Director of Operations · 19 September 2026

Four counters first. Then we hold.

15 October 2026 · 26 days. The board asked for automation, a different ICP, and a CTA. These are the machines that actually exist in 2026, with hours and capital. Defense is the second half. Adjustments are a vote, not a speech.

Fieldstone Advisory

Montgomery County, Texas · Addendum to the board packet · not a replacement

AI drafts. Human review. Then send. We are not a CPA firm. We are not a law firm. Circular 230 did not get an AI exception.

Self-eval

The offer is strong. The machine is the windshield.

The offer is a specialist practice. The go-to-market is still a windshield. The board is not wrong about compounding. They are wrong about which machine to buy.

Product 8 / 10
Go-to-market 4 / 10
Capital efficiency 9 / 10 vs the four counters
Capacity honesty 9 / 10 — the constraint is named

2026 stack

What lead-gen actually costs

Item	Published range
Bookkeeping Google CPC	$8 mid · $5–$18 working · $10–$30 metros
Bookkeeping Search CPL	$50–$100 agency · $120–$280 CPA-firm blogs
Fully-loaded mill CAC	$800–$2,000 (media is ½–⅓ of that)
Meta consult CPL (accounting)	$40–$120 · guides $8–$30 (cheap, soft)
GHL Starter	$97/mo. Unlimited $297 is agency vanity.
Google min spend to learn	$1,500/mo floor · $2,500–$5,000 predictable
LSA Tax Specialist	$30–$80 CPL. Bookkeeping LSA eligibility is disputed.
Bench 2026 list	Grow $199 · Core $399 · Core+Tax $599

Counters

What a reasonable DoO would table

Counter A

Texas Mill Search

Texas LLC / S-corp, $200k–$1.5M, searched “bookkeeping services Houston.” Not retainage.

$349/mo Core · $499 Plus · $199 onboard. Card on the call. GHL calendar CTA.

Capital 90d · $10,700 base / 90d (Search $6k + LSA $3k + GHL)
Hours · 14–18 wk 1–2, then 18–25. Speed-to-lead is a day job.
CAC · Media $450–$1,100 · loaded $800–$1,800 · first card 7–35 days
Cash fail / base / best · −$10.2k / −$5.6k / +$5.8k · exit $1,750 base
Breaks · 10 hours CT = 3–5 mill clients. Hire at #6 or the ads buy a queue you cannot close.
Why the board likes it · Same product Bench sells, $50 under Core, four metros, card not AR. Ads pause if it fails.

Kill as a practice strategy. It is a 4-client mill with a Google bill.

Counter B

Bookkeeper Factory

National career-switcher. Free workshop or $97 challenge → $2,499 course. 17k BKL grads already exist.

Meta/YouTube → Stripe. Not a bookkeeping firm. A media company.

Capital 90d · $18–28k / 90d. Creative dies in 2–5 weeks without a weekly show.
Hours · 18–28/wk. Cannot run this and a book of files.
CAC · $800–$4,000 per $2,499 buyer · first $97 in 2–14 days
Cash fail / base / best · −$14k / +$7k / +$27k · exit $0 (info product)
Breaks · FTC earnings claims, refunds 10–20%, CPA-referral reputation, NAICS code change.
Why the board likes it · Only 90-day cash-positive machine. Exits hours-for-dollars. Different company.

Second entity or no. Do not clone Launch inside Fieldstone.

Counter C

Service Line P&L

Independent Texas restaurant, Toast/Square, $800k–$4M. Houston led US city closures H1 2026.

$899/mo one location · $1,500 setup. Diagnostic screenshare, then ACH.

Capital 90d · $12,600 base (Search $7.5k + Meta $3k). Toast partner bounty $500, not a model.
Hours · 16–22/wk, ugly hours. One kitchen is 3–5 hrs/week. 10 hours CT ≈ 2 restaurants.
CAC · Blended $400–$900 · first deposit 10–30 days
Cash fail / base / best · −$10.2k / −$2.0k / +$9.4k · exit $3,600 base
Breaks · Churn is a feature (1,039 TX closures H1). Weekly Tuesday fire drills. Credit risk.
Why the board likes it · Best ACV of the service counters. POS is a channel Google cannot auction.

Keep as a named SKU. Do not make it primary. One prepaid seat, not ads.

Counter D

Catch-Up Calculator

US SMB 6–24 months dark. Loan, IRS, CPA refused the file. National remote. Stripe 50/50.

$1,490 / $2,900 / $4,900 SKUs. Instant quote. Card before the pack leaves.

Capital 90d · $15,100 base (Search $7.5k + Meta $3k + Clay/Instantly).
Hours · Sales 6–10. Production is the iceberg. 12-month medium file = 15–30 hours.
CAC · Media $250–$800 per deposit · first card 24–72 hours
Cash fail / base / best · −$12k / ~$0 / +$16k (WIP bomb) · exit $1.2–3.2k attach
Breaks · Sells past capacity in week 3. 16 files × 20 hrs = 320 hrs vs 120 hrs of CT. Unearned revenue + angry CPAs.
Why the board likes it · Fastest cash. Productized. The board’s “take the card” counter.

