Back in February I packaged up some old course notes as an email marketing skill. This is the bigger version of that idea, broadened out past email to the rest of a CRM programme.
crm-open-knowledge-wiki is a practitioner knowledge base covering email, lifecycle and CRM marketing: the customer data a programme runs on, the channels it sends through, the operations that keep it moving day to day, and the measurement that tells you whether any of it worked. Each concept carries a citations section grounded in primary sources: standards and RFCs, platform documentation, the regulators (FTC, ICO, FCC), and recognised industry and research references.
Written as an OKF bundle #
It follows the Open Knowledge Format, which Google published a couple of days ago. It's a spec for writing knowledge down as a directory of plain markdown files with a little YAML at the top of each one. Every file is one self-contained concept, and ordinary markdown links turn the directory into a graph you can enter from anywhere. There's no app and no database. You clone the repo, open any file, and follow the links. A person reads it the same way an agent does, which is the part I liked: the same files, no translation layer.
If you'd rather just read it, the same files are published as a browsable site at crmknowledgebase.com, built with Quartz.
How it's organised #
It's split into layers:
- Principles: the stances a recommendation should serve, like list quality, metrics discipline, and engagement as deliverability.
- Foundations: the cross-channel work, so identity and segmentation, consent, list building, lifecycle mapping, copywriting, automation, orchestration, ESP selection, and decisioning.
- Channels: the media you send through, email, SMS and RCS, push, in-app, and direct mail, each with its own permission model, filtering, technical specifics, constraints, and measurement.
- Measurement: the metric tree, holdouts and control groups, attribution, retention and LTV, experiment sizing, and incrementality.
- References: external platform behaviour, the legislation, and a glossary.
It's a work in progress #
The format is relaxed about that: broken links are allowed, and they just mark the bits I haven't written yet. It's MIT licensed, and the whole thing lives on GitHub.
Contributing #
Contributions are welcome, whether that's fixing an error, sharpening an explanation, adding a concept that's missing, or grounding a claim in a better source. Open an issue first for anything substantial; small fixes can go straight to a pull request. Keep each one focused on a single concept or change.
The sourcing bar is the main thing: every claim needs a primary or authoritative source, so standards and RFCs, platform documentation, the regulators, or real research, not a marketing page. The bundle is also vendor-neutral, which means disclosure. If you have a financial or employment relationship with a product, platform or blog your contribution mentions, say so in the PR. Disclosure is fine and won't sink the contribution; it's the undisclosed interest that gets things declined. Outbound commercial links are set to nofollow and carry no SEO value, so "add our URL as a source" PRs that don't improve the content get turned down.
The full conventions, frontmatter shape, house style and disclosure rules are on the contributing page.