Moves vs Multiple

RX 05.2 · REVISION 06 · MONTH 22 · 14 OCT 2024 · 36 LOCATIONS (42 PROJECTED AT PROCESS OPEN) · SOUTHEAST · SPECIMEN

EXIT CLOCK — 9 MONTHS TO PROCESS OPEN · TARGET Q3 '25 · RE-CUT QUARTERLY WITH CEO + CFO · 60 MIN · SUPERSEDED REVISIONS STAY IN THE FILE

CONFIDENTIAL · RX 05FORM 884-X

Which operational moves actually move the multiple — ranked by leverage, costed, and sequenced against the clock. Multiples are made two years before the banker calls; this revision is written with nine months left, and the clock is the reason two of the moves on this page are declined.

THE DISCIPLINE THIS DOCUMENT RUNS ON — READ IT BEFORE THE TABLES

Moves that add EBITDA are arithmetic. Moves that add multiple are arguments. We model the first and we prepare the second, and we never blend them into a single number, because blending them is how a growth model becomes a wish. Sheet 2 is the arithmetic: dollars of EBITDA, confidence-weighted, costed, dated. Sheet 3 is the arguments: what a buyer would have to accept, what evidence supports it, and what kills it. Sheet 3 carries no dollar figures at all.

Every enterprise-value figure in this document is EBITDA × a stated multiple, and the multiple is an input we do not control. Where a figure moves because the multiple moved rather than because the business did, we say so.

THE ONLY PUBLISHED BANDS THAT EXIST — AND WHERE THIS GROUP ACTUALLY SITS

NO PUBLISHED BAND EXISTS BETWEEN 6× AND 9× a group of 24–42 locations sits here, and no one publishes a number for it MODELLED 8.4× → 8.9× a model, not a comparable — and not credited to any move below 10× 11× 12× 13× 14× 15× 16× PUBLISHED · 5–6× INDIVIDUAL PRACTICES & ADD-ONS PUBLISHED · 9–10× LARGE DSO PLATFORMS 13–16× · THE PEAK PLATFORMS ARE REPORTED AS DOWN FROM

SOURCE — BRIAN COLAO, DYKEMA DSO INDUSTRY GROUP, VIA GROUP DENTISTRY NOW. WHAT THE BANDS COVER: US DENTAL, EBITDA MULTIPLES, TWO SIZE CLASSES. WHAT THEY DO NOT COVER: THE DEFINITION OF ADJUSTED EBITDA BEHIND THEM, ANY GROUP BETWEEN THE TWO CLASSES, MULTI-SPECIALTY OR MEDICAL, AND ANY DATE-STAMPED SERIES — SO "DOWN FROM 13–16×" CANNOT BE RE-DERIVED FROM PUBLISHED DATA. THEY ARE A PRACTITIONER'S REPORTED READ OF A TRANSACTION MARKET, NOT AN INDEX OF CLOSED TRANSACTIONS. EVERYTHING BETWEEN 6× AND 9× ON THIS PAGE IS A MODEL.

THE BASE — TRAILING TWELVE MONTHS TO MONTH 22 · CLIENT'S OWN LEDGER · CONTROLLER-SIGNED IN THE MONTH-22 BOARD PACK (FORM 610-M)
LINETTM TO MONTH 22PER LOCATIONDERIVATION / SOURCE
Gross production$61.5M$1.708MPMS · pre-GL
Contractual adjustments−$9.1M · 14.8%−$0.253MGL 4100 · 9.1 ÷ 61.5
Collections$52.4M$1.456MGL 4000 · 61.5 − 9.1
ADJUSTED EBITDA · 17.6% OF COLLECTIONS$9.2M$255.6KPER CFO · 9.2 ÷ 52.4
Annualised EBITDA per +1 new patient / month$6,758$971 collected value × 58% contribution × 12

36 LOCATIONS AT THIS REVISION · 42 PROJECTED AT PROCESS OPEN · 12 OF THEM DE NOVO AND STILL RAMPING, WHICH IS WHY EBITDA PER LOCATION IS BELOW WHAT THE MATURE BASE PRODUCES. THE $6,758 CONSTANT CONVERTS EVERY VOLUME MOVE ON SHEET 2 INTO DOLLARS AND IS THE CLIENT'S OWN NUMBER — GENERAL DENTISTRY PUBLISHES NO PATIENT-VALUE OR ACQUISITION-COST BENCHMARK, SO NONE IS USED.

