Bottleneck. Theory of constraint. The smallest set of changes that moves the multiple — ranked by leverage, not effort. If we hadn't found them, this document would say so, and we would leave. Engagement scope, if any, is quoted from this document and nothing else.
PRESENTING COMPLAINT — IN THE OWNER'S WORDS, AT INTAKE
"Growth has flattened at twelve locations. Marketing spend is up 22% year over year, new patients are flat, and the board no longer trusts the marketing slide. We are two years from a process we'd like to run with a banker."
| VITAL | DAY-0 READING | CONFIDENCE | READS AS |
|---|---|---|---|
| New patients / mo | 241 | HIGH | flat — demand reaching phones grew 11% YoY |
| Revenue (collections) / mo | $1.31M | HIGH | steady |
| NP acquisition cost, fully loaded | $246 | MODERATE | 2.1× what the board decks carry |
| Attribution coverage | 2 of 12 channels | HIGH | spend is steered blind |
| Answer rate, per desk | 78% blended · 68–96% range | HIGH | desk variance exceeds channel variance |
| Booking rate, per desk | 34% blended · 22–44% range | MODERATE | same pattern, same desks |
| No-show rate | 16% | LOW | real, but the instrument needs a quarter of clean data first |
FULL METHODS AND FLAGS IN THE BASELINE READING. NOTHING IN THIS DOCUMENT USES A NUMBER THAT DOCUMENT CANNOT DEFEND.
THE CONSTRAINT — SINGULAR, BY POLICY
The system's binding constraint is intake: demand already paid for is lost between the ring and the booking, so every dollar added above the leak buys less than the last one.
Theory of constraint: until the bottom-half desks answer and book at the top desk's demonstrated rate — a rate this network already achieves at North-04 — added acquisition spend subsidizes the leak. Attribution blindness is real but subordinate: it hides the leak, it does not cause it. One constraint is named because a diagnosis with two constraints is a to-do list.
THE SMALLEST SET OF CHANGES THAT MOVES THE MULTIPLE — RANKED BY LEVERAGE, NOT EFFORT
| CHANGE | LEVERAGE (I×C÷E) | EXPECTED MOVEMENT | COST / MO | SEQ | RX |
|---|---|---|---|---|---|
| Instrument every desk on the intake rubric; weekly per-desk readout reviewed with the people who took the calls | 21.3 | booking rate 34% → 40%+ in 90 days (~14 NP/mo) | $4K | 1 | Rx 03 |
| Overflow + after-hours routing to staffed desks; coverage for the two known lunch gaps | 16.0 | answer rate 78% → 90%+ (~11 NP/mo now going to voicemail) | $6K | 2 | Rx 03 |
| Attribution build-out: 2 → 10+ channels traceable click-to-chair | 8.4 | no direct NP gain — makes change № 4 safe to execute | $5K | 3 | Rx 02/04 |
| Reallocate spend by fully-loaded CAC per channel; zero-spend the untraceable bottom third | 6.7 | fully-loaded CAC $246 → ~$200 at flat volume | $0 net | 4 | Rx 01 |
| No-show instrument repair: one status definition across 12 locations, then reminder-cadence pilot | 3.1 | honest read first; movement priced after one clean quarter | $1K | 5 | Rx 03/04 |
LEVERAGE = 12-MO MARGIN IMPACT ($100K UNITS) × CONFIDENCE (0–1) ÷ EFFORT (INTERNAL DISRUPTION 1–5). RUBRIC: DIAGNOSIS-SOP.MD. FIVE CHANGES, NOT FIFTEEN — SMALLEST SET, BY POLICY. NOTHING HERE REQUIRES A SITE REBUILD OR A REBRAND, SO NONE IS PRESCRIBED.
Two paths, written with equal care. The no-go is not a failure state — it is the outcome that proves the go means something. Your call, made with the numbers in front of you.
PATH A — GO · OPERATE (RX 03)
If the numbers don't move by retest, the quarterly review will say so in our handwriting before you have to ask.
PATH B — NO-GO · WE SAY SO AND WE LEAVE
A no-go delivered well is the cheapest decision this document can hand you. We have written them before and will again.
RECOMMENDATION — STATED, NOT HEDGED
GO ☒NO-GO ☐ The constraint is movable, the network's own best desk proves the target rate is achievable in this payer mix, and the model clears the hurdle at conservative confidence. We recommend Path A, sequence as ranked. If change № 1 hasn't moved booking rate by day 90, treat that as the retest come early — and reread Path B.
WHAT WE WILL NOT DO — STRUCK BY POLICY, NOT BY OVERSIGHT
Anything on this list that a future vendor proposes should be asked to show the constraint it moves.
— specimen —
DIAGNOSING PARTNER, ADMEN
— specimen —
REVIEW GATE — SECOND PARTNER, PRE-DELIVERY
— specimen —
DECISION — CLIENT CFO / OWNER · GO / NO-GO · DATE