Data Literacy Apprenticeships: Standards & Funding

There is no apprenticeship called "data literacy". Search the term and providers will offer to sell you one, but Skills England, the body that approves apprenticeship standards in England, has never published a standard by that name. The capability you want, a workforce that can use data in its own work, is funded through named standards: Data Technician at Level 3 and Data Analyst at Level 4. This page names them, shows how the levy pays, and shows how to check that a programme is the real thing before you spend on it.

By James Cotton, Founder, iO-Sphere

Is there a data literacy apprenticeship?

No. "Data literacy apprenticeship" is search language, not a qualification. In England, every funded apprenticeship runs on a specific standard approved by Skills England, and the ones that build data capability across a general workforce are named things: Data Technician at Level 3, Data Analyst at Level 4, and a few adjacent routes for business analysis and governance.

So the label tells you little. When a provider offers you a "data literacy apprenticeship", ask which apprenticeship standard it runs on. A named standard, ST0795 or ST0118, is a real programme: funded to a set band, assessed independently, protected in working time. A name with no standard behind it is usually a library of modules wearing an apprenticeship label.

If it isn't "literacy", what are you building?

We call it essentials, and the word matters. "Literacy" carries a hint of remedial reading, which is the wrong posture for a capable professional who already reads a P&L or runs a rota. What a workforce is building is a set of foundational working skills, the way spreadsheets once were. The way we see the skill itself is a loop that runs in four steps: spot a problem in your own work, match it to a sensible method, do the analysis, then check the answer against what you already know to be true.

Most training teaches only the third step, the computation, which software now does for you. It skips the other three, which are the steps that make someone useful, and that is why so much data training washes off.

There is a reverse slogan too, that everyone must learn to code. It aims at the wrong target. Running the tools is where the work begins; the judgement that decides whether the work is any good is in choosing the question, matching it to a method, and knowing whether to trust the answer. None of that is programming.

The framing to drop is data-literacy-as-a-certificate: a library of modules and a multiple-choice test, sat by people whose real jobs never come up. It is easy to buy and easy to report on, and it rarely changes what anyone can do on Monday morning. Foundational skills are the output of doing real work; they are built in the doing itself. For the definitional groundwork on its own, our guide to what data literacy means is the companion piece.

Which standard builds the capability you want?

Two standards do most of the work for a general workforce.

Data Technician, Level 3 (ST0795) is the natural starting point for someone whose title has never said "analyst" but whose job touches data every day. It builds the foundations: collecting, structuring and cleaning data, running basic analysis, and communicating clearly what the numbers say. Its funding band is £13,000 and its typical duration is around 24 months.

Data Analyst, Level 4 (ST0118) goes further: working with SQL and Python, building analyses and visualisations in tools such as Power BI, and turning a business question into an answer someone can act on. Its funding band is £15,000, it carries 370 hours of off-the-job training, and its typical duration is also around 24 months. (Funding bands are set by Skills England and revised from time to time, so check the current band before you budget.)

A useful mental model when you are planning across a workforce: one standard covers many different job titles. You test a role against the standard's knowledge, skills and behaviours, its KSBs, not against the words on the door. Data now runs through most jobs even where no one's title says "analyst", so far more of your people fit an eligible role than the org chart suggests.

These standards have a ceiling. A Level 6 or 7 data science route, or a degree apprenticeship, is a different thing, and iO-Sphere delivers up to Level 5. A Level 4 Data Analyst builds toward data science and is a real step on the way, though it stops short of the destination; be clear with a candidate about which of the two they are getting.

What does the levy pay, and what does it cost you?

Apprenticeships in England are funded through the apprenticeship levy, and any start you plan now falls in the 2026-27 funding year, for apprentices starting from 1 August 2026. In plain money, for a start in England:

Apprentice aged 16 to 24 at the start: fully funded for any employer, whether you pay the levy or not, and even if a levy account is empty. You pay nothing toward the training.
Apprentice aged 25 or over, and you do not pay the levy: you pay 5% of the funding band; the government pays the other 95%.
Apprentice aged 25 or over, and you are a levy payer without enough in the account: you pay 25% of the band from your levy pot; the government pays 75%.

"Non-levy" is set by pay bill: an annual pay bill under £3 million, so plenty of mid-sized employers with large headcounts still count as non-levy and qualify for the co-invested rates. Such an employer can also receive a transfer of up to 50% of a larger employer's unused levy funds, which can cover the cost in full. Whatever the split, the apprentice is never charged for their own training, and the rules change each year, so confirm the current position on gov.uk before you commit.

Before you plan a start: eligibility

Two eligibility points catch employers out, so check them per role before you commit.

English and maths. A start aged 16 to 18 must reach Level 2 in English and maths to complete the apprenticeship, unless they already hold an accepted equivalent. Since 11 February 2025, that requirement has been optional for a start aged 19 or over, and it stays available and funded for anyone who wants it.

Prior learning. Funding does not pay twice for skills someone already has: where an apprentice already holds part of what the standard covers, the training content is reduced for that prior learning. So a role where the person already holds much of the standard, or a qualification that heavily overlaps it, is worth checking before you enrol them.

Should you use the short "units" instead?

The levy now also funds short apprenticeship units, roughly 30 to 140 delivery hours. It is tempting to read that as a cheaper way to reach more people, but 30 to 140 hours is too little time to build capability that changes how someone works: it funds exposure, not fluency. iO-Sphere does not deliver units.