HOW WRONG THIS MODEL HAS BEEN ON THIS ACCOUNT — FIVE MOVES ALREADY EXECUTED, REVISIONS 01–05 · ANNUALISED EBITDA
EXECUTED MOVEREV · MONTHMODELLEDDELIVEREDΔWHY
Desk instrumentation, three cohort regions01 · m7$410K$455K▲ 11.0%the leak was larger than the sit-week sample implied
Attribution build-out, 2 → 11 channels02 · m10$0$0enablerno direct EBITDA claimed then or now — it carries the pool the reallocation move spends from
Location-page template roll, 16 sites03 · m13$265K$172K▼ 35.1% — missthe hub kept outranking its own location pages; the template converted better and received less traffic than modelled
Broken-appointment deposit pilot04 · m16$190K$61K▼ 67.9% — misstwo regions declined the policy outright; it ran at 6 of 16 cohort sites. Adoption, not efficacy.
Dormant-patient reactivation programme05 · m19$145K$158K▲ 9.0%recall list was deeper than the PMS flag suggested
TOTAL — FIVE EXECUTED MOVES$1,010K$846K▼ 16.2%83.8% REALISATION

THIS TABLE IS PRINTED FIRST, BEFORE ANY FORECAST, BECAUSE A MODEL WITH NO PUBLISHED ERROR RATE IS A SALES DOCUMENT. THE MODEL HAS OVER-CALLED BY 16.2% ACROSS FIVE EXECUTED MOVES, AND BOTH MISSES HAVE THE SAME ROOT CAUSE — ADOPTION IN A NINE-PARTNER EQUITY STRUCTURE, NOT WHETHER THE CHANGE WORKS. THE CONFIDENCE COLUMN ON SHEET 2 IMPLIES 67% REALISATION AGAINST THE 84% WE ACTUALLY ACHIEVED. WE WOULD RATHER UNDER-CALL IT TWICE.

Ranked by leverage — not by size

SIX MOVES SCORED · FIVE RECOMMENDED · ONE DECLINED AT RANK 2 · 9 MONTHS TO PROCESS OPEN · SPECIMEN

CONFIDENTIAL · RX 05FORM 884-X

Leverage = twelve-month EBITDA impact ($K) × confidence (0–1) ÷ effort (internal disruption, 1–5). The move with the largest impact on this page by a factor of six is the move we recommend against, and the reason is the clock, the chairs, and the ownership structure — in that order.

SORTED BY LEVERAGE. RANK 2 IS STRUCK — IT IS THE LARGEST-IMPACT MOVE ON THE PAGE AND WE RECOMMEND AGAINST DOING IT. THE REASON IS UNDER THE TABLE, NOT IN A FOOTNOTE.
MOVE12-MO EBITDACONF.EFFORTLEVERAGECOST — ONE-TIME + /MOMO. TO EFFECTRX
M-01 · Close the answer-rate gap at the 14 below-median desks; overflow routing and coverage for the two known lunch gaps$338K0.752126.8$0 + $9.0K3Rx 03
M-06 · Roll the acquisition programme into the three regions that declined at the start — 26 locations · DO NOT DO$2,284K0.255114.2$95K + $34.0K12+
M-02 · Same-day capacity matching at the 12 de novo sites — route demand to open chairs before it expires$230K0.65274.8$18K + $4.0K4Rx 03
M-04 · Zero-spend the four markets that fail the demand/capacity read (Form 118-M); redeploy 60% of the budget to capacity-rich markets$167K0.80266.8$6K + $02Rx 01
M-03 · Membership-plan push to move 2 points of production out of PPO into FFS/cash$271K0.45340.7$9K + $3.0K6Rx 03
M-05 · Hand the desk huddle (247-K) and the desk→media loop (247-L) to client managers; ADMEN attends monthly and on exception$72K0.90232.4$0 − $6.0K3Rx 05
RECOMMENDED SET — FIVE MOVES, M-06 EXCLUDED$1,078K0.67 wtd$723K wtd$33K + $10.0K net

