WORKING DRAFT
WORKING DRAFT
Worked example · Ledgerly is a fictitious company
GX Fellowship Memo
Prepared for
Jesse · Adam · Nak
Date
August 2026
Author
Sam Okonkwo
Cover Letter
For Jesse, Adam & Nak —
Ledgerly is an AI-native bookkeeping and monthly-close practice for US SMB owner-operators — the people running $1M–$50M businesses with messy books and no finance team.
The problem: their monthly books are a fire drill that blocks taxes, financing and payroll — because incumbents burn 8–12 human hours a client against a fixed fee and can’t afford to answer every lead, let alone close on time.
Our solution: the machine turns bank feeds, receipts, invoices and uncategorized transactions into a close-ready exception queue and drafts the reconciliations and financials — 80% machine, 20% accountant — cutting a close from ~10 hours to under 6.
Everyone wants it because the demand already exists and is already paid for at a settled ~$1,500/mo: CPA firms like Harbor & Vale turn away 20–30 monthly-close inquiries a month, and the advisors whose own work depends on clean books are lined up to send us clients.
The economics: clients pay $1,500/mo — blended to ~$1,800 with add-ons and multi-entity books — at a 73% margin and ~$1,100 monthly contribution once our controller signs in January; the path to $10M EBITDA is 1,100 clients at a ~$24M run-rate, a rounding error against the millions of US SMBs that already outsource their books.
We stand up the org in in-house accountant pods delivered through our pilot firm until the controller co-founder signs in January, fed by two channels — CPA-firm referrals and fractional-CFO and franchise referrers — with Sam on demand, Dana on delivery, and Priya on product.
It’s the right business for Sam because Ledgerly is the service he wished he had: years running finance and ops inside SMBs, watching the most expensive people do the least leveraged work while the close decided whether payroll was calm and the bank trusted you. In two years that’s a ~$6M business at ~$2M EBITDA; in five, ~$24M at ~$10M EBITDA.
— Sam Okonkwo · Founder, Ledgerly · August 2026
Where we are · August 2026
1
client live & paying, via the pilot firm
6
interested · 1 verbal commit
5.5 hrs
on v1 (4 bookkeeper + 1.5 CPA) vs ~10 baseline
73%
contribution margin, from January
12
clients to break even · Q4 2026 · ~$30k cash
3 signed
Marcus Reyes BD · Priya Nair product · Dana Osei CPA
Question 1
Ledgerly helps SMB owner-operators close their books every month, fast, without the chaos that messy bookkeeping creates for them, their CPA, and their lender.
Category. AI-native bookkeeping and monthly-close practice, capturing existing spend by delivering with AI where incumbents bill human hours for data entry, categorization and reconciliation against a fixed monthly fee.
—
Core close: bank and credit-card feeds, receipts, invoices, statements, reconciliations, draft financials.
—
Customer profile: US SMBs with messy monthly transactions, multiple accounts, and no full-time finance team.
—
Add-ons: sales-tax support, payroll coordination, annual tax prep handoff, advisory and fractional-CFO work once the close is trusted.
Question 2
ICP
—
Who: owner-operators, and they are their own decision-maker, so the sales cycle is weeks, not quarters.
—
Business: $1M to $50M revenue, 10 to 100 employees.
—
Where: US services, trades, franchises, clinics, agencies and multi-location operators first.
—
Trigger: the moment they realize their books are blocking taxes, financing, payroll, or basic operating decisions.
Knowledge edge
Bookkeeping is the obstacle, not the goal. Their main goal is to run the business.
—
The process today is slow, mistake-prone, unpleasant and painful.
—
This buyer would pay more for clean books, fewer questions, and a monthly close that does not demand their time. I know because I ran finance and ops inside SMBs and lived inside the mess.
Distribution edge
10
active referrers today
5/mo
committed floor from Aug
15–20
referrers signed by Dec
Source: Marcus Reyes, recorded voice note, July 31; commitments on file.
—
Personal network plus Marcus’s advisor network. These advisors only keep trust if they solve the finance mess around the owner. It’s literally all they do.
