A start is an apprenticeship programme beginning. DfE counts starts across an academic year running from 1 August to 31 July.
The single most important thing about the measure: a start is not a person. DfE's own footnote is explicit – total starts are the count of apprenticeships started at any point in the period, and learners starting more than one apprenticeship will appear more than once.
So “308,770 apprenticeship starts” does not mean 308,770 people. If you want a count of people, the measure is participation, which counts learners and appears only once in the grand total. See starts, participation and achievements.
The nine-month trap
DfE publishes apprenticeship figures quarterly through the year, then finalises them. Current-year figures are provisional and cover only part of the year. The latest release covers August 2025 to April 2026 – nine months, not twelve.
This produces a specific and very common error: comparing a nine-month current-year figure with a previous full year and concluding that starts have collapsed. On that basis you would report 308,770 against 353,500 and announce a fall of 12.7%. DfE's own like-for-like comparison, against the same nine months a year earlier, shows starts up 8.7%.
The rule is simple: compare like periods, or use the change DfE publishes.
Other things to know about starts
- Age is measured at the start of the programme, not at any later point, and not academic age.
- Starts are rounded to the nearest 10, so component breakdowns may not sum exactly to the published total.
- A start is not a completion. Around 65.4% of those who left a programme in 2024/25 achieved it.
- Provisional figures get revised. When the full-year data arrives, the provisional figures are superseded, so any analysis should record which version it used.