Cumberland Valley Health Plan is being carved out of Cheatham Mutual Holdings, and it does not own everything it runs on. Payroll, the general ledger, network and end-user computing, security monitoring, data center hosting and core administration support are all provided by its parent. At closing those do not transfer — they continue on a meter, under a Transition Services Agreement, until ACME stands each one up itself. This document is the schedule of those six services, the criteria for exiting each, and the plan for doing so inside twelve months against an eighteen-month contractual wall. Approved August 28, 2023, before closing.
Exiting the TSA is not a workstream alongside the integration. It is the integration. Every month a service remains is money leaving the business for capability the combined entity was supposed to own, and the fee schedule is deliberately designed to make that hurt more over time. A program that treats the TSA as a safety net has misread it: it is a countdown with an invoice attached, and the seller wrote both.
Table of Contents
Part I — The Instrument
Part II — Getting Out
- Exit Sequence and Dependencies
- Exit Criteria — What "Done" Means
- Knowledge Transfer as a Deliverable
Part III — Governance
1. Why a TSA Exists Here
A whole-company acquisition rarely needs a Transition Services Agreement: the target owns its systems and its staff, and they transfer with it. A carve-out is different. Cumberland Valley operated as a subsidiary, consuming shared services from Cheatham Mutual the way any division consumes corporate infrastructure. Those services are not Cumberland Valley's to bring.
| Item | Provision |
|---|---|
| Provider | Cheatham Mutual Holdings — the divesting parent |
| Recipient | ACME Health, on behalf of the acquired operations |
| Commences | At closing, September 29, 2023 |
| Program exit target | September 30, 2024 — 12 months |
| Contractual maximum | March 31, 2025 — 18 months. Extension beyond requires provider consent, which is not assumed. |
| Services | Six, exited independently rather than as one event |
| Owner | G. Threadgill, TSA Manager, ACME Health |
| ⚠ Also required | A Business Associate Agreement with Cheatham Mutual — the provider handles protected health information. See the BAA Register, BAA-15. |
The last row is the one that gets missed, and the carve-out structure is what creates it. Cheatham Mutual is the party ACME just bought a company from — a counterparty, a negotiating adversary, in some sense the other side. It is also now hosting member data and supporting a claims platform, which makes it a business associate under HIPAA like any other vendor. Nobody files the seller under "vendors," which is exactly why that BAA is the most likely of the fifteen to be overlooked.
2. Service Schedule
| Ref | Service | Exit month | Exit depends on | What ACME must stand up |
|---|---|---|---|---|
| TS-01 | Payroll and HR administration | Month 3 | HRIS and payroll migration | Associates on ACME payroll and benefits, one full cycle run in parallel first |
| TS-02 | General ledger and financial close | Month 6 | GL migration at a period boundary | Combined chart of accounts, close calendar, statutory reporting path |
| TS-03 | Network, telecom and end-user compute | Month 8 | Identity and access migration | ACME network, directory, device management for acquired staff |
| TS-04 | Security operations and monitoring | Month 9 | Landing zone security controls live | Log ingestion, alerting, incident response covering acquired estate |
| TS-05 | Data center hosting and operations | Month 12 | Final cloud migration wave | Azure landing zone and all migrated workloads running |
| TS-06 | Core administration platform support | Month 12 | Core admin cutover and parallel run | Claims adjudicating on ACME's platform for the combined book |
2.1 The exit profile
| Service | M1 | M2 | M3 | M4 | M5 | M6 | M7 | M8 | M9 | M10 | M11 | M12 | M13 | M14 | M15 | M16 | M17 | M18 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TS-01 Payroll & HR | ||||||||||||||||||
| TS-02 General ledger | ||||||||||||||||||
| TS-03 Network & compute | ||||||||||||||||||
| TS-04 Security operations | ||||||||||||||||||
| TS-05 Data center hosting | ||||||||||||||||||
| TS-06 Core admin support | ||||||||||||||||||
| Negotiated margin | ||||||||||||||||||
Six services, six exits, not one. A single big-bang TSA termination would concentrate every dependency into one date and give the program no way to demonstrate progress until the very end. Independent exits mean payroll leaves the meter in month three while the platform work is still in planning — which reduces cost early, proves the exit machinery works on something small, and gives the Steering Committee something real to see at Day 100.
