Performance Improvement Plan

When someone keeps falling short, managers talk around it for a long time. "It's been a bit light lately." "It would be good to pick up the pace." The person saying it believes the message landed; the person hearing it files it as a minor note. The same conversation repeats for four months and nothing moves.

That is what happened at Pagedeck. Grace hinted at it all half, and the person wrote "made steady progress" in their own self-review. Two people remembered the same six months completely differently, and neither of them was lying. The standard had simply never been written down anywhere they could both look at it.

Then at some point the company decides something, and what reaches the person is a conclusion with no warning attached. They never learn what the actual problem was, and the company has no document saying what it asked for or why it didn't happen.

A performance improvement plan (PIP) exists to reverse that order. Write the standard down before reaching any conclusion, give the person a defined stretch of time to meet it, and record what the company will do to help during that stretch. Four months of vague remarks get converted into one page.

Six items. Leave any of them out and the document stops working as a standard.

1. The current state, stated as fact. Not "lacks initiative" but "responses to code review requests have averaged three days against a team norm of one day." It has to be a verifiable observation rather than a verdict, or the next item can't be written at all.
2. The expected level. What counts as clearing the bar, in measurable terms. For the example above: responses inside one day, sustained for four consecutive weeks.
3. The deadline. Thirty to ninety days is typical. Shorter when the thing to change is a habit, longer when it's a skill to be learned. Two weeks isn't long enough to observe any change; six months leaves the person anxious for half a year.
4. How and how often progress gets checked. Thirty minutes a week, with what will be looked at agreed in advance. Judge only once at the end and there is no chance to correct course in the middle.
5. What the company will provide. A senior colleague to pair with, one project taken off their plate, the cost of a course covered. Leave this item empty and the document is just a list of demands.
6. What happens when the window closes. Meeting the standard returns things to normal; missing it puts specific consequences on the table. Writing that down at the start is fairer to the person than revealing it later.

Items 1 and 2 are where the writing usually breaks. The two sentences have to be a matched pair. Put "collaboration attitude" in item 1 and "review responses inside a day" in item 2, and when the deadline arrives the argument about what to measure starts over.

The direct manager writes it. Someone in HR reviews the wording and the process, and a senior manager or founder signs off. A manager who writes and judges alone means the company applies a different standard to each person in the same situation.

Three audiences read it. The first is the person it's about, who has to be able to read it and know what to do. The second is the manager who will run the weekly checks. The third is the company later on: months from now, what remains of this period is the sentences in this document, not anyone's recollection of it.

Conversations during the window run as one-on-ones — the recurring private meeting between manager and employee — but tightened to weekly, with the agenda pinned to the items in the plan. What separates them from ordinary feedback, the running exchange about what went well and what should change, is that they get written down. Each week leaves a line recording what was checked.

When the deadline arrives, the judgment follows the document and nothing else. Other problems noticed along the way stay out of this decision. Add to the standard mid-window and the person is being held to a target that moves, which removes the reason the document was written. The outcome is also recorded in the next performance review.

"We need to decide this week whether this goes to a plan or gets one more round of feedback."

Drawing the line between what a conversation can still fix and what needs a document.

"Nothing outside the criteria in this document gets used to judge the result."

A commitment that the bar will not rise during the window.

"Friday, 30 minutes, these items only."

The check-in cadence written into the plan is being kept as written.

The item that took Grace longest the first time she wrote one of these was the second. Describing what was missing came quickly. Writing what would count as clearing the bar exposed the fact that the team had never held a standard at all. She ended up writing down the values the team actually lived by: review responses within a day, a weekly work log every Friday, and a reply to incidents on her own features within 30 minutes.

Sixty days later two of the three were met and one was not. Grace judged it exactly as the document specified, and nothing said in that meeting was new to either of them. The weekly checks had already left the whole record behind.

The most common reading is that a plan like this is a termination notice. A substantial share of people who receive one do end up leaving, so the impression is understandable. But the causation runs the other way. The document is the first attempt to write down a standard in a situation that had already deteriorated; it didn't create that situation. Read as a notice, it produces sixty days spent job hunting instead of improving, and the company never gets the improvement it wanted.

The other mistake is treating it as a period when only one person has work to do. That's why the support item is in the document. The manager also has obligations written in — checking weekly, supplying what's needed — and a window that closes with those obligations unmet leaves the judgment with nothing behind it. A plan with the support section blank isn't an improvement plan; it's a list of demands with a deadline on it.