Tendering for Tutoring After the NTP: A Practical Procurement Playbook for MATs and Schools

When the National Tutoring Programme ended, many schools were left with a familiar challenge: how do you keep intervention effective, compliant and affordable without the convenience of a central scheme? The answer is not to buy tutoring in the same way you bought it before. Post-NTP, tutoring procurement needs to be more rigorous, more evidence-led and more closely tied to school improvement priorities. Whether you are a single school leader, a trust central team, or a procurement lead in a MAT, you now need a clear tender guide that turns broad aspirations into measurable outcomes.

This playbook is designed to help you write a sharper tender specification, set realistic KPIs for tutoring, assess supplier claims, and compare the two models dominating the market: fixed-price AI programmes and per-hour marketplace tutoring. It also shows how to test whether a supplier can actually demonstrate pupil progress rather than simply attendance or satisfaction. For practical context on supplier variety and pricing, see our guides on the best online tutoring websites for UK schools, how to vet a marketplace or directory before you spend a dollar, and avoiding procurement pitfalls in martech, which offers useful lessons on vendor due diligence and contract discipline.

1. What changed after the National Tutoring Programme?

The NTP normalised tutoring, but also raised expectations

The NTP made tutoring mainstream for many schools by proving that targeted intervention could be delivered at scale. It also trained school leaders to look harder at service design, safeguarding, reporting and measurable impact. That legacy matters because buyers now expect more than a tutor who turns up and delivers a lesson. They expect alignment with curriculum, clarity on dosage, and data that can be explained to governors, trustees and parents.

In practical terms, this means procurement has shifted from “Can we get tutoring?” to “Which model gives us the best outcome for the lowest sustainable cost?” Schools need to consider whether a supplier supports reading catch-up, maths fluency, GCSE revision or post-16 transition. They also need to understand whether the service can operate flexibly across timetables, rooming constraints and exam windows. If you are balancing multiple priorities, our article on stakeholder engagement and governance offers a surprisingly relevant framework for managing different interests around a common objective.

Budget pressure has made value for money non-negotiable

Post-NTP spending is no longer shielded by a national wrapper. Schools and MATs must justify tutoring against other interventions, from reading schemes to teaching assistant deployment. That makes value for money more than a finance phrase; it becomes a performance question. A lower headline rate is not automatically cheaper if the programme requires extensive admin time, unstable staffing or weak progress evidence.

To think about value properly, compare total cost of ownership. That includes staff time to onboard, reporting burden, safeguarding checks, cancellation handling, and whether the model scales across year groups without new procurement each term. A platform with transparent pricing and consistent delivery may be more economical than a variable hourly arrangement that looks cheap until timetabling starts to fragment. For a broader lens on cost discipline, see build a budget in 30 minutes and pricing strategy lessons from Samsung.

Online tutoring is now the default for scale

The market has also changed operationally. As the source material notes, online tutoring has become the preferred format for pupils, parents, teachers and tutors, with a large majority of in-school tutoring now delivered online. That matters for procurement because online delivery broadens access to subject specialists and makes it easier to standardise quality assurance. It also introduces new questions around device access, platform reliability, safeguarding and session observation.

For schools in rural areas or MATs with dispersed sites, online provision can remove a major access barrier. But it only works well if the provider can evidence engagement and maintain continuity when pupils miss sessions. If your setting is also dealing with connectivity or device limitations, our guide on weathering network outages is a useful reminder that delivery systems are only as strong as the communication plan around them.

2. Start with the educational problem, not the supplier catalogue

Define the intervention purpose in plain language

The most common procurement mistake is to start with product features. Better procurement starts with a teaching problem. Are you trying to increase Year 6 maths fluency, close disadvantaged pupil gaps in reading, improve GCSE English Language outcomes, or support A-level resits? A good tender specification must say what success looks like in educational terms before it says what the supplier must provide. This prevents the “shiny platform” effect, where impressive software hides weak alignment to actual need.