—
Newest proof, August 1: Harbor & Vale CPAs, a regional tax firm with no appetite for monthly bookkeeping, turns away 20 to 30 monthly-close inquiries a month. Managing partner saw the demo and liked it; product test and security review next; follow-up 8/17. Marcus is landing the referral partnership.
ICP worksheet, completed on the GX template.
Question 3
Two reasons this doesn’t exist already
—
01 The technology didn’t exist, until now. Digesting a month of messy bank feeds, receipts, invoices, statements, uncategorized transactions and entity sprawl only became automatable on current models, at current prices.
—
02 Wrong economics. At 8 to 12 human hours per client per month against a fixed fee, an incumbent can’t afford to answer every lead, let alone build a better close. Example: clean monthly books for sub-$5M companies. The need obviously exists, yet good CPA firms punt it because it is too much work and the economics do not support it.
Where roughly 10 hours go on a messy monthly close · 15+ recorded practitioner conversations
| Where the hours go | Hours |
|---|---|
| Judgment and review | 4.5 hrs |
| Categorization, reconciliation, collection and draft financials | 5.5 hrs |
Half the work is judgment a CPA or senior bookkeeper should be doing. The other half is assembly a CPA should never have been doing. That half is the business.
| Alternative | Why it fails this customer |
|---|---|
| Solo bookkeepers and boutique firms | Capacity-capped, slow to respond, dependent on one person’s calendar; the close still slips. |
| Papertrail Books | Human-heavy monthly service; good operators but no structural cost advantage. |
| CloseKit | Receipt and categorization tooling; does not own the close or sign off the books. |
| ReconcileHQ | Bank-feed matching with outsourced reviewers; a competitor’s own operator said, “too many exceptions… the client still has to explain everything.” |
| LedgerDesk AI (+$249/user/mo) | Pilot firm demoed it: “useful for search, still not a close.” |
The gap over time
—
Intuit, payroll platforms and practice-management vendors sell software to the firms we compete with, not against us. The real gap-closer is incumbents getting better at AI for this use case, on a clock of roughly 18 to 36 months.
—
By then: trust in our corridors, referral networks compounding, and my personal brand, built in public, around SMB operators, CPAs and franchise systems.
—
Many winners: the incumbent shape is a $3M bookkeeping firm with five senior people and 20 offshore staff, so a number two and three still make real money.
Question 4
Why now
—
AI crossed practitioner quality. Models categorize transactions, match receipts, draft explanations, reconcile accounts and produce exception queues a human can actually review. This was impossible 24 months ago. Work that used to require an offshore team and three follow-up emails now costs a few dollars in API calls.
—
Rising compliance load is meeting worse books. Beneficial-ownership reporting, 1099-K changes, state sales-tax nexus and revenue-recognition complexity make DIY harder every year. More owners are running into a stricter, less forgiving operating environment, and every year a clean monthly close is worth more.
How big
~1,100 active clients is not millions of SMBs. At maturity, one client is worth a $1,800 blended monthly close, so I need about 1,100.
1,100
active clients
~$1,800
blended monthly fee
~$24M
run rate · ~40% EBITDA ≈ $10M
—
Price · add-ons make a client worth more. The base close is $1,500/mo. Add-on services and multi-entity books lift the blended fee to ~$1,800/mo across the book.
—
Volume · entities make more closes. A client with three LLCs, two locations or separate operating and real-estate entities is not three new customers, but it is more work and more revenue. Businesses do not stay simple as they grow.
1,100 clients × ~$1,800/mo = ~$2M a month. That is the whole model, and 1,100 clients is a rounding error against the millions of US SMBs that already outsource bookkeeping.
Both levers priced from our pilot firm’s rate card, and quoted live on a real sales call on August 2. Annual tax, advisory and fractional-CFO attach are not in this math. They compound on top: that’s Question 7b.
Question 5
Live URL
ledgerly.up.railway.app
demo@ledgerly.example
Password
GX_FAKE_DEMO_0826
Walkthrough
2-minute video
A shared demonstration workspace seeded with a fictitious messy client month.
