How to evidence jobs and skills social value: from commitment to sustained outcome

If jobs and skills are becoming more important in procurement, what should you actually measure beyond the number of jobs or training places created?

That question is about to get sharper. The proposed January 2027 Social Value Model narrows central government's scoring onto employment and skills outcomes, and the Cabinet Office's March 2026 consultation response confirms the direction: where an authority sets social value criteria for jobs, opportunities or skills, it will be expected to set at least one measurable KPI, monitor it, and report performance through the contract. The emphasis is shifting from what a supplier promises at bid stage to what actually happens for real people once the contract is live.

Credible jobs and skills evidence should show what opportunity was created, who benefited, its quality and duration, what changed for the person afterwards, and what evidence connects that outcome to the contract. This guide sets out a practical framework for collecting that evidence, whether you are a buyer specifying commitments before award, or a supplier delivering and reporting them.

One note on timing. PPN 002 remains the current published Social Value Model for central government. The new model, PPN 026, applies to procurements commenced on or after 1 January 2027, with fuller guidance expected in autumn 2026. Much of what follows is grounded in criteria PPN 002 already contains: fair work, skills, and support for people facing barriers to employment. You can start improving your evidence now rather than waiting for the new guidance to land.

What counts as jobs and skills social value?

'Jobs and skills' is a broad label covering several different things, and they are not interchangeable. Adding a training session, an apprenticeship and a permanent job into a single 'opportunities' figure hides more than it reveals. It helps to separate them:

Employment opportunities: a new job created and made available, whether openly advertised, ring-fenced or filled.

Retained employment: an existing role safeguarded that would otherwise have been at risk.

Apprenticeships: a formal apprenticeship start combining paid work with structured training toward a recognised standard.

Work placements: time-limited, often unpaid or bursaried experience, such as a T Level placement or supported internship.

Training and qualifications: structured learning that may or may not lead to an accredited qualification.

In-work progression: movement within employment: more responsibility, more hours, higher pay, or a new skill applied.

Employability support: help toward work that is not itself a job: careers advice, CV and interview coaching, or a sector-based work academy.

Each of these is a legitimate contribution. But they sit at different points on the journey from 'no opportunity' to 'sustained, better work', and they carry very different weight as evidence. A completed apprenticeship with a qualification and a permanent role at the end is a stronger outcome than a half-day taster session, even though both can be described as 'skills'. The rest of this guide is about measuring that difference.

Start with the commitment you are trying to evidence

Good evidence starts before delivery, not after it. If a commitment is defined loosely, no amount of careful data collection later will make it measurable. Two questions are worth settling up front: who the opportunity is for, and what counts as delivering it.

Define who the opportunity is for

A job or training place carries more social value when it reaches someone who would otherwise struggle to access it. So define the intended beneficiary before delivery, not after. Useful dimensions include:

• age or target cohort where relevant, for example young people not in education, employment or training, or workers over 50 returning to the labour market;
• barriers to employment: care leavers, prison leavers, disabled people or the long-term unemployed, the groups PPN 002 already singles out;
• geography: a named local authority, ward or travel-to-work area;
• existing workforce versus new entrant: whether the opportunity brings someone new into work or develops someone already employed.

Being explicit here matters for two reasons: it lets you target effort where local need is greatest, and it makes the eventual claim verifiable. 'We supported X' means little unless X was defined in advance. Use tools like Place Explorer to understand local employment, skills and deprivation context so that target cohorts reflect the actual labour-market conditions of the place the contract serves.

Define what counts before delivery starts

Then decide what will and will not count as delivering the commitment. Settle these criteria before the contract starts:

• New job or retained job? Both are valuable, but they are different claims and should never be merged into one number.
• Full-time, part-time or variable hours? Record actual contracted hours, not headcount alone.
• Minimum duration? Does a role lasting three weeks count the same as one lasting two years?
• Directly employed or supply chain? A job created several tiers down is real, but attribution and evidence are harder.
• Contract-specific or organisation-wide? Did the opportunity arise because of this contract, or would it have existed anyway?
• Genuinely additional? See additionality, below.

