The cost no one counts

Hiring

What does it really cost to say yes to a job offer?

Most offers are judged from one side. The employer asks whether the package fits the budget. Almost nobody asks whether it fits the life of the person accepting it.

In an illustrative Employer Report, a senior offer that looked affordable left the candidate almost nothing each month after housing, commute, school fees and debt. A modest rise in their living costs flipped the recommendation to reject, and the estimated chance of the hire leaving within a year was 38 per cent.

None of this means “don't hire”. It means hire with open eyes: adjust the package, the start date or the relocation support, or accept the risk knowingly and record why.

Try thisFor your next senior offer, list the candidate's likely fixed costs next to the net salary. If the gap is under ten per cent, have that conversation before signing.

Time

The one-day-in-five problem

Ask a team where its time goes and you'll hear “meetings” and “email”. Measure it properly and a different picture appears.

In an illustrative diagnostic of a 180-person claims department, about 22 per cent of working time went on four things: searching for answers nobody wrote down, redoing work after definition disputes, waiting for approvals, and meetings that read a dashboard aloud. Eight per cent was allowed as the normal cost of working together; the rest was drain.

The important finding wasn't the size. It was the cause. Almost none of it came from lack of effort. It came from things left unfinished upstream.

Try thisFor one week, ask your team to note each time they wait, search or redo. Group the notes by cause, not by person.

Knowledge

Your most important rule may live in two people's heads

Every organisation has a rule set that “just works”: how exceptions are approved, how a special case is priced. Often it works because one or two long-serving people carry it.

I measure this as a truck factor: how many people would need to leave before the rule can no longer be applied. In the sample, the exception rules had a truck factor of one, only five per cent of answers came from documentation, and new staff took seven months to reach full speed.

It's always reported for the asset, never for the people who hold it. One team's response: “Finally, someone's going to write this down so we can take leave.”

Try thisList your five most critical rule sets. For each, count who could apply it tomorrow without asking anyone.

Decisions

An open decision is not a free decision

Leaders sometimes delay a hard call to keep options open. The delay feels costless. It rarely is.

In the sample, a decision about outsourcing one function had been open for fourteen months. Two teams were staffing for both possible futures, the affected staff lived with uncertainty, and any later consultation would start from low trust.

The answer isn't always to decide immediately. Sometimes the right move is a small, funded option: a sixty-day design exercise with a clear review date and a trigger for reopening.

Try thisList every significant decision older than ninety days. Give each an owner, a date, and either a decision or a defined option.

Evidence

Why every number should say where it came from

A figure in a board pack carries authority whether or not it deserves it. A number someone typed in, a number from a published table and a number from a model all look the same on a slide.

Every figure I produce carries its pedigree: observed, declared, taken from a published source, or estimated. Estimates carry ranges. Even an audit trail marks each step as explicit, when someone recorded it, or inferred, when a system deduced it.

People stop arguing about whether a number is right and start asking how much weight it can bear.

Try thisIn your next management pack, add one small column to the key table: source.

Stakeholders

When the CFO and the shop steward read the same page

The CFO asks what something costs. The COO asks where the time goes. Employee relations asks whether a decision will stand at the CCMA. Organised labour asks whether workers will carry the burden. All four questions are fair.

One result, rendered for each chair, changes the conversation. In the sample, five of six stakeholder views ranked the same structural problem first, even though a purely financial ranking would have deferred it.

When everyone works from the same number, disagreement becomes about priorities, which can be discussed, rather than facts, which shouldn't need to be.

Try thisBefore your next contested decision, ask each stakeholder to rank cost, risk, people and long-term structure privately. Compare the rankings before the meeting.

Accountability

A good outcome is not a good decision

When something goes wrong, the search for who is to blame often starts before anyone has established what was knowable at the time. Choosing the forum that judges an event can quietly choose the verdict.