~10 hrs
baseline per monthly close
5–6 hrs
close one, logged, on v1
80 / 20
machine / accountant by design
What the machine does
—
Turns bank and credit-card feeds, receipts, invoices, statements and uncategorized transactions across accounts into a close-ready exception queue. Every practitioner we spoke to named this as the biggest time sink.
—
Drafts reconciliations and financials, and chases the client for whatever is still missing.
What stays human, permanently
Adjusting entries, accruals, judgment classification calls, review and sign-off. On screen, those gates will not turn green on their own. The bookkeeper or CPA is more of a checker than a data-entry worker.
How the ratio moves next
—
The reconciliation and draft-financials generator kills hours of offshore review per client per month.
—
QuickBooks, Xero and bank-feed integrations, which our pilot firm thinks could “maybe double the client capacity here.”
Question 6
Two demand sides: users, the bookkeepers and accountants who run the books, and end customers, the SMB owner-operators who pay.
Millions
of US SMBs outsource books
$1,500/mo
base close
~$24M
run-rate target, wedge category
Sources: pilot-firm rate card; partner conversations; demand plan.
—
The demand exists and is already being paid for. The price is close to a market constant, about $1,500/mo for the wedge customer.
—
The pricing conversation that determines our economics happens on the bookkeeping-partner side.
Two primary demand channels
—
01 Existing CPA-firm referrals.Harbor & Vale CPAs is a regional tax firm with no monthly bookkeeping bench. It gets 20 to 30 monthly-close inquiries a month and turns most of them away. The managing partner has seen the demo; product test and security review are next. That shape is not just one firm. Every tax practice serving SMBs has the same door in it, and the pitch that works on Harbor & Vale is the pitch that works on the next one.
—
02 Fractional-CFO, SMB-advisory and franchise-consultant referrers. Marcus’s network, plus mine. These advisors only keep the client if the numbers are reliable, so getting the books cleaned up is upstream of their own work. Ten active today, 15 to 20 signed by December, at 1 to 2 clients each a month.
6a · Users
20+
practitioner conversations in 45 days
10
recorded and on file · rest in person
Every single one said a variation of the same thing: transaction cleanup, document collection, reconciliation, and explaining exceptions to the client. That’s it.
Agreed with Dana: $850/mo to her firm, $550/mo contribution to us after machine delivery, her share coming down as the product improves and she comes on full time.
“Taking what the client gives us, understanding what clears, what does not, and where the missing support is. That’s the biggest time suck.”
Maya Chen · Partner, Northline CPAs
Unprompted, on pricing: “I’d gladly pay $600+ per client if the close came in clean.”
“If we can take away the transaction cleanup, we can run the client for maybe 60% of the fee instead of 100%.”
Nina Patel · Pilot firm bookkeeper
Her own counter-offer, against a 40% proposal.
“This is excellent. I can handle twice as many clients if this works well.”
Dana Osei · CPA · pilot firm principal, using it on a live client
Also: “more useful than LedgerDesk or ReconcileHQ.” Her complaint: “some issues with uncategorized transfers that the model made… so we needed more time to review,” caught by the accountant gate before close. Her asks: QuickBooks integration, client reminders, and better draft financials.
“With my overhead. No, that won’t work, we lose money.”
Talia Brooks · Ridgewater Back Office, competitor operator
Hardest objection. Restructured live to the owned-service and equity model; her answer became “maybe it works, if I own part of the tech company.” That objection is why Ledgerly owns the client relationship instead of selling software to bookkeepers.
6b · End customers
6
interested
1
verbal commit
1
paying
One close completed and paid, delivered through Dana. One close to signing, one on a follow-up call, three at first conversation.
What a real one sounds like · Malik, August 2
—
Owner of a three-location home-services business, ~$2.4M revenue.
—
Two LLCs today; third entity planned for a new crew and equipment lease. One lead, multiple sets of books.
“Isn’t that high for bookkeeping? I have a person who says she can do it for $600 a month.”
