Documentation
Plain-English notes on broken tracking, unclear attribution, slow launches, messy approvals, and growth numbers teams can trust.
Why growth work gets stuck across too many tools
Teams do not struggle because each tool is bad. They struggle because offers, pages, ads, email, spend, orders, and decisions live apart.
Read the documentMore docs
- 02 / 18Governance
Why approvals and audit logs cannot live in a separate tab
Governance fails when identity, authority, policy, approval, evidence, and audit live in separate tools instead of on the action itself.
7 min - 03 / 18Tracking & Attribution
Why tracking breaks before it reaches the dashboard
Tracking breaks across browser, network, identity, state, and trust. The final CAC or ROAS number is only as strong as the weakest layer.
7 min - 04 / 18Tracking & Attribution
The State of Web Tracking in 2026
Why web measurement broke — cookie decay, browser restrictions, and consent risk — and the blueprint for a tracking stack that degrades honestly.
18 min - 05 / 18Experimentation
Why growth teams should test the whole journey
Many growth moves span offer, page, creative, email, checkout, and follow-up. Testing one touchpoint often misses the real result.
7 min - 06 / 18AI & Networks
Why software has not learned from every customer
Shared software rarely turns every customer's data into better decisions for everyone else. Privacy, comparability, decay, applicability, and causality get in the way.
6 min - 07 / 18AI Architecture
Why AI answers need receipts
Finding relevant documents is not enough. Business AI needs to show which evidence supports which claim and what policy allowed the answer.
7 min - 08 / 18Software Markets
Why software categories stop meaning what buyers think
Category names are useful when they are young. Then they split, stretch, get acquired, get renamed, and stop describing the workflow.
6 min - 09 / 18Tracking & Attribution
What attribution has to prove
Attribution should prove identity, touchpoint, customer state, and cause. A model selector alone cannot make a budget number trustworthy.
7 min - 10 / 18Product Design
Why products end up with too many tabs
Products add tabs one reasonable request at a time. The result is a slower user and a navigation that mirrors the org chart.
6 min - 11 / 18Content & CMS
Why content tools need launch context
Modern content is part of campaigns, offers, experiments, approvals, and revenue. Storing the page is not enough.
6 min - 12 / 18Finance & Ops
Why month-end exists
Month-end exists because revenue, cash, payouts, refunds, and audit evidence drift during the month. The fix is fewer numbers that disagree.
5 min - 13 / 18Data & Honesty
Why a blank dashboard cell should never lie
A blank metric is not one thing. It may be a real zero, a failed calculation, an unsupported metric, or a hidden value.
7 min - 14 / 18AI & Trust
What every AI action needs to prove
If AI takes action, the company needs proof: who acted, what authority allowed it, which policy applied, what evidence was used, and where the audit record lives.
6 min - 15 / 18AI & Decisions
When the system should refuse to recommend
A recommendation engine that always answers is not safer. Useful systems know when the evidence is too thin and say so.
6 min - 16 / 18Benchmarks
Why benchmarks mislead good teams
Benchmarks only help when the cohort matches. Otherwise a clean median can push a good team toward the wrong target.
6 min - 17 / 18Engineering
Why slow systems often have too many meanings
Many systems get slow because the same business concept is represented several ways and the software spends its life translating.
6 min - 18 / 18Design System
A practical spacing system for product UI
A simple spacing system for product and marketing screens: fewer arbitrary values, clearer density rules, and less visual drift.
7 min
Curious how we use these in production?
Talk to the team