Five tests catch manufactured culpability: who chose the forum; whether the evidence standard was applied equally to everyone; when each piece of information became available; where it came from; and whether the question is being judged in the right domain at all.

Recording the frame before a decision, and what was knowable at the moment it was made, is what lets accountability land where it belongs rather than where it's easiest.

Try thisFor your next significant decision, write down three things before you choose: the frame, what you know now, and what you don't.

Culture

What gets paid is what is believed

Every organisation has two belief systems. One is written in the value statement. The other is written in the record of who gets promoted, whose objections get minuted, and what happens to someone after an honest mistake.

When the two disagree, people follow the second. Not out of cynicism, but because it is the one with consequences. Over time, those who acted on the stated values leave or learn, and the gap stops being visible from the inside.

The most telling moment is a failure. If a well-reasoned decision that went badly is treated the same as a careless one, the organisation is judging outcomes, not decisions, whatever it says about innovation.

Try thisTake one stated value. Find the last three occasions when honouring it cost something. Write down what the organisation chose.

Language

Words nobody chose

Some words arrive in an organisation without a decision. They come because respected peers use them, and soon they cannot be questioned without seeming difficult. They change what can be said long before they change anything that is done.

A useful word earns its place. It has a definition, someone who owns it, a decision it changes, and a way of telling whether it is working. A word with none of these is ceremony.

Try thisPick the three words that carry most weight in your strategy document. For each, ask which decision would have gone differently without it.

Human in the loop

The machines calculate. People do the thinking.

A general-purpose language model produces the most plausible next word. That's a reasonable way to draft a paragraph and an unsound way to price a hire, size a pay gap or estimate a year of rework. A board can't audit a sentence that merely sounded right.

So the line is drawn hard. Machines calculate, retrieve, propose and record. People frame the question, weigh what matters, decide and answer for it. No model sets a weight, asserts a constraint or issues a decision, and every machine contribution is logged as a proposal someone accepted or rejected.

Try thisFor one AI-assisted output you rely on, ask who accepted it, on what basis, and where that acceptance is written down.

One of three

Decisions are assets: govern them like capital

Boards govern capital with care. Every rand is budgeted, approved, recorded, audited and reviewed. Decisions receive almost none of that discipline, even though capital moves only because someone decided it should.

Ask most organisations three questions about a decision made last year and the answers fall away. What was known at the time? What alternatives were considered? Who had the authority, and when was it to be reviewed?

The Decision Governance Standard sets four conditions: integrity (sufficient, verified, traceable information), accountability (a named decision-maker, rationale, authority and review date), coherence (decisions don't quietly contradict each other) and learning (outcomes compared with intentions). The Decision Canvas scales with the stakes: a board resolution uses all twelve fields; a routine rule-governed decision records three.

Two of three

Clarity without surveillance

There's a fair fear about any tool that measures how work gets done: it can become a way of watching people rather than understanding systems. In South Africa, with its history and its labour law, that fear deserves a direct answer.

The answer is built into the instruments. Findings attach to systems, processes, decisions and roles. Time is sampled anonymously and reported only for groups of five or more. Diagnostic data is contractually barred from disciplinary, appraisal or retrenchment-selection use. Where labour is organised, a joint governance body holds audit rights.

The reward is trust, and trust is what makes measurement accurate. The findings are usually a relief: the friction was never people's fault, and now there's evidence to say so.

Three of three

Access is the point

South Africa's economy runs on organisations that have never had a strategy team: mid-sized manufacturers, family businesses, municipal entities, independent recruiters, schools, clinics and co-operatives. Most make their consequential decisions with the least support.

Three shifts change that. The method can be written down, so quality no longer depends on a large team. The instrument can be reused, so the second run costs a fraction of the first. And the standards can be open, using formats like DMN decision tables and YAML contracts, so nobody is locked into a black box.

A sixty-person business should see where its time goes, whether its offers are sound and whether its decisions can be defended, with the rigour a listed bank expects.