Malik · August 2
This objection is domestic and price-anchor specific. Malik is pricing us against a solo bookkeeper who closes late and does not reconcile all accounts. His alternative is not a high-quality CPA firm, it’s a cheaper person who absorbs his chaos until tax season. No franchise or multi-entity client has raised it the same way.
Raw transcripts on file, available on request.
Question 7
One client, today
| Line | Amount |
|---|---|
| Client pays | $1,500/mo |
| Dana’s firm, bookkeeping work | $850 |
| Ledgerly books | $650 |
| Delivery, all in: inference, OCR, bank-feed ingestion, retrieval, eval and retry | ~$100 |
| Contribution | ~$550/mo |
The bridge is the proof: a CPA with years of experience agreed to take above-market share on v1, because the machine genuinely alleviates her workload. Her share comes down as the product does more.
From January the in-house controller co-founder signs and we book the full $1,500.
One client, from January
| Line | Amount |
|---|---|
| Client pays | $1,500/mo |
| In-house pod, at full ramp: ~$18k/mo across ~60 clients | ~$300 |
| Delivery, machine | ~$100 |
| Contribution | ~$1,100/mo |
| After ~$1,200 of CAC, paid back in ~2 months | ~$1,100/mo ongoing |
One client, not one unit
| Rate card, quoted live on August 2 | Price |
|---|---|
| Base monthly close | $1,500/mo |
| Add-on services: payroll coordination, sales-tax support, annual tax handoff | +$200–$600/mo |
| Multi-entity books | priced per entity |
A client at maturity is a ~$1,800 monthly blend before any annual tax or advisory attach.
CAC is modelled at ~$1,200 per client, but paid once, on the first close. Additional entities and services arrive with near-zero incremental CAC, because we already hold the books.
Question 7b
01 Net revenue retention and expansion, from the books we hold
—
Clients add entities, locations, payroll complexity and services over time.
—
The books are already live, so expansion is cheaper to win than the first close.
02 Referrals, from the network
—
Partner firms send the monthly-close work they can’t serve.
—
Each signed firm is a standing stream of clients, not a one-off.
03 Annual tax and advisory, from the category we unlock
—
The clean close creates the right to do tax prep, planning, forecasting and fractional-CFO work.
—
Brings higher-value finance work into the book, not just bookkeeping.
Question 8
—
The wedge: the monthly close at the settled market price of $1,500/mo, delivered through the pilot firm while the bridge structure holds.
—
Time to value: the client signs and their books start moving the same day; document request inside 48 hours; value in weeks, not quarters.
—
The proof period pays for itself: break-even about 12 active clients, we cross it in Q4 2026, and H2 lands at roughly break-even before any investment.
01
Jan 27 · Controller signs
Contribution steps from ~$550 to ~$1,100/mo per client.
02
Q1 27 · $1M run-rate pod
One ramping pod, already EBITDA positive.
03
Q3 27 · Pod one at scale
Margin holds as add-ons lift blended ARPU toward ~$1,800/mo.
04
2028 · Multiply
~275 clients, ~$6M revenue, ~33% EBITDA ≈ $2M.
05
2029 · Scale
Pods six through twelve. ~$24M run rate at roughly 40%.
Question 9
My whole working life has been a story of messy books inside real operating companies.
—
I ran finance and ops inside SMBs where the close was never just an accounting task. It decided whether payroll was calm, whether the bank trusted us, whether the owner could sleep, and whether tax season became a fire drill.
—
I have been the person building the spreadsheet because the books were late, chasing receipts from a foreman on a Friday, explaining to an owner why cash was down while revenue was up, and translating a chart of accounts no one trusted.
—
The pattern was always the same: smart operators, decent accountants, bad system. The human judgment mattered. The human data entry did not. The close failed because the most expensive people were doing the least leveraged work.
—
Ledgerly is literally the service I wish I had inside those companies: clean the feeds, chase the missing support, reconcile the accounts, surface the judgment calls, and let a serious accountant sign off. I wrote that spec as an operator before I wrote it as a founder.
