Why direct employers need a middle path
Inside a company with an HR department, declining performance triggers a known machine: documentation, a formal conversation, a plan with a name and a form. When you hired a remote assistant directly, you are the machine. There is no HR partner to draft the document, no policy binder to consult, and no agency account manager to have the awkward conversation for you. So the awkward conversation tends not to happen. The employer notices the slipping quality in week one, mentions it vaguely in week three, and by week eight is browsing job boards while still paying for work they no longer trust.
That drift is expensive in a way that is easy to underestimate. By the time an assistant is six months in, you have paid for sourcing, interviews, a skills test, onboarding, and months of accumulated context about your business that lives nowhere else. Replacing them does not just cost the new search; it costs re-teaching everything the current person already knows. A thirty-day structured attempt to fix the problem is almost always cheaper than a restart, even when it fails, because at minimum it converts a vague dissatisfaction into a clear, documented reason to part ways, and at best it saves the hire entirely.
There is a fairness argument that matters just as much as the arithmetic. Remote work strips out the ambient feedback an office provides: nobody drops by an assistant's desk to say the reports have been sloppy lately. It is genuinely common for a remote worker to have no idea their standing has slipped, because every signal they receive, tasks still arriving, invoices still paid, says things are fine. A written plan is the honest correction to that silence. It tells the person exactly where they stand, exactly what would fix it, and exactly when they will hear about progress, which is everything the silence was denying them.
The anatomy of a plan that can actually be passed
Start with concerns specific enough to act on. There is a world of difference between communication needs to improve and the weekly report has been late or missing three of the last four weeks. The first is a mood; the second is a fact with a date on it, and facts with dates can be fixed and verified. Before writing the plan, gather your examples: which tasks, which dates, what impact. If you cannot produce concrete examples, that is a signal worth heeding, because it usually means the real problem is your own unstated expectations, and the document you need first is a clearer brief, not a PIP.
Expectations are the mirror image of the concerns, and the test for each one is simple: at the end of the plan, could you and your assistant independently agree whether it was met? Response times, delivery days, error rates you can count, checklists completed; these pass the test. Show more initiative and care more about quality do not, and every expectation that fails the test is a future argument stored up for the final review. The generator above ships measurable example standards for eight common assistant roles precisely because writing these from scratch is the step where most plans go vague.
The support section is what separates a plan from an ultimatum, and it is the section employers most often skimp. If the work has been going wrong, some of the fix usually lives on your side of the screen: clearer briefs, written procedures for the recurring tasks, faster answers to questions, paid time to practice a weak skill. Promise only what you will deliver, and then deliver it, because an employer who misses their own commitments in the support section has forfeited the plan's moral authority by the midpoint review. Finally, state the outcomes honestly, including the ending nobody enjoys writing down. The person on the plan already knows that ending exists; naming it in writing is not cruelty, it is the respect of treating them like an adult who deserves to know the stakes.
The remote factor: diagnose before you prescribe
With a remote hire, an apparent performance problem is regularly a systems problem wearing a performance problem's clothes. The assistant who keeps getting tasks wrong may be working from briefs that a colleague sitting next to you would have clarified with three quick questions across the desk. The one who seems slow may be losing a full day to every question because your working hours barely overlap and answers arrive after their day has ended. Before the plan blames the person, it should audit the system around them, which is why the generator includes a remote working factors section and switches it on by default.
The practical checks are unglamorous but decisive. Are the task briefs actually clear, or do they assume context that lives only in your head? Is there a real overlap window where questions get answered in minutes instead of days, and is it being used? For a South African assistant, the time zone maths is usually favourable, South Africa runs on UTC+2 with no daylight saving, so the overlap calculator will show most US and European employers a workable shared window; the question is whether your working rhythm actually exploits it. Is the internet and power setup holding, and is there a backup plan for outages? And is the workload itself sane, or has scope quietly grown past the paid hours?
Running this diagnosis inside the plan, at the first check-in, does two useful things at once. When a system problem surfaces, you fix the actual cause and the plan becomes easy to pass, which is the best possible outcome for everyone. And when the systems check comes back clean, the plan's legitimacy is strengthened: you have visibly ruled out the excuses, the support section addresses the real gaps, and what remains is a genuine performance question that the expectations will now measure fairly. Either way, the conversation stops being an exchange of impressions and becomes an inspection of specifics.
Running the plan: the month that decides it
Deliver the plan in a conversation, never by email alone. Send it shortly before a call so it is not an ambush, then walk through it together: the concerns with their examples, the expectations one by one, the support and what it will look like in practice, the schedule. Invite pushback and mean it; if an expectation is genuinely unrealistic, it is far better to hear that on day one and adjust than to discover it at the final review. The acknowledgment signature at the bottom of the generated plan records that this conversation happened, and it is worth saying out loud that signing means received and discussed, not I agree with every word.
Then the plan lives or dies on the check-ins. Short, on camera where possible, on the cadence you committed to, with notes in a shared document both of you can see. Cancelling a PIP check-in because a week got busy tells the person the plan was theatre, and you will not get their effort back after that. The midpoint review matters most: if progress is real but incomplete, say so plainly and keep going; if something is off track, adjust the support rather than silently letting the clock run. Nothing said at the final review should be new information, because every theme in it should have been raised in a check-in while there was still time to act on it. If you already run regular one on ones, fold the check-ins into them with the 1:1 agenda generator and give the plan its own standing agenda section.