TOTAL PROGRAMME COST THROUGH PROCESS OPEN: $33.0K ONE-TIME + $144.0K OVER NINE MONTHS, LESS $36.0K OF ADMEN FEE REMOVED BY M-05 = $141.0K. M-05 IS A MOVE WHOSE ENTIRE EBITDA CONTRIBUTION IS CUTTING OUR OWN FEE — IT IS ON THE PAGE BECAUSE A BUYER PRICES AN OPERATING SYSTEM HIGHER THAN AN AGENCY DEPENDENCY, AND BECAUSE WE WOULD RATHER PROPOSE IT THAN BE ASKED FOR IT IN DILIGENCE. M-04's REDEPLOYMENT LEG IS EBITDA-NEGATIVE IN YEAR ONE ($16.8K/MO OUT AGAINST $15.1K/MO OF FIRST-YEAR CONTRIBUTION); THE MOVE CLEARS ONLY BECAUSE THE WITHDRAWAL LEG IS LARGER. THE CHEAPEST MOVE ON THIS PAGE IS THE ONE THAT STOPS DOING SOMETHING. — and it's still only rank 4, because in a nine-partner structure "stop spending in your region" is not a low-effort conversation.

M-06 — RANKED SECOND BY LEVERAGE, LARGEST IMPACT ON THE PAGE, AND WE RECOMMEND AGAINST IT

Rolling the acquisition programme into the three regions that declined would put twenty-six more locations under it. At the cohort's demonstrated +13 new patients per location per month, that models to $2.28M of annualised EBITDA — six times the largest recommended move. We are not doing it. Three reasons, in the order that decides it:

1 · The clock. Nine months to process open. A programme started at month 22 produces its first clean quarter at month 28 and its first defensible trend at month 34 — after the process. What a buyer would see in the data room is spend up, volume not yet up, cost per new patient worse. That prices worse than not starting, and it is the single most common self-inflicted wound in a pre-exit year.

2 · The chairs. Two of the three regions are at or near chair capacity. Added spend there buys demand the schedule cannot absorb, and the only result is a worse cost per new patient — printed in the monthly pack, inside the diligence window, in the data room. Our own capacity read (Form 118-M) says so, and it would be read against us.

3 · It is not ours to decide. Partner-equity, nine owners, six regional P&Ls, no mandate. Two of our five executed moves already under-delivered for exactly this reason (sheet 1). Modelling a $2.28M move on an adoption assumption our own track record contradicts is how a model stops being useful.

This move is not deleted. It is handed to post-close integration (Rx 05.4) as documented headroom, with the cohort evidence attached, so the buyer can price the option rather than the seller claim the outcome. That is worth more in the data room than a half-run programme is worth on the P&L. Twenty-six unprogrammed locations is a better story told as headroom than as a nine-month scramble.

SEQUENCED AGAINST THE CLOCK · █ BUILDING · ▓ EFFECT LANDING IN THE NUMBERS · MONTHS 23 → 31 · PROCESS OPENS AT MONTH 31
MOVESTARTM23 24 25 26 27 28 29 30 31FIRST CLEAN READFITS THE CLOCK?
M-04 · zero-spend four markets, redeploy 60%m23██▓▓▓▓▓▓▓m25 · 610-Myes — 6 months of margin
M-01 · answer-rate gap, 14 desksm23███▓▓▓▓▓▓m26 · 610-Myes — 5 months of margin
M-05 · hand 247-K and 247-L to client managersm24░███▓▓▓▓▓m27 · 247-Lyes — and the evidence matters more than the dollars
M-02 · same-day capacity matching, 12 de novo sitesm24░████▓▓▓▓m28 · 610-Myes — 3 months of margin
M-03 · membership-plan push, 2 points of mixm25░░██████▓m31 · 610-Mtight — lands the month the process opens
M-06 · roll programme to the three declining regions█████████ → m35+m34 at bestNO — this is the argument

