01
the studio
[British Columbia, Canada] [remote] [full-time] [May 2026 to Oct 2026]
I have since left The Atelier. Everything on this page is past work.
Co-founder, The Atelier. The Substack positioning studio behind Home Abroad, which turned founders' offline credibility into a Substack worth following through brand positioning, systems and content support.
- clients
- 3
- revenue in the first 3 months
- $5,000+
- first Signature Build, in the pipeline
- $3,000
02
the business plan
I wrote Atelier's complete business plan with my co-founder, from the first idea through the offer, the workflow and the risks.
The gap it was built around: the credibility-visibility gap. Established people with authority elsewhere, whose Substack does not carry it (a founder with a proven consultancy and 50k LinkedIn followers, publishing on a 400-subscriber Substack). We did not sell "Substack help", which anyone can undercut. We sold an authority system: positioning, a redesigned publication, a command centre, a content calendar, working automations and the rhythm to keep it shipping.
The plan set out who the client was and was not, the three kinds of competitor (ghostwriters and content agencies, newsletter growth consultants, personal-brand coaches) and our edge: builders not consultants, a psychology layer, practitioners who run their own profiles, and warm access through the Home Abroad network. It also set out how to start: find the tension, ship a first version, gather the right few, and let reality edit the plan.
The ladder at launch: the Authority Audit at $500, the Positioning System at $1,500 and the Growth Retainer at $850 a month, with a menu of add-ons in between. The plan also covered who does what between the two founders, five rules for every client interaction, the tech stack, the go-to-market, the discovery call beat by beat, pricing psychology, the money, the risks, and a 2027 direction: a free Atelier Substack, a paid community, and a track for company Substacks.
03
whom we targeted, and what we delivered
We targeted founders, operators and established voices with authority elsewhere and a paying business, whose Substack had not caught up. The sweet spot was recent publishing effort, under 1,000 subscribers and thin engagement. We did not take anyone whose newsletter was their income, or who was already winning.
The Authority Audit (5 to 7 days, a Notion page plus a 60-minute walkthrough): a full Substack teardown, a short audience growth read, a 100-point scorecard across seven dimensions, a positioning verdict, and a ranked build blueprint of 10 to 15 moves.
The Positioning System (about 3 weeks): a positioning and brand pack, Substack design and clean-up, three visual identity directions and a final identity, a Notion command centre, an analytics dashboard, a 90-day content calendar with 90 written Notes, a build review call with QA, and a final handover walkthrough.
Add-ons, quoted separately: deeper audience analysis, a DM template library, a Pinterest starter kit, Writer's Desk sessions, publication tune-ups, archive rescue, existing-post optimisation, recommendations outreach and Substack Note scheduling. After the build came the Growth Retainer or advisory.
04
the business workflow
Every lead followed one workflow, whichever door they came through.
Two entry paths. Path A was a handpicked or researched lead: warm (friends, Home Abroad connections, referrals) or cold, found through research and approached with one specific gap named. Path B was inbound through our content: LinkedIn, Notes, the launch post, the Authority Scorecard or the landing-page quiz.
Every lead was scored and given a client card first. Then a discovery call before anything else, a tailored proposal after it, then agreement and payment, onboarding, delivery, and an upgrade only if the delivered work made it useful. There was no one-page overview, no quiz for people we had approached ourselves, and the website was optional context, never homework.
Qualification used a 25-point score with two gates. Follow-ups went out after five working days in the same thread and stopped after the second unanswered one.
05
my part, from first message to handover
I assisted with finding clients, emailing or messaging them and sending the follow-ups. Once a lead was booked for a discovery call, I worked on the pitch deck, and we took the discovery calls together.
For each discovery call I built the deck in Canva from one fixed template: about 17 slides that moved from the client's world to the gap, the fix, who we were, the method, where to start, proof and the invitation. The tailored slides and every speaker note were rewritten for that person, about 45 minutes of work, and each call had a call-prep card and an after-call checklist.
After the call I created the client proposals, the invoices and receipts, and the client contracts, signed electronically, with a showcase clause so that no case study could use a client's name without their written approval.
Once a client was in, I built two pages for them: a client-facing Build page, and an internal page with the full task board (owners, hours, deadlines and handling notes for that client).
06
the launch plan
We decided to launch Atelier with the Authority Scorecard, not the offer ladder, and to run the first audit in public. The ladder stayed live on the website for anyone who looked, but out of the launch content: three rungs on day one is a menu, and menus do not convert strangers.
