How to Track Apprenticeship Performance in 2026

The core problem: most apprenticeship metrics are lagging

Qualification Achievement Rate is the number the sector is judged on. It is also, by construction, a rear-view mirror.

QAR is calculated on leavers (those expected to complete in a given year) not on starts. By the time a cohort’s QAR is published, every decision that shaped it was made twelve to thirty-six months earlier. You cannot improve a published QAR. You can only improve the one being formed right now, quietly, inside your current cohorts.

So the practical question is not “what is our QAR” but “what is our QAR becoming, and which learners are moving it”.

Leading indicators worth tracking

These change while you can still act on them.

Engagement decay. Time since last meaningful contact, missed or rearranged reviews, declining submission frequency. Disengagement is rarely sudden. It is usually visible for weeks before a withdrawal request.

Off-the-job training shortfall. A learner falling behind on OTJ hours is both a funding compliance risk and a strong signal of a delivery problem. It compounds silently.

Review cadence. Not whether reviews happened, but whether they happened on time. Slipping review intervals across a coach’s caseload is often the earliest visible sign of capacity strain.

Gateway timeliness. Learners approaching planned end date without gateway readiness. Every week past planned end is funding at risk and a QAR outcome trending the wrong way.

Employer engagement. Changes of employer, redundancies and unresponsive workplace mentors are strong predictors of withdrawal, and they usually surface in the delivery team before they surface in any report.

Functional skills progress. For learners who need them, English and maths are the most common single cause of an otherwise-ready apprentice failing to achieve in year.

Lagging indicators worth tracking anyway

You still need these: just don’t manage on them alone.

• QAR: overall and by programme, employer, coach, cohort
• Retention and pass rates separately, because they fail for different reasons and have different fixes
• Withdrawal reasons, categorised consistently enough to aggregate
• Timely achievement versus overall achievement

The value in lagging measures is diagnostic. Segment them: a QAR that looks acceptable overall may be concealing one programme or one employer relationship performing badly enough to drag everything.

What changed for 2026

The inspection framework introduced in November 2025 organises evidence around Achieve, Belong, Thrive. That matters for performance tracking in three ways.

Achievement is no longer the whole story. Belonging and thriving are inspected alongside outcomes. A provider with strong QAR and weak learner experience is not automatically well-placed.

Evidence is expected to be current. The framework rewards providers who can show a continuous picture rather than one assembled ahead of a call. If your performance data only exists in a monthly pack, you are evidencing the pack, not the practice.

Report cards changed the presentation of judgements, which raises the value of knowing your own position precisely before anyone else forms a view of it.

Practically: tag your quality activity (learning walks, observations, case samples, learner conversations) to the framework principles as you go. Retrofitting that mapping later is slow and reads as retrofitted.

How often to look

Different questions have different natural rhythms.

The common failure is running everything at monthly cadence. Daily questions answered monthly are answered too late to matter, and annual questions answered monthly generate noise that gets ignored.

Making it act, not just report

Get from figure to name. A dashboard showing 12% at risk is interesting. The named list is actionable. If your reporting stops at percentages, someone is doing the real work in a spreadsheet.

Model the intervention, not just the risk. Knowing thirty learners are at risk is less useful than knowing which ten, if supported, would move the QAR most. Scenario modelling turns a risk register into a prioritised plan.

Make ownership explicit. Every at-risk learner should have someone accountable and a next action with a date. Risk lists without owners become wallpaper.

Close the loop. Track whether interventions worked. Without that, you repeat the ineffective ones indefinitely.

Where the data comes from

Nearly all of this already exists in systems you run: your e-portfolio or LMS holds progress and reviews, your MIS holds enrolment and funding, your safeguarding log holds concerns. The difficulty is almost never collection. It is that no single system sees all of it at once, so the joined-up view is assembled by hand, late, by whoever has time.

That joining is what a provider management system does. It doesn’t replace the e-portfolio your learners work in (AiVII integrates directly with Bud, Aptem and OneFile) it reads across them so the performance picture is a live state rather than a monthly reconstruction.

FAQ

How is QAR calculated? QAR measures the proportion of apprentices who achieve out of all those expected to complete in a given year. It combines retention and pass elements and is calculated on leavers rather than starts, which is why a year’s QAR is largely determined long before it is published.

What is the difference between timely and overall achievement? Overall achievement counts apprentices who achieve at any point. Timely achievement counts those who achieve by their planned end date. A large gap between the two usually indicates a gateway or end-point assessment bottleneck rather than a teaching problem.

How early can withdrawal risk be identified? Usually well in advance. Disengagement typically shows as declining contact, missed reviews and slowing submissions over weeks. The limiting factor is rarely the signal: it is whether those signals are visible together and early enough to act on.

What changed in the November 2025 Ofsted framework? Evidence is now organised around the Achieve, Belong, Thrive principles, so learner experience is inspected alongside outcomes, and judgements are presented through report cards. Providers are expected to evidence a continuous picture rather than one prepared ahead of an inspection.