HR guides that end with something you can use

Calculations carried through, templates you can fill in, procedures with an owner.

• Planning summer leave without emptying the department — July and August concentrate half the leave. How to set minimum coverage per team, what tie-break rule you announce in January, and how to avoid the May argument.
• Unused leave days: seeing them in September, not December — In December it is too late: there is nobody left to cover and nobody wants to go. The threshold worth watching, how to show each person their position, and what to do with refusals.
• Leave balances brought in from a spreadsheet: what to check before loading — A bad balance import is discovered in December and can no longer be corrected without contradicting somebody. The nine checks, choosing the anchor date, and what to do with mismatches.
• A three-shift rota, without rebuilding it five times — A rota gets rebuilt five times because it starts in the wrong place. The right order of the six steps, what gets fixed first, and why requests are collected before, not after.
• Shift swaps between colleagues: the rule that stops the chaos — People swap shifts among themselves anyway. The question is not whether you allow it, but how you record it — and who answers if nobody turns up.
• Overtime: recording it so it does not become a person-by-person argument — Hours worked beyond schedule are remembered differently by everyone and argued about at month end, from memory. The flow that makes them visible on the spot.
• An employee’s first day: the checklist, split by who is responsible — A first-day checklist with no owner against each line does not get done. The four roles, what falls to each, and what must be finished before the person arrives.
• The probation period: making it mean something rather than just pass — A probation where nobody checks anything until the final week does not assess the person, it assesses the surprise. The three checkpoints and how the decision is made.
• Why people leave in the first 90 days: the signals from week two — A departure at three months is usually decided in the first two weeks. The observable signals, what to check for each, and why the exit interview comes too late here.
• What a training certificate actually proves — A PDF with a name and a date can be edited in two minutes. What turns a certificate into proof: public verification, a link to the record, and a server-side verdict.
• Mandatory courses nobody completes: what to change, in order — A low completion rate has five causes, and four of them have nothing to do with motivation. How to tell them apart and which intervention works for each.
• Refresher training: tracking the deadlines without a parallel spreadsheet — Training expires individually, on different dates for each person. How to build the deadline register, what alert threshold works, and why the monthly list has to be short.
• The first 90 days of a newly promoted manager — Somebody promoted from within has a problem an external hire does not: they have to lead people who were colleagues yesterday. The three-stage plan and what gets decided in month one.
• The weekly team meeting: an agenda that does not bore people — Most weekly meetings are a round of status updates that everybody listens to for two minutes out of thirty. The agenda that changes that, and the removal rule.
• Delegation: deciding what to hand on and what to keep — Delegation usually fails not because the wrong task was chosen, but because the handover was incomplete. The matrix that decides what you hand on, and the five steps of a handover that holds.
• The exit interview that produces information, not pleasantries — An exit interview run by the direct manager produces polite answers. Who should run it, what to ask, when, and why the answers only get read six months later.
• The employee survey: what you ask, how often and what happens afterwards — A survey with no follow-through does more harm than none. What dimensions to measure, what cadence a team can bear, and the three things you must do with the results.
• The job description that actually helps you recruit — A description copied off the internet brings candidates who apply to anything. The five-section structure, written from the real work, and the test that tells you whether it is useful.
• The careers page: what one people actually apply through contains — A careers page with three photos and “we are a young team” convinces nobody. The seven elements that matter, in reading order, and what drives good candidates away.
• The monthly HR report to leadership: what goes in, what does not and why — An HR report with twenty indicators does not get read. The one-page structure, the six figures that stay, and the commentary rule that makes the difference.
• How to calculate employee turnover (and why you have three different figures) — Turnover is leavers divided by average headcount. Three variants of the formula give three figures from the same data — and the difference between them tells you where the problem is.
• What a hire actually costs — The cost of a hire is not the agency fee. The nine lines that make up the real figure, a full calculation on a concrete case, and what is excluded so it stays comparable.
• Absence: the formula, and what the figure hides when it looks normal — The absence rate is absence days over planned working days. The formula is simple; what matters is what goes in the numerator and how you cut the figure to make it mean something.
• Headcount or FTE: when confusing them costs money — Headcount counts people, FTE counts full-time equivalents. They are two different figures from the same data, and using one for the other breaks budgets, plans and comparisons.