A helpful way to frame the brief is to identify the target group, the expected dosage, the subject or domain, and the desired outcome window. For example, “Year 10 pupils who are one standard deviation below expected progress in algebra will receive 20 weeks of structured tutoring, with weekly evidence of improved accuracy on diagnostic tasks.” That is far more useful than “We need maths tutoring.” To sharpen your intervention logic, it can help to borrow planning discipline from other sectors, such as the evaluation principles discussed in cost, speed and reliability benchmarking and privacy-first analytics design.

Separate universal needs from targeted needs

Not every pupil needs the same support model. Some need high-frequency practice and diagnostic feedback; others need human reassurance, motivation and responsive explanation. If you are procuring for a MAT, it is wise to segment requirements into cohorts. Primary maths fluency, secondary GCSE intervention and post-16 exam preparation should not be bundled into a single undifferentiated requirement. The tender should state whether the provider must cover one cohort or several.

This segmentation also changes your supplier shortlist. A fixed-price AI product may suit high-volume, repeatable maths tuition, while a marketplace model may be better for bespoke subject choice or hard-to-fill niche subjects. If your trust runs a wide curriculum offer, it is sensible to compare models against subject depth and staffing resilience, not just unit cost. For the same reason, our article on AI rollout compliance is useful reading when technology is part of the tutoring proposal.

Turn school improvement priorities into procurement requirements

Strong tender specs echo the school improvement plan. If attendance is a problem, then the supplier must show how they minimise no-shows and communicate with schools. If SEND inclusion matters, the programme should explain how it adapts materials and tutor approach. If disadvantaged pupil progress is the priority, progress reporting must be granular enough to separate engagement from outcome. In other words, procurement should be built around what the school is accountable for, not what the vendor happens to sell.

A useful check is to ask: could a governor read this brief and understand why this intervention was selected over another? If the answer is no, the brief is probably too product-led. To strengthen your evaluation culture, you may find lessons from directory vetting and style assessment logic helpful, though your school version should always be grounded in education outcomes.

3. How to write a tender specification that produces usable bids

Describe the service in enough detail to compare like with like

A good specification removes ambiguity. It should define the age range, subjects, delivery mode, expected frequency, approximate pupil numbers, and whether sessions are 1:1, small group or mixed. It should also spell out operational constraints, such as school hours, booking windows, minimum tutor qualifications, safeguarding checks and data reporting expectations. When bids come back, you want to compare genuine alternatives, not apples with oranges.

Include explicit requirements for onboarding, teacher communication, cancellation policy, replacement tutor policy and support response times. Many schools underestimate how disruptive admin can be when these are unclear. If one provider requires manual coordination for every session while another automates scheduling, the delivery difference matters as much as the lesson content. This is a procurement lesson familiar from other categories too, such as procurement pitfalls in secure file transfer where process clarity prevents expensive surprises.

Ask suppliers to evidence curriculum alignment

Curriculum alignment should be an explicit section in the tender. For primary and secondary tutoring, ask how lessons map to national curriculum objectives, GCSE specifications or exam-board content. Providers should state whether they adapt to a school’s scheme of work or deliver a standardised programme. A vague promise of “personalisation” is not enough; you want to know what is personalised, how it is chosen, and who reviews it.

A strong supplier will be able to show sample lesson plans, diagnostic pathways, adaptive question sets and teacher dashboards. If the service is AI-led, ask how it avoids inappropriate pacing or superficial practice. If it is tutor-led, ask how tutor consistency is maintained across multiple pupils and sessions. For schools planning secure digital workflows, our piece on designing guardrails for AI workflows gives a useful model for setting controls around automated systems.

Set out reporting, escalation and safeguarding requirements

The tender should require progress reporting at intervals you can actually use. Weekly is useful for operational management; half-termly is often better for strategic review. Specify what the report must include: attendance, engagement, learning objectives covered, assessment gains, missed sessions, and concerns escalated to the school. If the provider cannot produce reports that teachers can act on, the intervention risks becoming a black box.