3. Fees and Why They Escalate
| Period | Fee basis | Effect |
|---|---|---|
| Months 1–6 | Cost plus a modest margin | Roughly what the service cost the parent to provide. Comfortable. |
| Months 7–12 | Stepped increase | The point at which staying becomes visibly more expensive than leaving. |
| Months 13–18 | Materially higher step | Priced to be unattractive. This is the negotiated margin, not a plan. |
The escalation is not the provider being difficult. Both parties wanted it. Cheatham Mutual wants its people back on its own work and does not want to run a service business for a competitor indefinitely. ACME wants internal pressure to exit that does not depend on the program manager nagging. An escalating fee converts "we should get off this" into a number the CFO sees every month — which is a far more reliable forcing function than good intentions. A flat-fee TSA is comfortable, and comfortable is how organizations arrive at month eighteen still running two of everything.
TSA fees are an operating cost carried by the receiving functions, not by the program budget. That placement is deliberate: the function that is slow to absorb a service should feel the cost of being slow.
4. Exit Sequence and Dependencies
The sequence is not chosen for convenience. Each exit is gated by a migration completing, and the order follows from which migrations can complete first.
| Exit | Gated by | Why it can go early / must go late |
|---|---|---|
| TS-01 Payroll | AD-17 HRIS and AD-18 payroll migration | Early. Self-contained, no member data, no regulatory calendar. Also the most personally visible failure in the program, so it runs in parallel for a full cycle first. |
| TS-02 General ledger | AD-15 GL migration at a fiscal period boundary | Early-ish. Constrained by the close calendar — a ledger cannot move mid-period — but otherwise independent. |
| TS-03 Network & compute | AD-19 identity and access migration | Middle. Blocks nothing downstream but depends on identity, which blocks a great deal. |
| TS-04 Security operations | Landing zone security controls operational | Middle. ⚠ Cannot precede TS-03 — monitoring a network you do not yet operate is theater. |
| TS-05 Data center hosting | Final cloud migration wave complete | Late. The last workload out of the parent's data center sets this date. Nothing accelerates it. |
| TS-06 Core admin support | Core administration cutover and parallel run complete | Late, and the critical path. Depends on member identity resolution, which depends on data profiling. The longest chain in the program. |
TS-06 is the whole program in one row. To exit core administration support, claims must adjudicate on ACME's platform for the combined book. That requires the target's members to exist correctly in ACME's systems, which requires identity resolution, which requires knowing the true condition of the target's member data — which the Clean Team Protocol made unknowable until after closing. The longest and least compressible chain in the program starts at the one thing nobody was permitted to measure in advance.
5. Exit Criteria — What "Done" Means
Each service exits on evidence, not on a date arriving. Criteria were agreed with the provider before closing, because a criterion negotiated while a service is failing is not a criterion.
| Service | Exit criteria |
|---|---|
| TS-01 Payroll & HR | One full pay cycle processed on ACME systems in parallel with zero variance; benefits enrollment complete; year-to-date balances reconciled and signed off by Finance |
| TS-02 General ledger | One full period closed on ACME's ledger; trial balance reconciled; statutory reporting produced from the combined ledger and accepted by the actuary |
| TS-03 Network & compute | All acquired-staff devices managed by ACME; directory migration complete; no residual dependency on parent network paths, evidenced by traffic analysis |
| TS-04 Security operations | Full log ingestion from the acquired estate; alerting tuned; one tabletop incident exercise run end to end with ACME staff |
| TS-05 Data center hosting | Zero workloads remaining; data destruction certificate issued by the provider; ⚠ backups and archives accounted for, not only live systems |
| TS-06 Core admin support | Claims adjudicating on ACME's platform at combined volume; parallel run reconciled to agreed tolerance; EDI trading partners re-registered; open appeals cases either closed or migrated with regulatory record intact |
Two of these criteria are about what remains behind rather than what has moved. TS-05 requires a data destruction certificate covering backups and archives — a migration can be complete while years of the target's member data still sit in the parent's backup estate, which is a live privacy exposure with no operational symptom. TS-06 requires open appeals to survive with their regulatory record intact. Exit criteria that only measure the new system are half a test.
6. Knowledge Transfer as a Deliverable
A TSA obliges the provider to perform a service. It does not oblige them to make you capable of performing it yourself. Those are different things, and the gap between them is where TSA exits fail. Twelve months can pass with every service running perfectly, every invoice paid, every service level met — and the receiving organization no more able to run the service on the last day than on the first. Then the meter stops and nobody knows how the month-end close works.