• Why your recruitment report shows almost zero on a busy stage — A report grouped by status shows a snapshot, not the period’s activity. How to read the difference, why the columns can exceed the total, and where the date filter goes wrong.
• Time to hire: how you measure it and where the delays really come from — Time to hire measured end to end tells you nothing actionable. Split by stage, it shows exactly where the days are lost — and usually not where everybody thinks.
• Twelve screening questions that remove pointless interviews — Good screening stops unsuitable candidates before they consume a manager’s hour. The 12 questions, how to set the pass threshold, and what you must not ask at this stage.
• The structured interview scorecard: building one for a specific role — A structured interview asks every candidate the same questions and rates them on the same scale. How to choose the six competencies, write the questions and shape the scale.
• How you reject a candidate: four templates, for four different stages — A rejection after a CV screen reads differently from one after a four-hour task. The four templates, the timing rule for each, and what you never write in a rejection.
• The review form: the scale, the criteria and why the overall average is useless — How you choose review criteria, how you describe each level of the scale so it means the same to every assessor, and why an average across five criteria hides what you needed to know.
• The 9-box grid: how to read each box and four mistakes that make it useless — The 9-box places people on performance and potential. What each of the nine positions means, what you concretely do with each, and the four mistakes that turn it into labelling.
• The skills matrix: from a list to a decision-making instrument — A skills matrix shows who can do what and where you are uncovered. How to choose the skills, what levels to use, and the three kinds of gap that call for completely different decisions.
• One-to-one questions, grouped by the situation you are in — A good one-to-one has three or four questions, not twenty, and the same ones each time. What to ask at the first, at a routine one, when somebody seems demotivated, and before a review.
• Keeping leave records without arguing with the employee in December — Year-end arguments about leave balances always have the same five causes. How to keep records nobody can contest, and what the employee has to see so they stop asking.
• The approval flow that does not jam when the manager is on leave themselves — An approval flow fails in exactly the busy periods, when the approver is away. How to build delegation, what deadline to put on each step, and when it escalates automatically.
• Planned and actual: the reconciliation that shows you where the hours go — The planned rota and the hours actually worked never match perfectly. The monthly reconciliation procedure, what to do with each type of difference, and what the pattern tells you.
• The rota as a photo in the group chat: what you replace it with — A photo of the rota sent to a group gets lost, gets changed, and never tells you who saw it. The three ways of sharing it, compared on what matters.
• A 30-60-90 onboarding plan: what goes in each stage — At 30 days the person should know how the place works, at 60 work with supervision, at 90 deliver alone. What goes in each stage, who owns it, and what a delayed month costs.
• The HR spreadsheet: exactly where it breaks, and at what headcount — A spreadsheet does not suddenly become unsuitable. It fails at six precise points, each with a different threshold — from leave records to change history. What they are and what to do.
• Choosing HR software: the question set for any vendor — Do not compare feature lists — they all look the same. The twelve uncomfortable questions to ask at any demo, including four that usually eliminate half the vendors.
• The leave request: what it contains and what it passes through — A leave request has five useful fields and three states. What it contains, who sees it at each step, and why the signed approval matters even if nobody asks for it.
• The team leave calendar: who sees it and what it shows — A shared calendar prevents half the clashes, but it also shows information that is not everybody’s business. What to display, to whom, and how to stop it becoming a comparison tool.
• What employee data you actually hold, and in how many places — Most companies do not know how many places hold their employee data. The five-category inventory, where it ends up without anyone deciding, and what to do with the copies.
• Who has access to salaries, and what is left behind when somebody opens them — Salary data leaks through files, not through systems. Who has a real reason to see it, how to limit access without blocking work, and why the log matters.
• What you do with an employee’s data after they leave — Access stays open for months after somebody leaves, and their data stays in places nobody checks. What you close on day one, what you keep, and what gets deleted.
• Who sees what in an HR system: the roles and the classic mistake — Permissions are usually granted by copying somebody else — which is how half the company ends up with access to everything. The four base roles and the rule that keeps them clean.
• The termination conversation: how you prepare it and how you hold it — The hardest conversation in HR usually goes badly because it was not prepared. What you prepare beforehand, the structure of the ten minutes, and what you never say.
• An employee’s last day: the checklist, by who is responsible — Departures are organised badly because nobody holds the whole list. The four roles, what falls to each, what happens before the last day and what stays open.