Three phases: a closed beta (12 hand-picked DMs and no public post) to test whether the score felt true, quoted the publication accurately and was something people would share; a soft launch in Chat first, then one Note; and the announcement Note only once the tool worked and the first audit was done.
The plan ruled out a public "want to be a beta tester?" Note, because it would spend our best-performing Note format at the product's weakest moment, and it kept everyone on the pitch list out of the beta, so an inaccurate score could not cost a real deal. It also adopted Substack comment mining as a lead channel: the comment sections of posts for new founders and writers going pro, where credible people describe their own stall in their own words.
Success was measured over 30 days, in this order: scorecard runs, scores shared publicly, discovery calls booked and one paid audit. Impressions were not on the list.
07
the client build system
I turned the build page into a repeatable system, so every client started from the same structure and only the substance changed.
The master page had nine parts before kickoff: a welcome, the build timeline and calls, how the page works, the client's details (prefilled, so they never researched for us), account access, material intake grouped by each deliverable, what they wanted from the build, how we would work together, and a notes space. A personalisation prompt filled it from the accepted proposal and client record, with a QA checklist before anything was shared. For Substack, access came by a team invitation to the studio, never the client's login.
After kickoff the page became a live delivery dashboard that answered four questions: where are we, what do you need from me, what has been delivered, and what happens next, with completed intake moved into a references section. A second prompt did that conversion.
Every call had two documents: client-facing call notes inside the build page, and an internal run sheet in the matching task card (roles, what the call must close, traps to avoid and a timed agenda).
08
the Notion dashboard
I built Atelier's operational backbone in Notion: departments, leads, clients, a content planner, deliverables and a time log, so the business runs on one system end to end.
09
the KPI scoreboard, the time log and the finances
At the end of every month, once a client build was done, we measured the business on the KPI scoreboard.
The north star: one Founding Build and one Signature Build signed a month, 2 to 3 active retainers near term, $4.25K a month in recurring revenue at steady state, 40% or more of discovery calls becoming clients, and a satisfaction score of 4.5 or higher. The scoreboard held a monthly table, the current priorities, a client pipeline (source, stage, deal value, hours and outcome) and a feedback log of testimonials.
The Deliverables and Time Log held every deliverable on every rung with its owner and the hours it really cost: sales and onboarding 8 hours, the Authority Audit 23, the Positioning System 42, and the Growth Retainer 35 a month. A full-climb client came to about 283 hours over seven months. When I modelled a target month against our capacity, it came out roughly 100 to 140 hours over, so the plan moved towards 2 to 3 retainers instead of 6.
I also kept detailed track of the finances of Home Abroad and Atelier, and of the financial goals for both.
10
the content planner
I built the Atelier Content Planner from scratch, a system for the full editorial cycle: plan, create, approve, publish, engage, learn. Atelier would have run two shared streams, Substack Notes and LinkedIn, with publishing slots alternating between the two founders. The monthly floor was 8 anchor Notes and 2 LinkedIn anchors, with up to 4 flex Notes that were allowed to be dropped, so visibility never ate the hours needed for outreach and client work. Every hook was scored out of six (specific, tense, relevant, unexpected, simple, honest) and had to reach four, a format never ran twice in a row, and a "receipt" of client proof only ran once the proof was approved. It also mapped the launch sequence onto the slots and set out who owned what. It was designed in full but never executed.
11
the batch week
Inside the planner, I designed the monthly Batch Week as five focused blocks, not five full production days: decide (lock one central argument and approve the anchors), draft the anchors, build the flex pieces selectively, a voice and proof pass, then pressure-test and queue. Scheduling could be automated; comments, replies, DMs and relationship-building stayed human. Not executed.
12
the engagement rhythm
I also built the engagement rhythm from scratch: how Atelier would stay visible in the right rooms between publishing days. Each day with no post had its own engagement run, with a checklist of a five-comment baseline, deeper relationship actions, connection requests, weekly goodwill, and logging each qualified signal with a follow-up date. The account owner always engaged from their own account, never the other founder's, and a short opener plus a five-message rhythm covered named leads. The test was simple: the right people return and answer in specific language, and a forced DM means the relationship still needs warming. Not executed.
13
the Practice, First Season and the Salon
We planned what happens after a build: the Atelier Practice, a method of four moves (publish, participate, connect, invite), First Season, a client's first 90 days of practising it, and the Atelier Salon, a paid community on Atelier's own Substack. It was planned in full and never launched.
