Three levels of data apprenticeship, three kinds of work. The right one is the one the job can keep busy for hundreds of hours of training.
Data Apprenticeships in England: Level 3, 4 and 5 Compared
By James Cotton · Last updated · 7 min read
Part of our topic guides on Data & AI Apprenticeships and Government-Funded Data & AI Training.
By James Cotton, Founder, iO-Sphere
The three levels at a glance
On the Digital route of the Skills England register there is one standard for a data occupation at each of Levels 3, 4 and 5. The figures below are from the register, read on 1 October 2026.
| Level 3 | Level 4 | Level 5 | |
|---|---|---|---|
| Standard | ST0795 | ST0118 | ST1386 |
| Occupation | Technician | Analyst | Engineer |
| Work with data | Prepare | Analyse | Build |
| Funding band | £13,000 | £15,000 | £19,000 |
| Minimum off-the-job hours | 348 | 370 | 418 |
The funding band is the most the government will put towards training and assessment. How much of it an employer meets depends on the apprentice's age and whether the employer pays the levy, which applies to pay bills over £3 million.
For starts from 1 August 2026 the government funds the whole cost, up to the band, for 16 to 24 year olds at employers that pay no levy or have used up their levy funds.
For apprentices aged 25 or over it funds 95% where the employer pays no levy and 75% where a levy payer's funds have run out. A levy payer with funds pays from its levy account. Our guide to the 2026 funding changes works through each case.
Skills England prints the hours as "Minimum hours for compliance", and the DfE apprenticeship funding rules for August 2026 to July 2027 point to that figure as each standard's minimum off-the-job training. Those hours are where the choice of level is won or lost.
Choosing by the work in front of them
Our test for the right level is whether its training hours can be filled with real work in the person's current role. Pitch it too low and the hours go on things they already do well. Pitch it too high and there is nothing in the job to practise on, so the evidence the assessment asks for has nothing to draw from.
Each rung is its own standard, written for its own occupation, so nobody has to climb from the bottom. Someone who already analyses data and explains it is a Level 4 candidate; what decides it is whether their job holds that work now.
A level number says nothing about seniority. The right rung for a senior manager who is new to data can sit below the one for a junior colleague who writes queries every day, and the same team can have people on different rungs.
With us, one or two people can join an open cohort, so each person can go to the level that fits them; a closed cohort for one organisation is around 10 to 12 people.
Level 3: the person who gets data ready
A data technician might be the operations or finance administrator who keeps the records everyone else reports from: exporting from two systems, matching the rows, clearing duplicates, keeping the monthly spreadsheet or dashboard feed accurate. It suits someone whose job is mostly making data usable, and someone new to data whatever their grade.
We train this rung as the Data & AI Essentials (Level 3) apprenticeship, and the Level 3 guide goes deeper into the standard. Under its current version, assessment is a scenario demonstration with questioning and a professional discussion underpinned by a portfolio of evidence. The role has to produce a steady run of preparation work for that portfolio to draw on.
Level 4: the person who answers questions with data
Here the work starts from a question: why returns rose in one region, which campaign paid back, where a process loses time. The analyst finds the answer in the data and explains it to the people deciding. It suits someone already producing reports who should move from supplying numbers to interpreting them.
The data analyst standard is assessed by a project with presentation and questioning, and a professional discussion with a portfolio. The Level 4 data analyst guide goes through how that assessment works.
We train this rung as Advanced Data & AI (Level 4). The role needs one piece of analysis big enough to present and enough smaller ones to fill a portfolio.
Where the job is changing how a team works with AI, AI Transformation (Level 4) is the alternative at the same level, and our note on standards with AI in the frame helps place it.
Level 5: the person who builds what analysts rely on
The data engineer builds and runs the plumbing: the pipeline that loads sales data every night, the warehouse tables reports are built on, the checks that catch a broken feed before anyone reads a wrong figure. It suits someone in IT or analytics who already maintains databases, scripts or integrations and is ready to design them.