Knowledge transfer is therefore specified as a deliverable with named receivers and acceptance criteria, not assumed as a by-product of the service running.
| Service | Named receiver | Transfer deliverable and acceptance |
|---|---|---|
| TS-01 Payroll & HR | D. Marchbanks, SVP HR | Documented procedures plus a parallel cycle run by ACME staff with provider observation only |
| TS-02 General ledger | S. Beauregard, VP FP&A | Close calendar, reconciliation procedures, one period closed by ACME staff unaided |
| TS-03 Network & compute | B. Trammell, VP Infrastructure | Topology, configuration baselines, escalation paths; shadowing then reverse-shadowing |
| TS-04 Security operations | A. Quintanilla, CISO | Runbooks, alert logic, tuning rationale; tabletop exercise led by ACME |
| TS-05 Data center hosting | H. Sandifer, Manager Platform Engineering | Asset inventory, dependency map, decommissioning evidence |
| TS-06 Core admin support | W. Ferriday, VP Claims Operations | Configuration documentation, custom edit inventory with policy rationale, defect history |
Note the pattern in the acceptance column: ACME staff perform, the provider watches. That reversal is the only acceptance test worth having. A knowledge transfer session where the provider demonstrates and the receiver takes notes proves that the provider knows how — which was never in doubt. Reverse-shadowing is uncomfortable, slower, and the only thing that establishes the receiving side can actually do the work.
⚠ This obligation interacts directly with retention. The provider's staff who know these services are Cheatham Mutual employees, and the target-side staff who know the core administration platform are covered by retention agreements precisely because they must survive to discharge this transfer. Attrition against those roles is not an HR statistic; it is a TSA exit risk with a date attached.
7. The Margin Between Plan and Wall
| Date | Months post-close | Status |
|---|---|---|
| September 30, 2024 | 12 | Program target. The plan, and what the synergy model assumes. |
| March 31, 2025 | 18 | Contractual maximum. A wall, not a target. Rank 1 in the Charter's constraint priority order. |
The six-month gap is negotiated margin, agreed at signing, and it exists for a specific reason. At the time the TSA term was negotiated, nobody could examine the target's member data — so nobody could size the identity resolution work that gates TS-06, the longest exit. Buying six months of headroom against an unmeasurable dependency was the correct response to that condition. The margin is protection against what discovery reveals, and the plan does not spend it.
Consuming margin and breaching a maximum are categorically different outcomes, and the Charter treats them that way. Spending buffer you deliberately purchased at signing is program management working — it retroactively justifies having negotiated the buffer. Blowing through a contractual maximum is a governance failure that would overshadow everything else the program achieved, and it hands the provider leverage at the worst possible moment. The margin may be spent. The wall may not be touched.
8. Governance, Disputes and Extension
| Mechanism | Provision |
|---|---|
| TSA Manager | G. Threadgill, ACME — single accountable owner for all six services. Not distributed across the receiving functions. |
| Cadence | Monthly service review with the provider; exit readiness reported to the Steering Committee |
| Service levels | Defined per service, measured by the provider, audited by ACME. ⚠ Measurement by the provider is normal and is the reason for the audit right. |
| Dispute resolution | Escalation to named executives on both sides before any formal mechanism. Most disputes are scope disputes about what "the service" includes. |
| Exit certification | Signed by the named receiver and the TSA Manager, evidenced against Section 5 criteria |
| Extension | Requires provider consent and Steering Committee approval. ⚠ Must state impact against the 18-month maximum and the remaining margin. |
| Reverse services | ACME provides limited transitional support to the provider where systems were genuinely shared. Small, and reciprocal goodwill is worth more than the fee. |
One accountable TSA owner rather than distributed ownership is a deliberate choice, and it runs against the instinct to let each function manage its own service. Distributed ownership produces six people who each believe their exit is on track and nobody who can see that four of them depend on the same identity migration. It also means that when the provider proposes a scope interpretation, six different people negotiate six different answers. The TSA is one commercial relationship with one counterparty, and it needs one throat to choke — on both sides.
Related artifacts: 1 — Integration Charter · 2 — Deal Summary · 5 — Clean Team Protocol · 7 — Due Diligence Findings (DD-08) · 19 — Retention & Key Talent · 20 — Application Disposition Matrix · 21 — Vendor & Contract Disposition Matrix · BAA Register · 45 — TSA Exit Readiness Report