A unit does have a place: an employer who needs a narrow, fully-funded intervention and cannot fund anything larger. That is a real and specific use. Building lasting data capability across a workforce is a bigger job, and for that the route is a full apprenticeship.

How do you tell a real programme from a content library?

The funding route says nothing about whether the delivery is any good, so inspect three things before you commit.

First, the standard. A real programme runs on a named, approved standard, ST0795 or ST0118, with a published set of knowledge, skills and behaviours behind it. A "data literacy apprenticeship" with no standard named is the warning sign.

Second, how a learner spends the time. The model to avoid is theory-first lecturing by tutors who have studied the subject more than they have practised it. Ask what your people will work on week to week, and who is coaching them.

Third, the assessment. For these standards the end-point assessment is a portfolio and a professional discussion, judged by an independent organisation on real work, with a written knowledge test only where the standard's plan requires one. Ask a provider to show how it builds toward that.

Why data and AI initiatives fail

Getting this right matters, because most data and AI investment does not deliver and the pattern in why is consistent. Three independent studies, each measured separately:

MIT NANDA's report The GenAI Divide: State of AI in Business 2025 found that 95% of enterprise GenAI pilots delivered no measurable profit-and-loss return, despite between $30 billion and $40 billion of investment, and named the core barrier as learning, not infrastructure (fieldwork January to June 2025).
RAND (report RRA2680-1, August 2024) found that more than 80% of AI projects fail, about twice the rate of non-AI IT projects, with the leading root cause a misunderstanding of the problem to be solved.
S&P Global Market Intelligence's Voice of the Enterprise: AI & Machine Learning Use Cases 2025 found the share of firms abandoning most of their AI initiatives rose from 17% to 42% in a year, and that the average organisation scraps 46% of its proofs of concept (fieldwork October to November 2024).

These are international enterprise figures, not UK ones. Read together, they point one way: what limits the return on data and AI is adoption inside the workforce, and adoption is a capability problem. A capability gap closes when people learn to do the work. More tools on their own do not close it. Which kind of training does that, and whether it needs to be a full apprenticeship or something shorter, depends on the need.

Which route fits your workforce

The way iO-Sphere builds the loop is on Prism, a simulated business we run for the programme. Learners tackle real analysis problems inside Prism, working the methods and judgement calls the job demands, in a place where a wrong query costs nothing and the problems are not toy versions. That is how someone practises all four steps, including the three a video course skips.

Match the standard to the capability you need built, and the route follows.

Foundational data skills across roles that already touch data. The Data Technician standard (ST0795) is the entry point, and it is our Data & AI Essentials programme. This is where most first-time learners should start. We deliver it as roughly 15 months of training plus a 3-month end-point assessment.
Roles whose core work is analysis. The Data Analyst standard (ST0118) is the next level up, and it is our Advanced Data & AI programme, delivered on the same 15-months-plus-assessment shape.
Planning across a whole workforce. Our employer page walks through mapping roles to standards and the funding routes, and it is the right place to start if you are working out the mix rather than a single hire.

If the need is broad awareness across many people quickly rather than deep capability in a few, an apprenticeship is the wrong tool for it, and shorter data and AI fluency training fits that better. Let the shape of the work decide.

FAQ

Is there a "data literacy apprenticeship" standard in the UK?

No. Skills England approves the apprenticeship standards used in England, and none of them is titled "data literacy apprenticeship". Employers reach that outcome through named standards, most often Data Technician (Level 3, ST0795) and Data Analyst (Level 4, ST0118). If a provider markets a "data literacy apprenticeship", the useful move is to ask for the standard code it delivers against.

Which standard should I fund for a workforce that needs to use data?

Match the role to what it needs to do. If the role mainly needs to gather, organise and make sense of data and explain it clearly, the Data Technician standard (Level 3, ST0795) fits. If it needs to query, model and turn business questions into analysis with tools such as SQL, Python and Power BI, the Data Analyst standard (Level 4, ST0118) is the level up. Many roles that never carried the word "analyst" still fit one of the two.

What does an apprenticeship cost the employer under the levy?

For a start in the 2026-27 funding year, from 1 August 2026, an apprentice aged 16 to 24 is fully funded for any employer, even on an empty levy account. For a start aged 25 or over, a non-levy employer pays 5% of the funding band and a levy payer without enough in the account pays 25%. The full split, and what "non-levy" means, is in the funding section above. Confirm the current rules on gov.uk before budgeting.

Should I use the short apprenticeship units instead?

Only for a narrow, one-off need. A unit runs roughly 30 to 140 hours: enough to introduce a topic, short of building capability someone carries into their work. It suits an employer with a specific gap, a fully-funded budget for it, and no room for more. To build lasting data skills across a workforce, the full apprenticeship is the route, and iO-Sphere does not deliver units.

How do I tell a real apprenticeship from a rebadged content library?

Three checks. Confirm the named standard it runs on (ST0795 or ST0118) and read the knowledge, skills and behaviours it commits to. Ask what learners work on day to day and who coaches them. And check that the end-point assessment judges real work through a portfolio and a professional discussion. A programme that can answer all three is the one to back.

How long does delivery take?

A standard's typical duration and a provider's delivery time are two different numbers. Both the Data Technician and Data Analyst standards carry a typical duration of around 24 months. iO-Sphere delivers each as roughly 15 months of training followed by a 3-month end-point assessment. That training includes protected off-the-job training inside paid working hours, which the apprenticeship requires; ask any provider how they schedule it.

Building data capability across your workforce? See how funded apprenticeships map to the roles you already have: talk to us about your team.