—
I’m still living it. The first customers are people whose businesses look like the ones I operated: thin admin teams, real revenue, multiple accounts, and owners who know the books matter but cannot afford a full finance department.
Team by archetype
| Archetype | Who | Status |
|---|---|---|
| CEO / sales · demand | Sam and Marcus | Marcus full time Jan 1, terms agreed |
| Product / delivery · value | Sam, Priya, Dana | Dana on a co-founder track |
| Ops / finance · leverage | The honest gap | First hire post-raise |
Followership
“Sam, I know this customer and I know this channel. The firms are already saying no to this work. If we are disciplined, we can win.”
Marcus Reyes · BD · advisor now, full time Jan 1
“This is excellent. I can handle twice as many clients if this works well. The interface is simple.”
Dana Osei · CPA · pilot firm principal
“I like that you are building the service and the system together. The workflow has a real user.”
Priya Nair · Product · ex-fintech, vetted by Adam
Neither Dana nor Marcus joined for the money, because there isn’t any yet. They joined because they think this works, and my job for the next six months is to make sure they were right.
—
Marcus ran partnerships inside a finance-services firm, left, joined as an advisor, and pushed his own start date up once the first CPA-firm conversations converted. Day one of his advisory produced the Harbor & Vale managing-partner call.
—
Contract first, convert on evidence: everyone starts contracted against defined scorecards, and co-founder equity follows demonstrated fit.
Iteration receipts
| What I heard | What I changed |
|---|---|
| “The money is in fractional CFO, not bookkeeping.” Early advisor conversations | Did not start with CFO. Started with the close, because no advisory work is trusted when the books are wrong. |
| “The exception queue. That’s the biggest time suck.” Maya Chen, Northline CPAs | Scrapped the dashboard-first build. Rebuilt around reconciliations and missing-document chase; that is what runs live clients today. |
Self-reflection
| Trait | Reflection |
|---|---|
| Compass points north | I would not sell a monthly close I would not have trusted when I was running ops. |
| Authentic salesman | When Ridgewater told me my price would lose her money, I restructured the deal honestly in the same conversation, and “that won’t work” became “maybe it works.” |
| Action-oriented | Idea on June 18th. Six weeks later: working product, a firm running a live client, a paying customer, two hires signed and dozens of conversations on record. |
| Doesn’t suffer fools | Marcus has more channel experience than I do, and I still asked for a written plan and ramp before I was sure I could pay him. |
| Competitive | I got a competitor’s own operator to tell me on tape where their product breaks, and then asked her to join us. |
| Relationship-driven | Most of the twenty-plus practitioner conversations happened in person, over coffee with CPAs and operators. |
| Admits unknowns | Ops and finance leadership is the honest gap in my team, and the only real paid data in this market sits at my model’s downside case, not its base case. |
| Takes feedback | Jesse told me to hold a high bar on the demand plan, so it stopped being a document and became a model you can change the inputs on. |
Question 10
Jesse · GTM and leadership
—
GTM strategy and experimentation across CPA-firm referrals, advisor referrers, organic and paid.
—
Introductions in two waves: SMB marketers and franchise advisors now, larger tax and advisory firms when the annual-tax attach launches.
—
Leadership coaching, the way you’ve already done it with me on Marcus, and Spencer before that. I’m managing people older and more experienced than me, and it has already helped a lot.
Adam · product
—
The discipline of turning practitioner feedback into product decisions. Dana’s list from August 1: QuickBooks and Xero integration, automated client reminders, better draft financials. I can build all of it; what I need is the product sense of what to build first and what to refuse.
—
Guidance on the coming compliance and data storage work we discussed.
Nak · the operating machine
—
Data hygiene and scorecards: Marcus’s five-month scorecard and the funnel instrumentation exist; I want your standard applied, and a reporting cadence for referrers, partner firms and marketing.
—
The GM hire, my version of Nak: help me design the role, source it, and vet it.
We make a great team. Every ask above is something you’ve already done once for me informally, and it has been very helpful.
Attached: demand plan · client-book worksheet · live demo · 2-minute walkthrough.