M-03 IS THE ONLY RECOMMENDED MOVE WITH NO MARGIN FOR ERROR: ITS FIRST CLEAN READ IS THE MONTH THE PROCESS OPENS, WHICH MEANS IF IT UNDER-DELIVERS THERE IS NO TIME TO CORRECT AND THE MISS IS VISIBLE. IT STAYS BECAUSE PAYER MIX IS THE ONLY MOVE ON THE PAGE THAT RAISES COLLECTIONS WITHOUT ADDING A SINGLE PATIENT — BUT IT IS FLAGGED HERE, NOT DISCOVERED IN MONTH 31. IF THE MONTH-28 PACK SHOWS FFS/CASH SHARE FLAT, WE PULL IT AND SAY SO.

Sensitivity, and the half we refuse to price

WHAT THE VALUE DOES IF A MOVE UNDER-DELIVERS · AND THE MULTIPLE ARGUMENTS, CARRYING NO DOLLARS AT ALL · SPECIMEN

CONFIDENTIAL · RX 05FORM 884-X

Every enterprise-value figure below is EBITDA × a stated multiple. The multiple is an input we do not control and no move on sheet 2 is credited with moving it.

PER-MOVE SENSITIVITY — WHAT COMES OFF THE VALUE IF A MOVE DELIVERS HALF, AND THE INSTRUMENT THAT WOULD TELL US FIRST
MOVEMODELLED EBITDAAT HALF DELIVERYEV SWING AT 8.4×EARLY INDICATOR — WHAT WE WATCH, AND WHERE
M-01 · answer-rate gap$338K$169K−$1.42Manswer rate at the 14 flagged desks · month 25 board pack (610-M)
M-03 · payer-mix push$271K$136K−$1.14MFFS/cash share of production · month 28 pack — the go/kill read
M-02 · same-day capacity matching$230K$115K−$0.97Msame-day fill rate at the 12 de novo sites · month 26 pack
M-04 · zero-spend + redeploy$167K$84K−$0.70Mloaded cost per new patient in the four withdrawn markets · month 24 pack
M-05 · hand the loops to the client$72K$36K−$0.30Mhuddle held without ADMEN in the room · month 25 loop minutes (247-L)
ALL FIVE AT HALF DELIVERY$1,078K$539K−$4.53MEVERY INDICATOR LANDS BEFORE MONTH 29 EXCEPT M-03's
SCENARIO GRID — EBITDA × MULTIPLE. THE 9.0× COLUMN IS THE ONLY ONE ANCHORED TO A PUBLISHED FIGURE, AND IT IS ANCHORED TO THE BOTTOM OF A BAND PUBLISHED FOR PLATFORMS LARGER THAN THIS GROUP.
SCENARIOEBITDA AT PROCESS OPENAT 8.4× — MODELAT 8.9× — MODELAT 9.0× — PUBLISHED FLOORΔ vs BASE AT 8.4×
BASE — no moves executed$9.20M$77.3M$81.9M$82.8M
DOWNSIDE — M-02 and M-03 deliver nothing, the rest land at confidence weight$9.65M$81.1M$85.9M$86.9M+$3.8M
MODEL — all five at confidence weight (67% realisation)$9.92M$83.3M$88.3M$89.3M+$6.0M
UPSIDE — all five delivered in full$10.28M$86.4M$91.5M$92.5M+$9.1M

PROGRAMME COST $141.0K AGAINST $6.0M OF MODELLED ENTERPRISE VALUE AT A CONSTANT MULTIPLE — WHICH IS AN ASSUMPTION, NOT A FINDING. AND THE RATIO NOBODY PUTS IN THESE DOCUMENTS: ALL FIVE MOVES DELIVERED PERFECTLY BUY $9.07M OF VALUE AT 8.4×. ONE FULL TURN OF MULTIPLE ON $9.92M OF EBITDA IS WORTH $9.92M. THE ENTIRE ARITHMETIC HALF OF THIS DOCUMENT IS WORTH 0.91 TURNS. EVERYTHING BELOW IS WORTH MORE IF IT LANDS — WHICH IS EXACTLY WHY WE WILL NOT PUT A NUMBER ON IT.

