a change actually caused, usually with holdouts or controlled tests. Read →
Citation share
For a fixed set of prompts, the share of AI answers that cite your domain. Read →
Claim lock
The rule that every page, schema block and brief must match one canonical claim library. Read →
Comparison page
A page on your own domain that honestly compares you with alternatives, including when you're not the right pick. Read →
Creative fatigue
When an ad stops performing because the same audience has seen it too often. Read →
Distribution loop
A system where each channel's output becomes another channel's input. Read →
E-E-A-T
Experience, expertise, authoritativeness and trust: how engines judge whom to believe. Read →
Entity truth
One canonical value for every attribute of your company, products and people. Read →
False dichotomy
Treating two options as either-or when the answer is both, like organic versus paid. Read →
Frequency cap
A limit on how many messages one person gets across all flows in a period. Read →
Gap list
Questions where a third party supplies the answer you should be giving. Read →
GEO
Generative engine optimization. Earning a citation when a chatbot writes the answer. Read →
Geo test
Turning a channel on in some regions and off in comparable ones to measure its effect. Read →
Holdout
A random group that doesn't get a channel or flow, used to measure what it actually caused. Read →
K-factor
New paying customers each customer brings in, discounted for retention. Read →
Kill list
The regular, written list of channels and projects you're going to stop. Read →
llms.txt
A proposed file that guides language models to a site's most useful content. Plausible, unproven. Read →
Necessary but not sufficient
Required for success, but not enough on its own. Like a product without distribution. Read →
Opportunity cost
What you give up by spending time or runway on one thing instead of another. Read →
Question map
Every question a buyer asks, grouped by funnel stage and priority. Read →
Runway
How long the company can operate before the money runs out. Every wasted sprint shortens it. Read →
Schema
Structured data that labels what a page contains so machines don't have to guess. Read →
Second-order effect
The consequence of the consequence. Where most of the interesting strategy lives. Read →
Share-worthy moment
The first moment a customer feels the value, and the right time to ask for a referral. Read →
Signal-to-noise ratio
How much of a metric or channel is meaningful versus distracting. Read →
Stop rule
A condition, set before a test starts, that ends it. Read →
Successful referral
A referral that paid or retained. Never a click. Read →
Viral cycle time
Time from one customer to the next customer they bring in. Read →