Compare two versions of the same promise. 'We will create five jobs' is unmeasurable: five jobs of what kind, where, for whom, lasting how long, and how would anyone check? 'We will create five new full-time roles (minimum 12 months) on the contract site in [local authority], at least three ring-fenced for local residents currently unemployed, recruited within the first six months' can be planned, delivered, evidenced and audited. The second is a commitment; the first is a slogan.

Measure job quality as well as job quantity

A job count says nothing about whether the job is any good. The new model's emphasis on 'high quality jobs' and PPN 002's existing Fair Work criteria point the same way: quality is part of the outcome, not a nice-to-have. Three dimensions are worth capturing.

Pay

Record what people are actually paid, and be clear about the benchmark. Meeting the statutory National Minimum Wage or National Living Wage is a floor, not an achievement: it is a legal requirement. If you want pay to count as social value, measure against a more ambitious standard: the real Living Wage, the going rate for the role locally, or a defined uplift over the statutory minimum. Keep 'we comply with the law' and 'we pay well' clearly apart.

Security and working conditions

Quality also lives in the terms of the work. Fields worth recording include:

• contractual hours (guaranteed, or zero-hours);
• contract type (permanent, fixed-term, agency, self-employed);
• retention: is the person still in post after a defined period?;
• predictable working patterns and reasonable notice of shifts.

These map directly onto PPN 002's Fair Work criteria, so capturing them also helps evidence the current model.

Progression

The strongest employment outcomes are the ones that move someone forward. Track:

• qualifications gained;
• specific skills acquired and used;
• increased responsibility;
• movement into higher-paid or more secure work.

In-work progression is already named in PPN 002, so this section is grounded in current policy rather than speculation about the replacement model.

Go beyond training attendance

Training is where reporting most often overstates impact, because attendance is easy to count and outcomes are not.

Distinguish participation, completion and outcome

A single training figure can mean very different things. Picture the chain: training offered → training started → training completed → qualification achieved → skill used → progression.

Reporting the number at the left-hand end ('120 training opportunities offered') while implying impact at the right-hand end ('120 people upskilled') is one of the commonest ways social value reporting loses credibility. Say where on the chain your number sits, and report drop-off honestly. '120 offered, 90 started, 70 completed, 55 qualified' tells a truthful and more useful story than a single round number.

When training hours are still useful

Output metrics such as hours delivered, sessions run and places offered are not worthless. Early in a contract they may be the only data available; they show effort and reach; and for short awareness or safety training the 'output' genuinely is the point. Use them, but label them as activity rather than outcome, and don't let an impressive hours figure stand in for evidence that anyone was actually better off afterwards.

How do you evidence sustained employment?

'We got someone a job' is a weaker claim than 'we got someone a job they kept'. Sustained employment is where social value is really created, and it is also the hardest thing to evidence, because it means following up after the initial placement. A proportionate set of measurement points is:

• job start;
• 3 months;
• 6 months;
• 12 months, where the contract length and stakes make it proportionate.

The practical tension is between rigour and burden. Tracking someone for a year gives strong evidence but requires their consent, a way to stay in contact, and someone whose job it is to follow up. Chasing 12-month data on a small, short contract may cost more than it proves. Set the follow-up period to match the scale of the commitment, agree it up front, and be honest about attrition: people move, change jobs and drop out of contact, and a follow-up rate of 70% reported transparently is more credible than 100% that quietly assumes everyone stayed.

What evidence should suppliers retain?

Every claim should be substantiable, without over-collecting personal data. Two principles keep this proportionate. First, collect the minimum needed to substantiate the claim: a referral confirmation is usually enough to show someone met a target cohort; you rarely need their full history. Second, apply data protection and data minimisation from the outset: hold personal data lawfully, keep it only as long as needed, and evidence outcomes in aggregate or anonymised form wherever you can. In short, keep evidence requirements proportionate to the size and stakes of the contract.

We've distilled this into a one-page Jobs and Skills Evidence Matrix you can download and adapt for your own contracts. See the download at the end of this guide.

