What the agreement is actually for
In nearly every carve-out the divested business cannot operate alone on day one. It depends on the parent for identity systems, payroll, finance, security monitoring and sometimes customer-facing services. The transition services agreement governs that dependency and sets the clock on ending it.
Treated as a legal afterthought, it becomes the thing both sides argue about for the following two years. Treated as a product, with a defined scope, a price and a service level, it becomes a selling point, because it tells a buyer that separation risk has been thought through by someone who has done this before.
Scoping each service properly
Every service in the agreement needs five things stated: what is provided, at what volume, to what standard, for how long, and at what price.
Vague scope is the leading cause of post-closing disputes. An entry that says "IT support" without hours, response times and a definition of what is in scope guarantees that the two parties have different expectations, and the disagreement surfaces at the worst moment, when the seller has already been paid and has no commercial reason to be generous.
Build the catalogue from the dependency map produced during preparation. If a service is not on the list, the buyer will assume they are not getting it, and if they need it after close they will be negotiating from a weak position.
Pricing
Cost plus a modest margin, usually five to ten percent, is the market standard and the right answer for most services.
Price below true cost and the seller subsidises the buyer while their own people carry the work, which produces exactly the resentment that makes service quality slip. Price punitively above and it becomes a negotiating point that costs goodwill during the period when both sides most need to cooperate.
Cost means fully loaded cost, including a share of the overhead that supports the function. A seller who prices at direct salary only will find that the agreement loses money for two years.
Duration and the exit ramp
Twelve months with two three month extensions concentrates minds better than a flat twenty four month term, because a deadline that arrives twice creates two moments where somebody has to make a decision.
Set the extension price higher than the base price. The buyer should feel a cost to delay, because without one the incentive to invest in standalone infrastructure is weak and the seller ends up running a service business they did not want.
Include a mechanism for terminating individual services early as the buyer builds capability. Services should fall away one by one rather than all at once, and the buyer should not pay for a service they have already replaced.
The provisions people forget
Who does the work. Naming the team, or at least the function and the escalation path, prevents the situation where the service is delivered by whoever has time, which after a divestiture is usually nobody.
Change control. Volumes change and requirements change. A mechanism for pricing a change is better than an argument about whether the change was in scope.
Audit rights for the buyer, so they can verify that cost-plus pricing reflects actual cost.
A joint steering group with a defined cadence. Most disputes are operational rather than legal, and a monthly meeting resolves things a contract clause cannot.
Data and security obligations, particularly where the parent continues to process the divested business's customer data. This is a regulatory exposure for both sides and it does not end when the service does.
What sellers get wrong
Starting the negotiation late. Transition services should be scoped during preparation, alongside the standalone cost model, because the two are the same analysis viewed from different ends.
Staffing the delivery with people who have already mentally moved on. Assign named resources with the work in their objectives, or accept that service levels will slip and the buyer will have a legitimate complaint.
Treating the agreement as a way to keep revenue. It is not a business line, it is a bridge, and a seller who tries to extend it for margin damages a relationship they may need again.
What buyers get wrong
Signing without a migration plan. The agreement buys time to build standalone capability, and a buyer who has not budgeted for that capability arrives at month eleven with nothing built and no leverage.
Assuming the service level will match what the division experienced inside the group. It rarely does, because inside the group the service was provided by colleagues with shared incentives. Benchmark against the pre-transaction level explicitly and write it down.
Underestimating what happens after the agreement ends. The steady state cost of running those functions independently is what belongs in the valuation model, not the transition price.
The test of a good agreement
Both sides should be able to read it a year after signing and agree on what is owed without calling their lawyers. If the document requires interpretation, it was not specific enough, and the cost of that vagueness will be paid by whichever party has less leverage at the moment the disagreement surfaces.
Editorial Team · Published for orientation, not as advice on a specific transaction. Any figure cited is orientation, not a valuation. See market notes.