• Where AI helps in HR and where it has no business being — AI helps on tasks where a mistake is cheap and visible, and harms where a right follows from the result. The three-question test and the list by category.
• What you do not put in an AI chat, when you work in HR — The most frequent HR data incident of recent years is not an attack, it is a table pasted into a chat. What not to send, how to get the same result without personal data.
• Turning a slide deck into a course people actually finish — A 60-slide deck does not become a course by uploading it. What to cut, how to split it into modules, what to add, and how long it realistically takes.
• The end-of-course test: writing questions that measure something — Most internal tests measure short-term memory rather than understanding. The six question patterns that work, the ones that do not, and how to set the pass mark.
• A learning plan for one person: how to build it in 30 minutes — A learning plan with eight objectives never happens. How to pick the two that matter, where the material comes from, and how to check it was worth anything.
• The training budget: how to build it and how to defend it with figures — The training budget is the first thing cut, because it is the worst argued. The four-line structure, how to calculate the real cost of training, and which figure convinces.
• Paper timesheets: going digital without losing people along the way — The switch usually fails because everything changes at once. What you digitise first, what stays on paper for a while, and how you handle people without a phone.
• How to check a timesheet in five minutes, before you sign it — Nobody reads 1,500 cells. The seven controls that catch almost every timesheet error, in the order worth doing them, and what to do with what you find.
• The job advert that does not sound like every other one — Adverts get read in twenty seconds and almost all say the same thing. The structure that works, what to cut first, and one advert rewritten, before and after.
• The work sample: how long, what to ask for and how to score it — A task that is too long screens out the good candidates, who have other options. The length rule, what to ask, how to score, and why feedback is mandatory.
• Candidate sources: finding out which one brings people who stay — The source that brings the most candidates is rarely the one that brings the best. How to attribute correctly, what to measure beyond volume, and why the useful figure arrives a year later.
• The hiring manager’s interview: what they ask and what they stop repeating — The second interview usually repeats the first. What the manager should cover, what they receive from recruitment, and the four questions only they can ask.
• What to do with candidates who do not fit right now — Almost every talent pool is a CV graveyard. What makes one useful: three tags, a reactivation rule, and a six-month contact that is not spam.
• Objectives you can verify: from “better” to a threshold — An objective you cannot verify at the end is not an objective, it is an intention. The four-step conversion method and eight real examples, before and after.
• Critical feedback: how to give it without damaging the working relationship — Critical feedback usually fails from preparation, not from a lack of courage. The four-step structure, what you say in the first thirty seconds, and two full conversations.
• How to distribute work in a team with uneven loads — In almost any team, two people carry half the work. How to measure load without counting hours, why it happens, and how to rebalance without punishing anyone.
• Conflict between two colleagues: when to step in and in what order — An ignored conflict does not fade, it moves into the work. When to intervene, what to do before the joint conversation, and why direct mediation is rarely the first step.
• The signs of overload that show in your data, not in conversation — People do not say they are struggling, and by the time they do it is late. Five signals that appear in data you already hold, their thresholds, and what to do.
• How to give a team bad news — Bad news communicated badly does more damage than the news itself. The order in which people find out, the structure of the announcement, and what you never say.
• What you do differently when you lead a team remotely — Remote work does not remove communication, it removes accidental communication — which was the most useful kind. What has to be replaced deliberately, and what is not worth imitating.
• Why recognition programmes fail, and what holds up — Employee of the month dies within six months in almost every company. The three causes, what works instead, and why peer recognition beats top-down.
• A team’s unwritten rules: how to find them and when to write them down — Every team has rules nobody wrote and everybody follows. How to find them, which ones deserve writing down, which do not, and why new joiners learn them by getting them wrong.
• First-year retention: what you influence and what you do not — Half of early departures have causes you fully control, and the other half almost none. How to separate them and where the effort is worth spending.
• The personnel spreadsheet on somebody’s laptop — The most common HR data incident is not an attack, it is a file downloaded and forgotten. Why it happens, what risks it creates, and four measures that need no budget.
• AI CV screening: what to check before you rely on it — A model that ranks CVs is useful; one that rejects is risky, and the line between them is crossed without anyone noticing. How to check the score and what not to delegate.
• How to check a generated text, in two minutes — A generated text is credible even when it is wrong — that is the problem. The five controls that catch almost every error, and what cannot be checked at all.