We already have the widget on Quote. Do not buy $15k of national ads. Cap WIP at two named Texas files.

Other ICPs

Who a CTA can actually close

ICP	CTA?	Vs our floor
1040 mill / DIY extension4.1 as a 4-week cash product · 1.5 as a practice	Yes, gated, card	Inferior LTV · superior cash this fortnight
Woodlands / Conroe restaurant3.6	No. Toast dump, then invoice.	At/under $1,200. Houston #1 closures. Credit risk.
Dental / medical GP3.3	No. 2–6 week sit-down.	Only ICP that clears $1,500 without apology. Q1 play.
Shopify / Amazon $500k–2M2.8	Yes at $299. Then you drown.	They want 1/5 of the floor. Q4 txn explosion.
STR / Schedule E2.6	Books maybe. Cost-seg no.	Stessa is $0. Do not sell cost-seg lite.
BKL wholesale1.7	Yes, and that is the trap.	Friday Close at 13¢ on the dollar.

Defense

After the counters, not instead of them

The binding constraint is ten weekly hours, Central Time. Every counter that “works” sells past that in 90 days. Ads do not create capacity. They create WIP.
A mill client at $349 consumes 1.5–6 hours/month. Ten hours holds 3–5 mill files — $1,047–$1,745 MRR — less than one Friday Close. You will have spent $11k to get there.
Media CAC $800–$2,000 to buy a stranger at Bench’s price. Transition $1,000 from a name you already have is $0 media. The board is comparing a windshield to a machine. Compare cash collected per steward-hour.
Houston construction added 11,200 jobs Feb 2025–Feb 2026, #1 US metro. Starts +6% in 2026. “Switch ICPs because trades are dead” is a bad reason. Switch only if the sales motion cannot produce a deposit this month.
Bookkeeping is not a reliable LSA category. Tax Specialist is. The funnel the board is picturing (Google badge, phone, $349) may not even be eligible.
The Quote desk is already the catch-up calculator. Funding $15k of national Search to fill it without a two-file cap is how a boutique becomes a refund shop.
A course mill can be cash-positive in 90 days. It is not Fieldstone. It poisons the CPAs who send K-1 overflow. Second entity or no.

Our 90 days, no ad buy

Capital $0–$400 (domain + GBP). No ad buy.

10 CT, named. ABC Supply, five warm texts, two Extension, one wrap if real.

Fail $0 if the names are cold and no 4868 is in reach. Base $1,000–$2,900 (one Transition + one Extension). Best $8,500–$10,400 (one wrap + two Extension + one Transition).

Fail is cheaper than Mill Search fail (−$10k) and does not create WIP you cannot finish.

Worth asking

Adjust the rail. Do not sell the mill.

Adjust

Should Extension File take a card?

Yes. Different ICP, 26 days to the wall, DIY overflow will not wait on an invoice. 50% Stripe to start, documents in seven days or the file dies. Cap two. Invoice remains the rail on Friday Close.

Adjust the rail, not the product

Is the Quote desk the CTA the board asked for?

Yes. Volume, accounts, AR/AP, payroll, ProAxis 0–12, rush windows. What it does not do is take the card. Do not buy GoHighLevel to rebuild it. Put a “invoice after the call” line on the result, and a Stripe link only on Extension.

Hold

Do we spend $2,500/mo on “bookkeeping services”?

No. That keyword buys the mill buyer. CPC $5–$14, they want $199–$399. We lose the auction and the file.

Named seat only

Unpause Kitchen because restaurants need books?

Need is not credit. Houston led US city closures H1 2026. One prepaid Kitchen Start on a named owner is a test. Mixed-beverage SEO stays later.

Park

Is dental the better $1,500 ICP?

Economically yes. Funnel-wise no. Sit-down, 2–6 weeks, HIPAA, Dentrix. One study-club breakfast in Q1 2027. Not a September ad.

Hold

Should we white-label for BKL grads?

No. Sell them a Transition seat for their own file, or take the end client at full freight.

Reallocate hours, not capital

What if the names at ABC Supply are not warm?

Then the windshield fails this week, and the honest move is CPA-overflow texts for Extension — not a $11k mill. Two hours of the ten, five named CPAs, K-1 overflow, card on.

Vote

Six motions. Staple behind the board packet.

Motion 1. Hold Friday Close $1,500 / founding $6,500 / Transition $1,000. Invoice. No card on that desk.
Motion 2. Adjust: Extension File takes card, 50% to start, cap two, drop-dead 29 Sep. Different rail.
Motion 3. Do not fund Texas Mill Search ($11k) or Catch-Up national ads ($15k) or Bookkeeper Factory ($18–28k).
Motion 4. Quote is the CTA. Do not buy GHL to clone it. Stripe only on Extension this fortnight.
Motion 5. Kitchen: one prepaid named seat if the owner exists. No ads. No Toast blog this fortnight.
Motion 6. BKL clone is a second entity or no. Workshop $1,500 stays teaching, not a $97 challenge.

Director of Operations — recommend

Officer — implement