By the MeridianCogent team — built from working with integration offices, separation leaders, and PE operating partners on M&A execution. This perspective is drawn from published carve-out and TSA research across the industry.
TSA overruns almost never arrive as a single, obvious failure. They arrive as a stream of small, individually-defensible decisions — one more month on IT, a slightly delayed payroll cutover — each of which looks minor in isolation. The cost only becomes visible in aggregate.
The exact escalation math
TSA pricing is typically set once, at signing, on the assumption the relationship is short-term. Once a workstream runs past its base term, most agreements step the price up on a defined schedule — not linearly, but in bands designed to make lingering progressively more painful:
| Period | Pricing Basis | Illustrative Rate |
|---|---|---|
| Base term (months 1–12) | Cost + fixed margin | Cost + 10% |
| First extension (months 13–15) | Cost + escalated margin | Cost + 35–50% |
| Second extension (months 16–18) | Cost + penalty margin | Cost + 75–100% |
| Beyond month 18 | Cost + severe penalty margin | Cost + 100–150%+ |
This structure is deliberate. A service still running 6 months past its base term should cost meaningfully more than double the original rate — otherwise there's no real incentive to finish the cutover.
What this looks like in euros, using a €135k/month base TSA (ERP, IT infrastructure, finance, payroll, help desk combined):
| Period | Rate | Total |
|---|---|---|
| Months 1–12 (base) | €135k/month | €1.62M |
| Months 13–15 (extension) | ~€196k/month | €588k |
| Months 16–18 (extension) | ~€250k/month | €750k |
An 18-month TSA that should have closed at 12 costs an additional €1.34M beyond the base term.
A 3-month extension doesn't cost 3× the base monthly rate. It costs closer to 4.3× once escalation is applied.
The structural reason TSA terms are always too short
The most common root cause is that the TSA term was set against the legal closing calendar — what both sides needed the number to be in order to sign — rather than the actual separation plan's realistic timeline. It's a known gap at signing that both sides often accept anyway, because closing on time matters more in the room than the operational reality six months out.
Carve-out complexity compounds the exposure
McKinsey's analysis of carve-out cost structures has found that standalone costs for services previously delivered at parent-company scale can run as much as 200 percent higher once replicated independently. A finance function allocated at €24k/year in the parent's books can cost €70k+ standalone once you actually build it — a gap that typically surfaces mid-TSA, right when there's no runway left to absorb it cleanly.
The value destruction math
Bain's longitudinal data on carve-out outcomes since 2012 shows an average MOIC of roughly 1.5x, with top-quartile execution closer to 2.5x. For a $400M acquisition targeting 2.0x, a realistic $25M in TSA overruns, discovery costs, and extended services drops the MOIC to 1.94x — a real dent in fund-level returns, not a rounding error.
Why the fix usually isn't "negotiate a longer TSA"
A longer TSA still lacks the defined exit criteria, ownership, and acceptance testing that determine whether a service is actually transitioned or just still running. What works instead: every service inside the TSA needs a named owner on both sides, a defined completion standard, a tracked exit date with real consequences, and live visibility into where each workstream stands. That discipline is exactly what separates the carve-outs that exit in 2–3 months post-readiness from the ones still running TSAs at month 18.
For where the underlying scope gaps actually come from — the systems and services nobody priced in at signing — see Scope Creep in M&A Carve-Outs. For the TSA structure itself, What Is a Transition Service Agreement covers the base mechanics.
Lessons learned
- Extensions are pricing failures, not plan failures. You didn't fail to exit — you failed to price for the reality. Build 20–30% contingency into the TSA cost model from day one.
- Month 13 is exponentially more expensive than month 12. A hard call in month 11 (cut over with temporary workarounds, or extend now and lock in known pricing) beats discovering you need to extend in month 13 with no leverage left.
- Stranded costs are invisible until the seller's P&L is separate. Model for the fact that the seller's IT budget doesn't shrink just because the divested unit left.
- Escalation should be aggressive by design. 15–25% uplift per month on extensions isn't punishment — it's the only mechanism that keeps a "temporary" cost temporary.
Sources: McKinsey & Company (carve-out cost structure analysis); Bain & Company (carve-out MOIC research, data since 2012); BD Emerson (TSA and IT separation cost analysis); FTI Consulting (carve-out economics and value creation research).