THE QUICK READ

The story in brief

A percentage can be mathematically correct and still leave a reader with the wrong impression. Before a startup publishes a growth announcement, attach the metric definition, baseline, comparison period and source. The result is a claim an editor can understand and investigate.

Source starting point: Office for Statistics Regulation: Short guide to Standard Nine, Be Clear — 29 October 2025 ↗. Full sources and context appear below.

Make the claim complete before making it short

The Office for Statistics Regulation’s Be Clear guide emphasises context, useful comparisons and preventing misinterpretation. Published on 29 October 2025, it addresses official statistics. The communication principle is useful more broadly: readers need enough information to understand what a number means.

For a startup announcement, begin with a full internal sentence. Name the metric, its starting and ending values, the relevant dates, the population being counted and the source used. Only then decide which parts belong in the headline and which must stay immediately beside it.

A reader should not have to guess whether growth refers to sign-ups, paying organisations, transactions or usage. These measures can move differently. If the headline says customers, but the underlying count includes everyone who created an account, change the wording or the calculation before publication.

Show the base behind the percentage

Consider a fictional startup whose paying business accounts rose from 40 on 31 March 2026 to 60 on 30 June 2026. The increase is 20 accounts. Dividing that increase by the starting count of 40 gives 50% growth.

A useful announcement sentence would say: “Paying business accounts increased from 40 to 60 between 31 March and 30 June 2026, a rise of 50%, according to the company’s billing records.” This is an illustrative sentence, not a report about a real company.

The percentage and the counts answer different questions. The percentage expresses the change relative to the starting point; the counts reveal its scale. Keep both available. If the starting count is zero, this percentage-growth calculation is undefined. Report the movement from zero to the new count rather than inventing a percentage.

Do not confuse percentages with percentage points

The ONS content style guide distinguishes a percentage change from a difference measured in percentage points. That distinction matters when a founder reports a conversion rate or another proportion.

In a fictional example, a conversion rate rises from 10% to 15%. The increase is 5 percentage points. Relative to the original 10% rate, it is a 50% increase. Writing that conversion rose by 5% would describe a different change.

Include the underlying counts where they help interpretation. If one period had 10 conversions from 100 eligible attempts and the next had 30 from 200, readers can see both the rate and the changing volume. Define an eligible attempt consistently. A shift in who enters the denominator can alter the rate without establishing that the product itself improved.

Match the period and the counting method

Choose comparison windows that answer the question the announcement is making. A partial month and a complete month are different observation periods. A total accumulated since launch and the number active this month are different measures. Do not let a convenient label hide either distinction.

Record the exact cutoff dates and how the count was produced. For an activity measure, explain the action that qualifies someone as active and the window in which it must occur. For an organisation count, decide whether several team members belong to one organisation or several accounts.

If the measurement method changed, investigate whether the earlier figure can be calculated on the same basis. If it cannot, explain the break instead of presenting the numbers as a clean trend. Keep a record of exclusions and adjustments so that a later question can be answered from evidence rather than memory.

Keep survey findings within the survey

The CIPR/MRS/RSS guide to statistics in communications recommends making survey context available, including who commissioned it, who was represented, sample size, selection method, fieldwork dates and the numbers behind percentages. It also recommends defining methods in the release or accompanying notes.

For a founder, a customer survey and a product-usage count need different explanations. A survey of respondents cannot automatically describe every customer, every business in a city or an entire industry. Identify who answered and avoid expanding the conclusion beyond the group the evidence supports.

Put the essential limitation near the claim. A broad headline does not become precise merely because a distant footnote names a narrow sample. If you cannot explain who the result represents in ordinary language, resolve that question before turning the result into an announcement.

Separate a result from an explanation

A product change followed by higher usage gives the team a sequence of events to investigate. By itself, that sequence does not establish that the change caused the increase. Promotions, customer mix, seasonality or changes in measurement may also matter.

Use language that matches what was observed. The team can describe when it shipped a feature and what the measured activity was before and after. A stronger causal claim needs evidence suited to that question. Do not replace an uncertain explanation with a confident founder quote.

Keep future goals separate too. A target for the next quarter is not a result for the last quarter. Label forecasts and targets clearly, retain their assumptions, and avoid combining them with achieved figures in a headline that makes the whole statement sound completed.

Give an editor a reproducible evidence note

Before release, have someone outside the drafting process reconstruct the headline calculation from the approved figures. Ask them to explain the claim in their own words. A mismatch between their interpretation and the intended meaning is a reason to revise the sentence.

Save the metric definition, dates, aggregate inputs, calculation and data owner together. A short public methodology note can explain the basis without exposing individual customer records. Keep the headline, release body, chart caption and social copy tied to the same approved claim.

The final test is straightforward: could a reader say what changed and what the statement does not establish? Useful context does not weaken a meaningful result. It lets the audience judge that result on the right basis.

YOUR WORKING NOTES

Audit one number before sending the release

Use these questions to investigate the idea in your own context.

  1. Write the full metric definition and who or what is counted.
  2. Record starting and ending values with exact comparison dates.
  3. Recalculate the percentage and distinguish percentage points where relevant.
  4. Check whether the population or measurement method changed.
  5. Ask a second person to interpret the headline and verify it against the evidence note.

A closer look

Is 50% growth enough information for a headline?

It needs a named metric and comparison period, with the starting and ending values readily available. The same percentage can describe very different scales.

What is the difference between percent and percentage points?

Percentage points express the difference between two percentages. A rate moving from 10% to 15% rises by 5 percentage points, or 50% relative to its starting rate.

Are the numerical examples real startup results?

No. All figures and company scenarios in this guide are fictional teaching examples. Replace them with verified figures and definitions before publishing your own announcement.

Sources & editorial note

Original Founderline editorial guidance researched on 18 November 2026 using current ONS guidance, the Office for Statistics Regulation’s guide published on 29 October 2025, and the CIPR/MRS/RSS guide to statistics in communications. These are reference materials, not new announcements today. Every company scenario and numerical example below is fictional. The practical workflow is Founderline analysis; official-statistics guidance is used as a communication reference, not presented as a rule binding startups.

See our editorial standards or request a correction. For a useful comparison, read When your startup announcement is wrong: how to correct the record. Explore the founder topic guides for connected reading paths.