Manage a warehouse without spreadsheets

Let us clear up a misunderstanding first. The spreadsheet is not the enemy. It is a remarkable tool, and every organisation will always use one.

The problem begins when the spreadsheet becomes the system of record — when the true value of an indicator exists nowhere except in a file on one person’s computer, populated by hand, with calculation rules embedded in formulas.

Three symptoms signal this shift:

• two departments present different figures for the same thing, and the discussion is about the method rather than the action;
• one person goes on leave and management stops;
• no one can explain precisely how a figure is calculated, but everyone uses it.

This document explains how to move beyond that situation. It is not about tools: it is about definitions, responsibilities and sources.

A warehouse management team that tracks thirty indicators manages none of them. The useful number is around five — indicators that people genuinely review every week and that trigger decisions.

The other twenty-five do not disappear: they become diagnostic indicators, brought out when one of the five starts to drift.

An indicator deserves a place among the five if it answers the following four questions:

• What decision do we make when it moves? If the answer is ‘none’, it is a decorative figure.
• Who owns it? Not who calculates it — who is accountable for its value.
• Can it be calculated without manual work? If not, it will be abandoned in three months.
• Can its definition fit into one sentence that two people will understand in the same way?

Reliability. Stock accuracy, and above all the time taken to detect a variance. The latter is more actionable than the former: it shows whether the organisation learns.

Service. The proportion of orders fulfilled in full and within the promised lead time. Pay attention to the definition of ‘promised lead time’: the one communicated to the customer, the contractual one, or the one the warehouse set itself? Three different answers, three different indicators.

Productivity. Volume processed in relation to time worked. This is the most treacherous: it only makes sense with a constant scope. A drop in productivity may indicate deterioration — or simply a change in product mix.

Quality. Errors detected by stage, disputes and returns caused by an internal error. For obvious human reasons, this is the hardest indicator to report honestly.

Utilisation. Occupancy by zone, dock saturation and fill rate. These are capacity indicators: they show when an investment decision is approaching.

This is the part everyone skips, and the only one that truly matters.

The anatomy of an indicator definition

Take an ordinary example: service level.

Worded that way, it means nothing. A usable definition specifies seven elements:

Two companies in the same sector can display very different service levels with exactly the same actual performance. The difference lies in these seven lines.

A half-page document for each indicator that answers the seven questions above plus three others:

INDICATOR · [name]

ONE-SENTENCE DEFINITION ·...

CALCULATION · numerator / denominator, with exclusions

SOURCE · where the data comes from, which system is authoritative

FREQUENCY ·...

OWNER · [role] is accountable for the value

ASSOCIATED DECISION · when this indicator drifts, we...

HISTORY · value measured on [date]:... (baseline)

LAST REVIEW · [date] by [name]

Five sheets are enough to start. They can be written in a half-day group session and will resolve half of the disagreements to come.

For each data item used in an indicator, ask one question: which system is authoritative?

Not ‘where can the data be found’ — several systems may hold it. The question is which one is right when they disagree.

This table can be completed in one meeting. The difficult cells — where two departments claim to be authoritative — are precisely the places where your figures diverge today.

Data entered twice will diverge. This is not a matter of discipline: it is arithmetic.

The practical consequence: any identified re-entry puts an indicator on borrowed time. If calculating your productivity requires a team leader to copy figures into a file every day, that indicator will die — at the first activity peak, the first holiday or the first reorganisation.

Take an inventory of re-entry before choosing your indicators. You will know which ones are sustainable.

What can be measured without installing anything

Many useful indicators require no system:

• Time to detect a variance — the difference between the date of the incorrect movement and the date of the count that revealed it.
• Truck turnaround time — arrival and departure times recorded in a logbook at the gatehouse.
• Incomplete-order rate — a manual count over two weeks, sufficient to establish a baseline.
• The share of orders picked after the cut-off time — visible in any management system, however old.

Start with those. They provide a baseline before any project, and that baseline is invaluable when the time comes to assess the result.

A dashboard sent by email is not management. It is publication.

Management requires a short, regular moment when people review the figures together and make decisions. Thirty minutes a week is enough, with three questions:

• Which of our five indicators moved?
• Do we know why?
• Who will do what by next week?

This ritual has an unexpected virtue: it quickly reveals useless indicators. A figure reviewed every week without ever prompting action removes itself after a quarter.

The trap of an indicator used to assess people

As soon as an indicator is used to assess someone, it stops being reliable. This is not a moral issue: it is a rational and universal reaction.

A concrete example: measuring picking errors by picker produces a spectacular fall in reported errors within a few weeks. Actual errors do not change — they are simply corrected discreetly before anyone sees them.

The rule: indicators measure processes; managers assess people. Both are done, but not with the same tool.

Review once a year

Indicators become outdated. Operations change, and so do priorities.

Hold a one-hour annual review, with three questions for each indicator: is it still useful? Is its definition still correct? Is someone still accountable for it?

That is a small effort, and it prevents fossilised dashboards being displayed out of habit.

THE FIVE INDICATORS WE RECOMMEND AS A STARTING POINT

They do not suit everyone — but they are a defensible starting point for most warehouses.

1. Time taken to detect a stock variance

Why: it shows whether your organisation learns from its errors.

Associated decision: if it increases, review the frequency or targeting of counts.

2. Proportion of orders fulfilled in full

Why: this is what the customer experiences.

Associated decision: if it falls, work back through the chain — availability, allocation, picking.

3. Vehicle turnaround time at the dock

Why: it reflects your absorption capacity and is the first sign of saturation.

Associated decision: if it drifts, choose between better time-slot organisation and more physical capacity.

4. Error rate detected during checking

Why: it measures quality upstream, not complaints downstream.

Associated decision: if it rises, look for the cause by zone and item — never by person.

5. Occupancy by zone

Why: it anticipates layout and investment decisions.

Associated decision: above a threshold that you set, review slotting before considering more floor space.

SELF-DIAGNOSIS SHEET

Seven ‘yes’ answers or more: your management system is robust; a tool will save you time without changing the fundamentals.

Between four and six: start with definitions and responsibilities. It costs nothing and delivers the greatest return.

Three or fewer: do not look for a tool yet. Choose three indicators, write their definitions and name their owners. Repeat this diagnosis in three months.

To conclude

An information system does not create management discipline. It makes it possible without effort — which matters, because effort is what kills initiatives.

But an automated dashboard built on vague definitions produces incorrect figures faster. That is the only real risk of digitising management, and it is entirely avoidable: define first.