At the heart of it was a bank of 12 challenges, one a week for a 90-day season, each a single doable action: name five writers you already know, leave three comments that add something, send one short warm ask, ask one well-matched writer to recommend you. The same bank would feed the build emails, the Salon challenges and every later season. Clients would get four short emails during the build and a Season Map at handover: what the client does, and what Atelier could carry for them.
Three phases: a free Atelier Substack from January 2027 (two essays a month, each adapted into a LinkedIn post), the paid Salon from February (a monthly insider lesson from real builds, a weekly challenge and a voluntary Restack Round), then extras held until April. Behind it sat one scheduled Claude task with three jobs (a weekly insight sweep with a privacy flag, a monthly batch and the build emails), with a person checking everything and publishing by hand.
It also reshaped the ladder. The Growth Retainer was not selling, because it sold upkeep the build had already covered, at about $24 an hour. The plan proposed retiring it for new clients in favour of Advisory, a Second Season at half the scope, and a fixed-term Launch Season for authors with a launch date. The 2027 ideas still open were more scorecard follow-up emails, corporate and brand Substack outreach, and a part-time operations teammate.
14
the admin
I ran almost all of the admin, owning the day-to-day operational and administrative load that keeps the studio running.
15
the first clients
We generated more than $5,000 in revenue in the first 3 months of starting Atelier from scratch. I worked with all three clients: Mikelann, Carolina and Bhakta. Mikelann bought the Founding Build for $1,000, paid in full (her build is below). More was in the pipeline at the time, including a first Signature Build worth $3,000. At kickoff, Mikelann had 95 subscribers on Substack (and about 2,700 stuck in Kajabi) and Carolina had 67. By mid-September they stood at 2,600+ and 433, nearly all of it their existing lists moved across. read the post on all three builds →
16
the post that told the story
"We Rebuilt 3 Substacks. Here's What Actually Worked." was published on Home Abroad's Substack on 24 September 2026. Atelier was introduced in it as "a Substack positioning studio", and the first three rebuilt publications were named: Bhakta Angelika Lorenz (a storyteller with two published books), Mikelann Valterra and Carolina Aponte. Before and after captures of each build were shown, along with Bhakta's Notion command centre, and six lessons from the work were set out: give the publication one sentence people can repeat, map what you already own, build a front door, make the visual identity earn its keep, choose a rhythm your real life can survive, and put the operation outside your head. Mikelann commented that she was delighted with her rebuilt Substack. read the post →
17
the landing page
I designed and built Atelier's landing page myself: a 12-section "ascent" up a painted cliffside maison, with a lantern-lit pricing climb, cursor-following light and page transitions built to read like a story, not a sales page. how I built it →
18
the Authority Scorecard
A free lead-gen tool I worked on that reads a visitor's actual Substack and scores it out of 100 across seven dimensions, graded by "The Editor", an AI persona. It fetches the publication server-side, weighs writing, visual identity and subscriber data, and hands back an honest, quotable read, with every run landing as a lead.
The flow: a prep screen, a gate, a short brief about you, then "show your work" (best posts, recent posts, the text of your last Notes, a welcome email and up to three screenshots). An optional subscriber CSV is read entirely in the browser, and only six aggregate numbers ever travel. Then five questions only the owner can answer, a waiting screen, and the result.
The scoring: seven dimensions, with positioning clarity and post quality worth 20 points each, cadence 15, the About page and welcome 13, proof and visual identity 12 each, and name and tagline 8. The Editor works to a strict brief: tell the owner what they cannot see, quote their own pages, never guess, and grade against what this owner uses Substack for, whether a business or craft and community. The server recalculates the total rather than trusting the model's arithmetic, so the score is always out of 100 across the dimensions that could be evidenced.
Beta testers reshaped it. A new "Taking Shape" band (40 to 59) split the one band that seven of eight real publications had landed in. A calibration rule stopped a coherent publication without a funnel scoring in the 30s. Two new purposes served writers who write for craft or community. And a clear "this takes up to 5 minutes" screen followed a tester who closed the tab and lost a read that was still running.
The result: an Authority Score and band, a card for each dimension (what is working and what is leaking), what the publication does for the owner's business, a "fix this first" move, three ranked actions, a shareable Authority ID card and a PDF of the full read, a receipt email carrying both, a WhatsApp button that opens with the score pre-filled, and a follow-up email three days later. A live read of Home Abroad's own Substack scored 64 out of 100 in about 46 seconds.
see the build →
