Briefing: the vision for CRM and comms
Briefing: the vision for CRM and comms
TL;DR
We're moving our contacts and newsletters onto a small setup we own: Twenty (our contact database, or CRM), Resend (which sends the email) and a small mailer for signup forms and email courses. Everyone who signs up for anything should land in one list, automatically, so we can find the right people for a launch. AI prepares newsletters; a person approves them in a couple of minutes. Signups and email courses already work, and the first real newsletter (Way Into AI) went out on 5 Oct. Next: make sure it reaches inboxes, then move the Life Itself and Seeds of Renaissance newsletters off Substack.
What I need from you
- A volunteer to do a real newsletter send without me (
crm-dxb.4.4). The guides exist; this is the test of whether they're good enough. - Views on the open questions at the end, especially whether we keep Substack for Seeds of Renaissance paid subscribers, and which team mailboxes (if any) should connect to Twenty.
The problem
People reach us through lots of doors (Substack, Luma, courses, forms, retreats, hubs, email), and each door keeps its own list, or none. When 27 Teal Estate emails arrived recently there was nowhere to put them. So we can't answer a simple question like "who might care about this new course or white paper?"
Newsletters often don't get sent either. Substack has no way for a program to send a post, so someone has to log in and click through it. As the use-cases doc puts it: "That final manual step means posts often do not get sent." And we're a tiny team: "any workflow requiring manual data entry or manual imports will not happen."
The vision
- One record per person in Twenty: who they are, where they came from (which form, event or project, and when), what they've signed up for, and what they've agreed to. New activity gets added; nothing is overwritten.
- Contacts arrive automatically from forms, Luma, imports and email. Manual import happens once, to bring in old lists.
- We use it to pull together the right people for a launch, look someone up before a meeting, and leave a note and a follow-up owner.
- We write newsletters where we already write (Google Docs, Markdown, the website). AI tidies and assembles the issue, the recipient list and a preview; a person approves the exact content and recipients in under two minutes, and then it sends. As the use-cases doc says, "Programmatic publishing does not mean ungated publishing."
- Setting up is one request: "Add a newsletter/course signup to this site" should create the form, the list, the link into Twenty, and check it works. Each project keeps its own name, domain and lists, while we run one shared setup, and unsubscribes and bounces are respected everywhere. The guides should be good enough that someone else, or another AI session, can run all this without me.
Substack stays only where we deliberately want its network, app or paid subscriptions.
Where it stands
The first real newsletter, Way Into AI week 40, went out on 5 Oct to 30 people, approved by me (crm-dxb.3.18). We haven't yet confirmed it reached inboxes, and an earlier test landed in spam.
| Job | Status (as of 9 Oct) |
|---|---|
| Newsletter signup on any site | Working since 26 Sep. Way Into AI's 29 Substack subscribers are imported. |
| Interest form on a project site | Working; each site's form still needs checking. |
| Email course (confirmation, lessons, status in Twenty) | Working, for three Developmental Spaces courses. |
| Turn notes into an approved newsletter and send it | Partly: first real send done; inbox delivery not confirmed. |
| Capture event attendees and old lists | Partly: past Luma events imported (106 events, 1,748 attendances); regular sync not set up. |
| Find the right people for a launch and contact them | Partly: the pieces exist, not yet shown end to end. |
| Look up someone's history, keep notes, assign follow-ups | Not yet a team habit. |
| Someone other than me runs these | Partly: guides exist, no real send yet. |
| Send updates to a course or event group | Not started. |
What's next
Tracked in the crm Beads epic crm-dxb:
- Confirm the newsletter reaches inboxes and fix spam placement (
crm-dxb.3.9,crm-dxb.3.20). - Put the lifeitself.org signup form back and remove the Substack one (
crm-au0). - Someone other than me does a real send (
crm-dxb.4.4). - Show a real "find the right people" and "look someone up" in action (
crm-dxb.5.3). - Set up regular Luma import (
crm-dxb.6.3). - Decide how long we keep data after someone asks to be deleted or opts out (
crm-dxb.11.4). - Move the Life Itself and Seeds of Renaissance lists off Substack (not yet in Beads).
We're done when "Both newsletters (Life Itself, Seeds of Renaissance) run on Twenty + Resend, embedded on the websites, with Substack lists migrated without subscriber loss and written operating procedures." Way Into AI is the pilot that proves the path.
Open questions
- Seeds of Renaissance paid subscribers bring in about 1,000 a year (currency to confirm) via Substack. Do we keep Substack for paid subscriptions, or move? Paid membership needs its own research and "must not determine the current email-infrastructure conclusion".
- Should this sit under Media House (the provisional answer) or Ops?
- What happens to Open Revolution's Brevo email sequence, and to Life Itself Research's Substack (and its future name)?
- Which team mailboxes connect to Twenty, and under what privacy rules?
See decisions still needed and still to confirm.
More detail
- Comms Infrastructure project: the planning record.
- Communications use cases: the current state, cross-project needs and communication stories.
- CRM motivation and job stories: why we need one contact list, how contacts get captured, stored and used and the nine job stories. The two docs will be merged into one once the open questions settle.
- How it's built: the crm repo, including email infrastructure, where contact data lives and the job-story scorecard. Open tasks:
bd ready --parent crm-dxbin the crm repo. - The channel inventory lives in the comms repo.