For new starts from 18 December 2026, a revised assessment plan requires a project and grades pass or distinction. Its six outcomes describe the job well: designing data products; acquiring and integrating data through pipelines; quality, governance and security; running and improving systems; sustainability and strategy; working with stakeholders. A role that touches only one or two of those will struggle to fill 418 hours.
We train this rung as Data Engineering (Level 5), and the Level 5 guide goes through the standard. Starts before 18 December 2026 are assessed under the current version, graded fail, pass, merit or distinction.
What the level number means
The number puts an apprenticeship on the same scale as other qualifications. GOV.UK's Become an apprentice page calls Level 3 an advanced apprenticeship, at the level of an A level, and Levels 4 to 7 higher apprenticeships, from foundation degree upwards; degree apprenticeships sit at Levels 6 and 7.
GOV.UK's list of what qualification levels mean puts a higher apprenticeship at Level 4 alongside the certificate of higher education (CertHE) and the higher national certificate (HNC). At Level 5 it lists the foundation degree, the higher national diploma (HND) and the diploma of higher education (DipHE), so a data engineer apprenticeship sits at the same level as those.
How many take each, and how many finish
In 2024/25 about 7,600 data technician and 7,140 data analyst apprenticeships started in England, against 440 data engineer, according to DfE's statistics release Apprenticeships, Academic year 2025/26.
DfE's national achievement rates for 2024/25, published March 2026, were 60.4% for data technician and 58.2% for data analyst, against 65.4% across all apprenticeships. Pass rates were 98.2% and 98.3%, and retention 61.5% and 59.2%, so nearly all of the distance from 100% lies in retention, not in the pass rate.
These are national figures for every provider together; comparing data analyst apprenticeship providers one by one is a separate job. DfE publishes no data engineer rate for 2024/25: only about 10 apprentices had left the programme.
Changes to assessment, by level
Each rung is at a different stage of what Skills England calls the 2025/2026 apprenticeship assessment reforms, and only one has a date.
- Level 5, from 18 December 2026: the revised data engineer assessment plan replaces the current one for new starts. The band, typical duration and minimum hours are unchanged.
- Level 3, no date: Skills England has published an approved revised assessment plan for the data technician standard, but has not yet set the date it takes effect; until it does, the current end-point assessment stays in place.
- Level 4, no date: the data analyst standard's record says it "is in the process of being revised or adjusted", and the revisions and adjustments status report has listed a revision to its assessment plan since 1 February 2026.
Above Level 5
At Level 6, the Data scientist (integrated degree) standard leads to a bachelor's degree in data science. At Level 7 sits the standard titled Artificial intelligence (AI) data specialist.
Skills England's Level 7 pages state that from 1 January 2026 these apprenticeships are government-funded only for people aged 16 to 21 at the start of training, or under 25 with an Education, Health and Care plan or who have been, or are, in the care of their local authority; earlier starters are funded to completion. The guide to levy-funded graduate programmes takes the rule further.
Frequently asked questions
Does the apprentice pay for the training, as a student would at university?
No. Paragraph 220 of the 2026 to 2027 funding rules bars the provider and the employer from asking the apprentice to contribute to the eligible costs of training or assessment, whether they finish or leave early. A new full-time student in England can borrow up to £9,790 a year in tuition fees for 2026 to 2027, a loan they have to pay back, according to GOV.UK's student finance page.
Can someone aged 25 or over do a data apprenticeship?
Yes. GOV.UK's conditions for starting are being 16 or over, living in England and not in full-time education, and DfE's figures for 2024/25 show 6,440 of the 7,600 data technician starts and 6,400 of the 7,140 data analyst starts were by people aged 25 or over. Age can change how much of the band the government meets, which the funding guide linked above explains.
Build a data-literate workforce
Data apprenticeships funded through the Growth & Skills Levy: £0 for any apprentice under 25, and most of the cost covered for older starts.