Fable's end-to-end service plan, digested · the full document is 828 lines in the vault
Fable came back with a service plan you could hand to an operator, not a deck: offer, price table, the thirteen-step migration runbook, the posting pipeline, contracts, a pilot calendar through January, and the risk register. This is the whole thing in one screen. Two lines carry the strategy: the migration is the door, priced near cost on purpose; the monthly retainer is the business. Six decisions are waiting in week zero.
The offer
Exit S · up to 12 pages, no blog. SZL is the shape.$1,200–1,500
Exit M · 13 to 25 pages. Cliff's firm is the shape.$2,500–3,000
Exit L · 26 to 40 pages or a blog port. RJ is the shape.$3,500–4,000
Care, organizations · 2 posts a month, ≤400 words, 30 min edits$200–400 / mo
Care, firms · 4 posts a month, ≤600 words, attorney review built in$500–800 / mo
Sizes and tiers come from the one-hour audit, not from negotiation. Pilot prices, fixed: D&W $2,500 + $600/mo (a test case, invoiced for real, zero friction), RJ $3,500 + $650/mo (the external proof; the blog port is the hard part), SZL $1,350 + $300/mo (the org template; $1,500 if the .org question adds work).
The economics
The exits run at or below cost at the floors, and Exit S loses money at any price in its band. That is the design, not a mistake: the retainer carries the business. Consequences: an Exit is never sold without Care attached, the first Care term is three months, and payment is 50/50 (signature / cutover).
At target, a firm retainer bills $650 against $140–230 of work: 65 to 78 percent. The org tier pays for itself at the $300 target, so the plan quotes that, not the $200 floor.
Year one, three-site pilot, modeled: est
D&W · $9,700 revenue$4,290–6,120 margin (44–63%)
RJ · $11,300 revenue$4,790–6,920 margin (42–61%)
SZL · $4,950 revenue$1,320–2,830 margin (27–57%)
Three-site pilot, year 1$25,950 revenue · 40–61% margin
Every number here is modeled. The pilot's stated job is to replace each one with logged actuals, site by site.
The door pays for itself. The monthly retainer is the business.
How it runs
Sell
The intake form asks the questions that decide the job: registrar, who holds the Squarespace login, whether email runs on the domain.
One-hour audit, then a one-page quote, valid 30 days. What we found, the Exit, the Care terms, what is not included.
One closer per deal: Amantu on firms and anything web, Kenzie on orgs she already knows. Cliff refers and never quotes his own firm.
Post
Weekly batch: drafted Monday, reviewed internally Tuesday, sent Wednesday. The client's window is three business days.
Nothing publishes without written approval. For firms, the attorney signs off every post, and the approval log is the record.
Publish lands within 48 to 72 hours of approval. The monthly digest goes out by the third business day.
Migrate
The live crawl decides the URL inventory, never the Squarespace export. Every old address gets a redirect, tested 100 percent before and after cutover.
Cutover runs Tuesday to Thursday, morning, never Friday. Squarespace is cancelled only after 14 verified days, and only the site plan.
DNS records are inventoried before anything changes, and email gets a send-and-receive test at cutover. A law firm without email is the scenario this rule exists to prevent.
Operate
Cloudflare hosting costs $0 a month: no CMS, no plugins, no admin login to attack. WordPress is the obvious alternative, and its upkeep is the exact work this service is sold against.
Analytics and Search Console stand up on staging before cutover; Squarespace numbers get snapshotted before cancellation.
Incidents run S1/S2/S3 with response targets, and rollback is reverting a commit. A runbook card per site means an S1 does not wait on one person.
The Exit: nine steps from handshake to steady state.
What Fable found that the research missed
Ops tooling was invisible in the cost model. The research costed three toolkit pieces at $1.5–2.3K. Running the service needs eight more (forms kit, digest generator, intake, approval tooling, uptime, SOW variant, runbook kit, QA kit): another $1.8–2.7K est. Recommendation: treat those as Design Garden operations investment that serves every static build, and say so out loud at the margin recompute.
The five-site toolkit amortization is optimistic. Only three sites are in the paid pilot. On a three-site basis, toolkit costs $500–770 per site, and Exit M at the floor slips to −$470. Ishango starts from the three-site figure at the first gate.
The org floor is a 30 percent month. $200 against a $140 cost month. That finding alone is why the target quote for orgs is $300.
The pilot
Week 0 · Sep 15: decisions locked; D&W SOW signed and half invoiced; toolkit builds start.
Week 4 · Oct 13–17: D&W cutover. Care bills Nov 1.
Gate G1 · late Oct: D&W live, redirects 100 percent, zero S1 incidents in 14 days, every hour logged. Then RJ starts.
Early Dec: RJ live, blog port and per-post redirects included. Gate G2, then SZL.
Dec 22 to Jan 2: SZL cutover before the break if staging approves; otherwise the first week of January.
~Jan 19 · Gate G3: actuals replace every modeled number; price bands, name check, and whether to open beyond the circle all get decided together.
One migration in flight, ever. The next one does not start until the last one is verified live.
Capacity: the real constraint is attention, not agents
Budget during the pilot: 8 hours a month of Amantu-time, about two hours a week. A firm client takes 1.25–2 hours a month; an org takes about one.
That math caps the book at five Care clients and one migration in flight. The cap holds until instrumentation says otherwise.
Levers for later, in order: a trust ladder that cuts firm review to one post in four after three clean months, Cliff as a paid second reviewer for non-D&W firms legal, and a client approval surface, not before six clients.
Decisions waiting in week zero
Pricing floors and pilot prices as written in the plan.
Who closes: Amantu firms, Kenzie orgs, Cliff refers.
Entity: Design Garden LLC. Action items: a Design Garden SOW variant and Stripe under DG.
Default DNS path: nameservers to Cloudflare.
Care billing starts the first full month after cutover.
GA4 stays off by default; Cloudflare Web Analytics plus Search Console is the stack.
Still open from the research: what szlalphas.org is meant to be. Needed before the SZL quote in week 8; the plan covers both answers. Week 3 brings the Care terms and portfolio rights through Cliff's legal pass. A week without these six costs the pilot a week.
Top risks, and the rule that answers each
Scope creep at staging (likely): one correction round only; a scripted reply for design asks; new pages are add-ons or a ladder quote.
The sign-off treadmill (likely, for firms): weekly batching, a three-day window, and drafting pauses after two silent batches rather than chasing forever.
Orphaned DNS records (email breaks after the nameserver change): full record inventory before any change, records rebuilt first, send-and-receive test at cutover.
Capacity (certain): the five-client cap, one migration in flight, and Yvonne does all pre-review so Amantu's ten minutes are the only ten minutes.
Holiday timing: no cutover near a holiday or on a Friday; the calendar already pads those weeks.
The toolkit
Eleven reusable pieces, 45–66 hours to build, about $3.5–5.1K at the internal rate est. The core three (content extractor, page-template family, redirect mapper) are 21–31 hours; the rest is operations tooling. Build order starts week zero with the minimum Cliff's site needs, and every piece serves all of Design Garden's static work, not just Perennial.