MOVES THAT ADD MULTIPLE — ARGUMENTS, NOT ARITHMETIC. NO DOLLAR FIGURES APPEAR IN THIS SECTION, BY POLICY.

ARGUMENTSTRENGTHWHAT A BUYER MUST ACCEPT · AND THE EVIDENCE THAT SUPPORTS ITWHAT KILLS IT
B-01Owner-independence
STRONG
That the operating cadence runs on client staff, not on us. Evidence: Form 247-L loop minutes for months 7–34 with a named client owner per form; the desk huddle run by the client's own managers; eleven quarterly recommitment memos (610-Q), each containing a pre-filled case for ending the engagement.
The desk managers leaving before close. Key-person risk sits with the operators and a data room cannot fix it. There is also no post-close evidence — nobody has yet run this cadence under new ownership.
B-02Attribution provability
STRONG
That marketing-attributed collections have tied to GL 4000 every month for thirty-two consecutive closes, each signed by the client's controller before publication, with the standing rule that where attribution and the PMS disagree the PMS wins.
A quality-of-earnings team that wants incrementality rather than attribution. We never ran a post-rollout holdout and we cannot supply one. That refusal is written into Form 884-N as declined claim D-03 rather than argued in the room.
B-03De novo ramp legibility
MODERATE
That the blended per-location decline is a ramp, not decay — because twelve sites opened inside the window. Evidence: cohort-separated curves per site in tab 06 of the diligence index, plus National Vision's filed ramp (55% of year-five sales in the first full year, profitable in year two) as the only external anchor.
Small n. Twelve sites, five of them with under twelve months of data. And National Vision is a chain of optical stores — it bounds the shape of a ramp, not its depth. A buyer is entitled to mark the blended number anyway.
B-04Regional adoption — 16 of 42
WEAK · CUTS BOTH WAYS
Two readings exist and we present both. Headroom: twenty-six locations have never run the programme, and the cohort evidence shows what it does when it runs. Governance risk: the programme could not be mandated in a nine-partner equity structure, and two executed moves already under-delivered for exactly that reason.
Presenting only the headroom reading. A buyer who finds the governance reading themselves discounts everything else in the book, and they will find it — the adoption number is on the face of the record. We would rather concede it than defend it.

DEFERRED PAST THE CLOCK — HANDED TO POST-CLOSE INTEGRATION (RX 05.4), NOT DROPPED

Hub re-point and location-template consolidation across all 42 sites. The hub domain outranks its own location pages on the money queries; the month-13 template roll delivered 35% under model for that reason. The fix is real and the impact is high. It is not executed inside the clock because a six-to-nine month ranking-volatility window landing inside diligence is a valuation event whether or not the change is correct — a buyer reading a traffic decline in month 30 does not care that it is temporary and deliberate. We also decline to attach a percentage to the cannibalisation itself, because no published measure of self-cannibalisation in local search exists and we would have to invent the denominator (Form 884-N, declined claim D-04).

Both deferred items — this and M-06 — go into the data room as documented headroom with the evidence attached, so the buyer prices an option instead of the seller claiming an outcome. A seller who claims it gets challenged. A seller who documents it and declines to claim it gets believed.

— specimen —

PARTNER, ADMEN · MODEL OWNER

— specimen —

REVIEW GATE — SECOND PARTNER, PRE-DELIVERY

— specimen —

CLIENT CEO + CFO · ACCEPT SEQUENCE AND THE M-06 DECISION · DATE