• AI in assessing people: where the usefulness stops — A generated score cannot be explained to the person it describes, and a review that cannot be explained is not a review. What still helps, and the line not to cross.
• How to tell the team that someone is leaving — How you announce a departure tells the team how you will talk about each of them when they leave. The order, the four sentences and what you never say.
• The new hire’s mentor: how to choose one and what to ask of them — A mentor assigned on paper who does not know they are a mentor is the most common onboarding defect. How to choose one, what exactly to ask, and how much time to give.
• Onboarding in production and retail: what is different — An onboarding plan written for an office does not work on a shop floor: there is no email, no computer, and day one is a real shift. What changes, concretely.
• What a new hire gets in writing, in the first week — Beyond the paperwork, there are five things a new joiner should receive in writing and that are usually said out loud and forgotten. The list, and why each one.
• Lateness: what to measure, what to ignore and when it becomes a problem — Measured wrongly, lateness produces conflict without solving anything. What to count, what threshold makes sense, and why the pattern matters more than the minutes.
• Annual or continuous reviews: what you gain and lose with each — The debate is usually held in ideological terms. The useful comparison is on five practical criteria, plus the mixed model that works in most smaller companies.
• The manager who never finishes their reviews: the three real causes — The tenth reminder solves nothing, because the problem is not memory. The three real causes, how to tell them apart, and the right intervention for each.
• 360 feedback: who answers, who sees the result and what you do with it — A badly run 360 does more harm than not running one. Who is asked, how many, what is visible and how the result is handed back — with the rules that matter.
• When an employee asks to see what data you hold on them — The request comes rarely and, when it does, catches everyone unprepared. What to prepare in advance, what the answer looks like, and why it is simpler than it seems.
• Peer learning: the cheapest form, and the worst organised — Most of what people know they learned from a colleague, not from a course. How to organise that deliberately, without producing another programme that dies in three months.
• The internal trainer: how to choose one and what to ask of them — The best specialist is not automatically the best trainer, and the difference shows in the first session. The real criteria, what you give them, and how to avoid depending on one person.
• Hybrid work: what to measure when people stop coming in — Start time no longer says anything, and online presence measures who leaves the app open. What is still worth measuring, and what is recorded only out of obligation.
• Seasonal staff: what changes in your records and your planning — Seasonal workers mixed in with permanent staff distort every report. What to keep separate, what to prepare before the season, and why the order matters.
• Handover when someone leaves: the list with three columns — A handover done as a two-hour conversation is not retained. The three-column list, what to document in the right order, and what is lost anyway.
• When a key person leaves: the first 48 hours and the next three months — When someone load-bearing leaves, you get two different risks: an operational one, immediately, and a morale one, a few weeks later. What to do about each.
• How to hear what nobody tells you — The higher you sit, the less you hear — and not out of ill will. The four channels that produce real information, what each one can do, and why none is enough alone.
• Someone comes back to the company: what changes and what does not — A returning employee is usually treated as if they had never left — and that is the mistake. What changed while they were away, and what onboarding they need.
• What to ask a software vendor about your employee data — Nine questions that separate a prepared vendor from one improvising, with the answers that should worry you and the ones that are fine.
• A file sent to the wrong person: the first hours after a data incident — Most HR data incidents are not attacks — they are a file sent to the wrong address. What to do in the first hour, what to write down, and what does not help.
• Remote onboarding: what does not travel through a screen — Remote work loses exactly the information nobody ever teaches formally. What replaces what used to be absorbed in an office, and how the first days should run.
• When thirty people start at once: onboarding in cohorts — A plan designed for one person a month breaks at thirty. What happens in a group, what stays individual, and why a fixed start date solves half of it.
• Learning something new when you have no time: the 20-minute method — Courses fail not from lack of motivation but because learning has no slot in the calendar and no immediate application. What works, concretely.
• “They completed the course”: what the completion rate actually measures — A 94% completion rate can mean nobody learned anything. What each indicator measures, and the three that actually tell you something.
• Travel time: how to record it without inventing rules — Three travel situations that get confused with one another, how each is recorded, and why the unwritten rule always produces arguments.
• Moving your data to a new system: what to take and what to leave — Migrations rarely fail for technical reasons and almost always because nobody decided which version is true. The right order, and what is not worth moving.