Microsoft PIP: the 16-week offer and the rehire ban

Every figure here is reported, and traces to a small number of sources

This page rests on reporting of internal Microsoft documents from 2025, now repeated across several outlets, plus employee accounts. Repetition across outlets is not the same as independent corroboration, and the underlying source is narrow. The policy was introduced in 2025 and may since have been revised.

Treat the shape as reliable and the numbers as indicative.

The choice: enter the plan, or take sixteen weeks' pay

The reported sequence: around 2,000 people terminated for underperformance in early 2025 without severance, followed by a new globally standardised improvement plan communicated to managers in late April 2025.

An employee flagged for low performance is reportedly offered a straight choice: enter the plan, or take a Global Voluntary Separation Agreement worth 16 weeks' pay.

Unlike Amazon's descending ladder, the number is fixed and the choice is presented once. That makes it a cleaner decision problem and a harder deadline.

The window is reportedly five business days. What that window forecloses is the subject of the next section.

Five business days, and a door that does not reopen

The procedural detail on this page comes from a single coaching account dated March 2026, re-read in September. One source, and a firm with an interest in being consulted, so weigh it accordingly. What it describes, though, is a decision shaped quite differently from the one people assume:

• Five business days to choose. As it puts it, "before that clock runs out, you face a binary decision: take approximately 16 weeks of base pay as a lump sum and leave immediately, or decline the money and attempt the improvement plan."
• Declining is permanent. "Once you decline the offer, it is permanently off the table." The separation agreement is not held open behind the plan.
• Failing the plan pays nothing. Termination at the end of a failed plan is reported to carry no additional severance.
• Unvested stock does not vest for somebody who takes the offer.

Read those together and the arithmetic is not the one the framing suggests. Sixteen weeks is not a floor underneath an attempt at the plan. It is the alternative to attempting it, and the attempt has a floor of zero.

That does not mean take the money. It means the comparison is between a certain sixteen weeks and an uncertain outcome worth either your salary continuing or nothing, and the second term is the one worth thinking hard about rather than the first.

Two things change that arithmetic for individual people, and both are worth working out before the five days rather than during them:

• Unvested stock can dwarf the cash. For somebody several years in, the RSUs that do not vest may be worth considerably more than sixteen weeks of base pay, and they do not appear anywhere in the phrase "sixteen weeks". Look up your vesting schedule before you decide, not after.
• Local law may put a floor under the plan route that the offer does not reflect. The section below on working outside the United States is about that, and in several European jurisdictions the statutory entitlement for long-tenured staff exceeds sixteen weeks on its own.

The plan itself, and the absence at the end of it

The plan is reported to run up to 90 days, with goals that have to be met to keep the job.

What is missing from every account of this process is the thing Amazon has and Microsoft reportedly does not: no appeal is described anywhere. At Amazon a failed plan can go to a panel of peers who hear the employee and the manager both. Nothing in the reporting on Microsoft describes an equivalent, which means the manager's assessment at the end of the ninety days is reported to be the end of it.

That absence is worth naming rather than leaving as a gap in the page, because it changes where effort is worth spending. Where there is an appeal, evidence collected during the plan has a forum to be heard in later. Where there is not, the only audience for that evidence is the manager writing the assessment, and anything you want considered has to reach them while the plan is running and in a form they have to respond to.

So the practice is the same as everywhere else in this guide and the reason for it is sharper: put the deliverables against the goals in writing, on the record, at each check-in rather than at the end. Not to build a case for a hearing that does not exist, but because the person deciding is reading as they go.

What actually matters more: the transfer block

Employees in the lowest reward outcomes are reportedly ineligible for internal moves.

The escape hatch closes before the conversation happens. Transferring to another team is the classic survival move, and it is blocked by the rating rather than by the plan — so by the time the plan arrives, mobility is already gone.

That reframes the first question worth asking. It is not "how do I pass this plan" but "what is my rating, and am I still transfer-eligible". One reported ambiguity is worth flagging: some coverage describes the block as applying to the lowest reward payout buckets, other coverage as applying to the bottom 60% of ratings. Those are very different nets, and the narrower reading is the one this page takes.

Can I be rehired after leaving on a plan?

Anyone leaving under low-performance circumstances — or during or after a plan — is reportedly ineligible for rehire for two years.

It reportedly attaches to both doors: taking the separation agreement and failing the plan. So it should not tilt the decision at all. Choose on cash, on time, on the visa clock if you have one, and on an honest reading of the plan. Preserving the option to return is not something either path preserves.

"Good attrition"

Departures of low performers are counted under a metric mirroring Amazon's unregretted attrition. Reporting notes that no formal targets were set when it was introduced, though the metric is discussed at senior levels — so the quota logic is present in a weaker form than at Amazon.

The same reading applies: a metric existing at all means a placement may say less about the person than it appears to. The same three-part shape repeats across the large employers this guide covers: a counted attrition metric, a stage before the formal plan, and a priced way out. It is visible at Amazon, Meta and Google as well as here.

The number is knowable in advance, which is unusual

Sixteen weeks is a concrete anchor. That lets you do arithmetic most people in this situation cannot:

• Sixteen weeks, against
• the remaining salary if you attempt the plan, weighted by a realistic view of passing, plus
• whatever your local law would require, plus
• for a visa holder, the fact that the grace period starts on the employment end date whichever door you take

If you are not in the United States

The policy is described as standardised across geographies, but the local dismissal regime still governs the legal exit. A Microsoft plan in Dublin runs into Irish fair-procedures law; in Germany into the Abmahnung requirement and works-council consultation; in the Netherlands into the verbetertraject preconditions.

A globally uniform "plan or 16 weeks" offer can be a weak offer where local law would require more — and in several European jurisdictions the statutory floor for long-tenured staff exceeds sixteen weeks. Read the page for where you actually work — the guide lists them by country — before deciding the door is the whole choice.

One more trap: a "voluntary separation agreement" may be characterised as a voluntary exit for benefit purposes in some jurisdictions. How it is papered matters.

None of this comes from Microsoft

Fairhanded is not a law firm, is not affiliated with Microsoft, and has no access to its internal policies. Before signing a separation agreement — particularly outside the United States — take it to an employment lawyer where you work.