A percentage can look precise while leaving the most important questions unanswered. To interpret a training outcome statistic, identify who was measured, what counted as success, when it happened and who checked the result.

Find the denominator

Every percentage is a fraction. If a provider says 90 per cent succeeded, ask 90 per cent of whom. The denominator might be everyone who enrolled, everyone who completed, everyone who answered a survey or a small selected group. Those populations can produce very different results.

Suppose 100 people enrol, 70 complete and 45 respond to a survey. If 36 respondents report a positive outcome, the result is 80 per cent of respondents, 51 per cent of completers and 36 per cent of starters. Each calculation is mathematically correct, but each answers a different question.

Define the outcome

Course completion, assessment pass, work placement participation, interview activity, job offer and sustained paid employment are not interchangeable. A headline should name the measured outcome rather than using broad words such as success or progression.

Ask whether employment must be related to the training subject, whether it can be part time or temporary and how long after completion it is measured. For placement figures, establish whether participation means starting, completing or merely being introduced to an opportunity.

Check the cohort and period

Results should identify the learners included and the dates they began or completed. A rolling twelve-month cohort may be more current than an all-time figure, but seasonality and programme changes matter. Results from one course should not be presented as if they apply to every course.

Look for a publication or last-updated date. A historic statistic may still be informative when labelled accurately, but it should not silently represent current performance.

Understand missing data

People who do not respond may have different outcomes from those who do. A survey response rate helps you judge this uncertainty. If 20 of 100 learners reply, even unanimous positive feedback cannot describe the unobserved 80 with confidence.

Ask how withdrawals, deferrals, unreachable learners and learners who decline to disclose employment are treated. Excluding them may have a reasonable methodological basis, but that basis should be visible.

Separate collection from verification

Self-reported outcomes can be useful, especially when privacy limits access to employment records. They are not the same as employer-confirmed or independently audited outcomes. A transparent report states how data was collected, what evidence was requested and whether an independent party reviewed it.

Independent verification should name the verifier, scope and date. A platform or software provider processing survey responses is not automatically auditing the accuracy of those responses.

Test comparisons

Two providers' figures should only be compared when the cohort, definitions, time horizon and collection method are sufficiently alike. An employment rate measured six months after completion cannot be directly ranked against an interview rate measured immediately after training.

Beware of tiny cohorts. If nine of ten learners report a positive outcome, one additional result changes the rate by ten percentage points. Show the count alongside the percentage so the scale remains visible.

A responsible presentation

A clear statement might read: Of 80 learners who completed Programme X between January and June, 60 answered a follow-up survey three months later. Thirty respondents reported starting paid work related to the subject. The result is therefore 50 per cent of respondents and 37.5 per cent of all completers, based on self-reported data.

That wording is longer than a slogan, but it gives a prospective learner what they need to make a reasoned judgement. If a provider will not supply basic definitions, treat the headline as marketing rather than decision-grade evidence.

Disclosure: Workplace Ready UK is owned by SRH CORP LIMITED, the organisation behind Excellent Pathways. We do not publish outcome rankings based on hidden criteria.