Close the plan in writing whatever the outcome. If the expectations were met, say so unambiguously and let the document end; a plan that just fades out leaves the person permanently unsure of their standing, which corrodes exactly the confidence the plan was meant to rebuild. If progress was real but short, a single extension with the same support is often the right call. And if the honest conclusion is that it is not working, the plan has done its quieter job: the decision is documented, expected by both sides, and can be executed cleanly under the notice terms of your written agreement, followed by a proper handover with the offboarding checklist.
Contractors, employees, and keeping it fair
Most directly hired South African assistants work as independent contractors, and for a contractor the formal legal machinery of performance management mostly does not apply: the relationship is governed by the contract, which typically lets either side end the engagement with notice. That means a PIP for a contractor is not usually a legal requirement; it is a management choice, and the case for it is practical rather than statutory. You are protecting an investment, giving a person a genuine chance, and building a reputation as an employer worth working for in a talent market where word travels. The document can accordingly be lighter than a corporate HR template: clear, humane, and free of legalese.
The caveat that belongs in writing: employment status is a legal question decided by the substance of the relationship, not the label on the contract. If your assistant is, in substance or in law, an employee, then the rules change, and in some jurisdictions, including South Africa for its employees, dismissing someone for poor performance requires a specific fair process. The generated plan says plainly that it is a template and not legal advice, and that anything uncertain belongs in front of a professional. That sentence is not boilerplate; it is the honest boundary of what a free tool can responsibly do for you.
Fairness also has a plain-dealing dimension that no jurisdiction enforces but everyone can see. Do not put someone on a plan you have already decided they will fail; it wastes their month and your money, and the pretence shows. Do not stack the expectations with standards you never previously communicated, and do not move the goalposts mid-plan. The test worth applying to every line of the document is simple: if a fair-minded outsider read this plan and watched you run it, would they say the person got a real chance? When the answer is yes, every outcome of the plan, including the hard one, is defensible, explainable, and clean.
Where the improvement plan fits in your toolkit
The improvement plan is the missing rung between two tools this site already provides. The performance review generator is the periodic measurement: it tells you, on a rhythm, where the work stands. The offboarding checklist is the clean ending when an engagement runs its course. The PIP is what belongs between them when a review comes back poor: the structured attempt to fix before you conclude, which most direct employers currently improvise or skip.
Upstream, the best improvement plan is the one you never need, and the earlier tools in the kit are the prevention. A precise job description and a real skills test stop mismatched hires at the door. A proper onboarding plan and written SOPs prevent the unclear-expectations problem that a shocking share of PIPs are quietly trying to cure. And a standing remote work policy means hours, updates, and availability were agreed in writing before they could become grievances.
And if you are reading this page before you have hired anyone, take the meta-lesson: the employers who rarely need improvement plans are the ones who set expectations in writing from day one and check in on a rhythm. That is a learnable system, not a personality trait, and every piece of it is covered by a free tool on this page's list. Start with the hiring tools, keep the management rhythm, and let the PIP be the tool you own but seldom open.
Performance improvement plan FAQs
What is a performance improvement plan?
A performance improvement plan, usually shortened to PIP, is a short written document you use when the work of someone you hired has slipped below what you agreed and you want to fix it rather than end the engagement. A good one has five parts: the specific concerns with dates and examples, a clear statement of what good performance looks like, the support you will provide, a timeline with check-ins, and an honest description of the possible outcomes. It turns a vague sense that things are not working into a concrete, fair, two-sided plan.
How long should a performance improvement plan be?
Thirty days is the right default for most assistant roles: long enough to show a genuine change in work habits, short enough that neither side is living in limbo for a quarter. Choose 45 or 60 days when the expectations involve skills that take time to rebuild, such as a new tool or a quality standard that only shows up across many pieces of work, and reserve 90 days for senior roles with long feedback cycles. Whatever you pick, add a midpoint review; the worst plans are the ones nobody looks at until the deadline.
What should a performance improvement plan include?
Five things. First, the specific concerns: observed work, with dates and impact, not impressions or personality. Second, expectations written so clearly that both of you could independently agree whether each one was met. Third, the support you will provide, because a plan that demands change while offering nothing is a countdown, not a plan. Fourth, a schedule of short check-ins with shared notes. Fifth, the possible outcomes stated honestly, including the one nobody enjoys writing down. The generator on this page builds all five sections and lets you toggle the rest.
Is a PIP just a step before firing someone?
It has that reputation, and it earns it whenever a manager writes one after already deciding the outcome. Used honestly, it is the opposite: a structured attempt to keep a hire you have already invested in finding, onboarding, and training. The difference shows in the details. A real plan has support the manager actually delivers, check-ins that actually happen, and expectations a reasonable person could meet. If you have already decided, skip the theatre and part ways cleanly with proper notice; a fake PIP wastes weeks and poisons the reference you could otherwise honestly give.
Do I need a formal PIP for an independent contractor?
Legally, usually not: an independent contractor relationship is governed by the contract, and most contracts let either side end the engagement with notice. Practically, a light PIP is still often worth it. You spent real money and weeks finding and training this person; a structured month of clear expectations and support is usually cheaper than restarting the search. It also protects your reputation in the talent market you will hire from again. If your assistant is legally an employee rather than a contractor, local labour law may require a specific fair process, so take advice before acting.
Is this performance improvement plan generator free, and is anything stored?
Completely free, no signup, no watermark, and nothing you type is stored or sent anywhere; the whole tool runs in your browser. Pick the plan length and check-in cadence, write your specific concerns, keep or edit the role-specific expectations and support list, and copy the finished plan into a document you can share and sign. It is a starting template, not legal advice: your written agreement governs notice, and anything you are unsure about belongs in front of a professional.