Numbers a Judge Can Check

No affiliate links on this site, and no competition is named here as good or bad.

A figure earns points when a judge can tell what it measures, over what period, against what. Without those three, it is decoration. For additional operational context, see the overview here, offered here only as a separate reference.

What a usable figure contains

The measure. What was counted, precisely. "Customers" is ambiguous; "customers who ordered at least once in the year" is not.

The period. A start and an end, or a duration. Related industry context can be found at Wikipedia.

The comparison. Against the year before, against the position at the start, against a sector figure where one exists.

And the source. Internal records are fine and should be labelled as such — "from our own booking system" — because a bare number with no provenance is discounted.

Four elements, one sentence, and it scores where a paragraph of assertion does not.

The four that undermine

Percentages without a base. "Up 300%" from three to twelve is arithmetic. State both numbers and let the percentage follow, or omit it.

Round figures. "Over 10,000" reads as an estimate; 10,412 reads as a count, and the precision is itself evidence that somebody measured.

Cumulative totals presented as annual. "We have helped 5,000 people" over eleven years is a different claim from the same figure in one, and a judge who works it out has found an inflation.

And unattributed causation. State what happened and let the judge infer why, which is more persuasive than claiming the link.

The numbers most entries could use and do not

Retention. How many customers, members or users stayed — frequently the most telling figure any organisation holds and rarely calculated.

Time. How long something took before and after.

Error or failure rates, which sound negative and demonstrate measurement.

Cost per unit of whatever you produce, which shows understanding of the operation.

And a figure from somebody else — a client's own reporting, a supplier's data — which is the strongest of all because it is not yours.

Where the numbers come from

Records that already exist and are not organised.

Invoices, job notes, tickets, the accounts, the calendar. The entry does not create evidence; it retrieves it — and how long that takes depends entirely on whether anybody was recording.

Which is why the first serious entry is much harder than the second, and why the useful preparation happens months before any deadline.

One page, updated quarterly: what changed, by how much, over what period, and where the figure came from. That page is the actual asset, and it serves proposals and pricing as well as entries.

What to do when the numbers are unflattering

Use them anyway, with the context.

A figure that went the wrong way, explained, scores better than silence — because silence in a section about results is read as an absence of results.

"Volume fell 12% while margin rose from X to Y, following a deliberate decision to..." is a stronger answer than an omission, and judges recognise the shape of a business that measures itself.

What does not work is selecting only the favourable period, which is visible to anybody who reads the dates.

Presenting several figures

A short table beats a paragraph of numbers, where the form allows it.

Judges scan tables. A row per measure, with before, after and the period, is read in seconds and scored immediately.

Where the form is plain text, one figure per sentence rather than three in a clause — a sentence containing four numbers is read as one impression rather than four facts.

And do not repeat the same figure across sections, which reads as thin rather than emphatic.

Rounding, honestly

Some rounding is normal and one kind is not.

Rounding to a sensible precision — 10,412 to "about 10,400" — is ordinary and loses nothing.

Rounding upward to a threshold — 9,600 to "nearly 10,000", then "10,000" in the next section — is where an entry starts to drift, and it is the mechanism by which honest people end up with figures they cannot support.

Write the real number in your notes and round only in the final text, which keeps the drift from compounding.

The figure you should not include

Anything you could not produce if asked.

Some programmes verify, some interview, and some ask for supporting documents at the shortlist stage — and a figure that cannot be evidenced at that point is worse than one never claimed.

The test before including any number: where would I get this from if somebody asked tomorrow?

Where the answer is "I would have to work it out", work it out now. Where it is "I could not", remove it.

Small organisations and small numbers

A common anxiety and mostly misplaced.

Judges score against the category, and a category for small organisations is judged on what small organisations achieve.

A figure that seems modest in absolute terms is frequently strong as a proportion — and stating both is the answer: "three of our eleven staff", rather than "three staff".

What reads badly is disguising the scale, which is transparent and invites the judge to assume worse than the truth.

Keeping the page

One page, quarterly, four columns: what changed, by how much, over what period, and where the figure came from.

Fifteen minutes each quarter, and it removes the hardest part of every future entry.

It is also the document that answers a client asking what you have achieved, a lender asking about trajectory, and yourself asking whether a year went well — which makes it worth keeping whether or not you ever enter anything.

The short version

• A usable figure has four elements: the measure, the period, the comparison, and the source — one sentence
• Four that undermine: percentages without a base, round numbers, cumulative totals presented as annual, and claimed causation
• 10,412 reads as a count where "over 10,000" reads as an estimate, and the precision is itself evidence of measurement
• The most telling figures are usually retention, time before and after, error rates, unit cost, and anything from a third party
• Entries retrieve evidence rather than creating it, so the preparation happens months earlier in whether anybody recorded anything
• Unflattering figures with context score better than silence, and selecting only the favourable period is visible from the dates