How much can the contract actually claim?

Not every job or qualification that appears during a contract was caused by it. Before a contract claims credit, three questions sharpen the number.

Additionality and deadweight

Would this job or training have happened anyway? If a supplier was going to hire five people regardless of the contract, counting those five as social value is deadweight: it inflates the figure without reflecting any change the contract caused. Count what is genuinely additional.

Attribution

Who else contributed? Someone who moved into sustained work may have been helped by a college, a jobcentre, a charity and the employer. The contract can reasonably claim a share, but claiming the whole outcome double-counts against everyone else doing the same.

Displacement

Could the opportunity have displaced something else? A 'local job' that simply moved employment from the next town, or an apprentice who replaced an existing worker, has created less net value than the headline suggests.

These are exactly the questions Social Return on Investment (SROI) is built to handle. You do not need a full SROI study on every KPI to take them seriously. Applying additionality, attribution and displacement as sense-checks keeps contract-level reporting honest. A full SROI analysis becomes worthwhile when you want to understand and value the wider outcomes a programme creates, not simply count them.

A practical jobs and skills evidence framework

Pulling the threads together, one reusable chain works for almost any jobs or skills commitment: commitment → person/place → activity → quality → evidence → outcome → duration → contribution.

Worked example (fictional). A facilities-management contract commits to 'three new full-time roles, minimum 12 months, ring-fenced for local residents who are long-term unemployed'.

• Commitment: three new full-time roles, 12-month minimum, local, long-term unemployed.
• Person / place: three residents of a named ward, each unemployed 12+ months, referred by a named employment programme.
• Activity: recruited and employed on the contract from month two.
• Quality: permanent contracts, guaranteed 37.5 hours, paid at the real Living Wage.
• Evidence: HR records and start dates; referral confirmations for cohort eligibility.
• Outcome: all three in post at 3 and 6 months; two at 12 months, one moved to a higher-paid role with another employer.
• Duration: followed up to 12 months.
• Contribution: additional to business-as-usual hiring; recruitment run jointly with the referral programme, so credit is shared, not claimed in full.

The headline 'we created three jobs' is true, but the chain is what makes it credible.

What should buyers put in place before contract award?

Most evidence problems are designed in at the buyer's end, before a single job is created. Five steps at specification and award stage prevent them:

1. Define terms. State what 'job', 'local', 'sustained' and 'training outcome' mean in this contract, so bidders and evaluators share one definition.
2. Agree reporting metrics. Decide the small number of metrics you will actually monitor. Under the direction confirmed in March 2026, set at least one measurable social value KPI where you set jobs, opportunities or skills criteria.
3. Specify evidence. Say what evidence you expect for each claim, and keep the requirement proportionate to the contract.
4. Allocate ownership. Name who is responsible for collecting, reporting and assuring the data on both sides.
5. Set review frequency and escalation. Fix how often performance is reviewed and what happens if commitments slip, increasingly reported through contract performance mechanisms rather than left to the final account.

What should suppliers do now?

You do not need to wait for the autumn 2026 guidance to prepare. A short readiness checklist:

• Separate your numbers. Stop reporting jobs, apprenticeships, placements and training under one 'opportunities' total.
• Define beneficiaries and additionality now, on live bids and contracts.
• Capture quality, not just counts: pay benchmark, hours, contract type.
• Build a light follow-up process for sustained-employment data, with consent, before you need it.
• Keep proportionate evidence for every claim, and mind data protection.

Remember that PPN 002 remains the current published central-government model, and much of the above already helps you evidence its Fair Work, skills and 'barriers to employment' criteria. For the broader picture, see our guide to evidencing social value commitments in public sector bids. This is preparation that pays off immediately, not only in 2027.

Keeping commitments, evidence and outcomes in one place

Managing these commitments across multiple contracts? See how SVE Connect helps commissioners and delivery partners keep commitments, evidence and reported outcomes in one traceable process, so the story from promise to sustained outcome holds together when someone asks to see it. Learn more about monitoring social value after contract award.