Engagements end. They end because the work ended, because the business changed direction, because it did not work, or because you are bringing the work in-house. Ending well is cheaper than ending badly and is mostly a matter of sequence.
Give more notice than the contract requires
The contractual minimum is a floor. The provider has employment obligations that the entry on Philippine employment describes, and more notice gives them time to redeploy somebody rather than make them redundant.
It costs you very little and it determines whether a person who worked for you for three years has a job in six weeks.
Ask what happens to them
Directly, and before you give notice. Providers of any size redeploy; smaller ones may not be able to. If the answer is redundancy, the entry on statutory pay explains what is owed and the agreement should say who bears it.
This is a question about people and it should be asked out loud rather than left to the contract.
The handover, in week one
Not week four. Ask for it at the start of the notice period, while the person still has both the knowledge and the motivation. By the final fortnight they have mentally left and the document will show it.
What you want: the current procedures, the exceptions that are not in them, the contacts, the things that go wrong periodically and what to do about them.
Tell the person, but in the right order
They should hear it from their employer first. That is the provider's job and it is not yours to pre-empt.
Then say something yourself. Somebody who has worked on your account for years deserves a conversation rather than a silence, and a client who disappears at the end is remembered as one.
The administrative part
Revoke access on the last day, using the offboarding list from the entry on access. Retrieve anything held on provider systems.
Confirm that the documentation is on your systems, which it is if the arrangement in the entry on what to keep onshore was set up at the start, and is a scramble if it was not.
Bringing work back in-house
Plan an overlap of several weeks and expect a productivity dip regardless. The incoming person is learning a process that has been running without them, from documentation that is good rather than perfect.
Budget the dip rather than being surprised by it, and do not schedule the transition into a peak.
Ending because it did not work
Establish which of three things happened before deciding what to do next: the role was wrong, the brief was wrong, or the arrangement was wrong. The entry on choosing the first role sets out how to tell.
Concluding that offshoring does not work, on one engagement with an undefined role, is a conclusion about that engagement.
Ending well, and why it matters commercially
This industry is small and people move between providers. A client known for giving notice, funding a proper handover and treating people decently at the end recruits more easily the next time, including through the same provider.
That is a self-interested argument for behaving well and it is a real one.
The last entry in this journal
Thirty entries, and the argument in all of them is the same. The difference between an offshore team that works and one that does not is almost never the people. It is the definition of the work, the preparation, the documentation, the cadence and the attention, and every one of those is decided by the business engaging the team.
Which is inconvenient for a provider to say and remains the most useful thing we know about this.
The reference
Offer one, in writing, to anybody who worked well on your account. It costs ten minutes, it matters a great deal in a labour market where outsourced staff struggle to evidence what they did, and it is the easiest decent thing available at the end.