Safeguarding is equally important. Ask how tutors are vetted, how they are trained, what checks are in place, and how online sessions are monitored. If the supplier is engaging directly with pupils, insist on clear DSL escalation routes and session recording or observation policies where appropriate. Schools should not assume that all “enhanced DBS” claims mean the same thing. Our guide to safe online tutoring models and the practical advice in vetting marketplaces are relevant here.

4. KPIs for tutoring: what to measure and what to ignore

Use a balanced scorecard, not a single metric

Too many schools judge tutoring on attendance alone, or on one end-of-programme test. That is not enough. Good KPIs for tutoring should capture inputs, engagement, learning and impact. A balanced scorecard helps you avoid overreacting to a single data point. For example, a programme can show excellent attendance but weak progress if the curriculum is misaligned; equally, a short-term test gain may mask poor retention.

At minimum, build KPIs across four categories: delivery fidelity, engagement, progress and satisfaction. Delivery fidelity asks whether the agreed dosage was delivered on time. Engagement looks at participation and completion. Progress measures diagnostic improvement or assessment movement. Satisfaction captures teacher and pupil feedback, but should never replace outcome data. This approach mirrors good performance design in other sectors, where operational reliability and user experience are tracked alongside results.

Examples of practical tutoring KPIs

For schools and MATs, the most useful KPIs are specific and auditable. A few examples include: 90% of scheduled sessions delivered; 85% of pupils attending at least 80% of sessions; average diagnostic improvement of one curriculum step per half-term; 75% of teachers rating progress reports as actionable; and all sessions supported by appropriate safeguarding checks. These are more meaningful than “positive feedback” because they link directly to educational delivery.

When comparing suppliers, ask how each KPI is measured and who owns the data. If a platform produces its own assessments, does it benchmark against external curriculum expectations? If tutors self-report engagement, is there independent quality assurance? The more objective the evidence, the more confidence you can have in value for money. For another perspective on vendor metrics and trust, see how to vet a marketplace or directory.

Do not confuse proxy measures with progress

Attendance, logins and satisfaction are all useful, but they are still proxy measures. They tell you whether the machine is running, not whether the pupil is learning. A supplier that excels at nudging pupils to turn up may still deliver weak academic impact. That is why your tender should distinguish between operational KPIs and educational KPIs, then give greater weight to the latter in evaluation.

Ask for pre- and post-intervention data, but also ask how gains are sustained. A pupil who improves in week eight and regresses by week twelve has not truly benefited. Similarly, a supplier that reports average uplift without subgroup analysis may be hiding uneven outcomes for SEND pupils, EAL learners or disadvantaged cohorts. Strong progress reporting should allow you to see who is improving, by how much, and in what context.

5. Fixed-price AI tutoring vs per-hour marketplace models

What fixed-price AI programmes are good at

Fixed-price AI tutoring programmes are attractive because they make budgeting simpler and scaling easier. If a provider offers unlimited or high-volume one-to-one support for a set annual fee, the cost per pupil can fall sharply as usage rises. That is especially valuable for maths intervention, where practice, repetition and immediate feedback matter. It also reduces the staffing volatility that can hit traditional tutor pools during exam season.

The downside is that AI systems may work best in narrow subject domains and may not be equally strong for complex explanation, motivation or essay-based feedback. Schools should therefore test whether the system genuinely adapts to the pupil or simply branches through a pre-written pathway. Ask to see evidence of learning gains, not just user engagement. The source material’s note that Third Space Learning’s AI maths tutor uses fixed annual pricing is a useful example of how schools are now being asked to assess value at scale.

Where per-hour marketplace models win

Per-hour marketplace models are more flexible. They can provide subject breadth, local nuance, and human tutors for a wide range of key stages and exam needs. They are often well-suited to one-off support, specialist subjects, or schools that need a live tutor quickly. For example, a school may need A-level Chemistry, GCSE Spanish, and primary SATs support in the same term. Marketplace models can sometimes cover that breadth better than a single-programme provider.

The challenge is variability. Quality can differ by tutor, by session and by subject. Procurement therefore needs stronger controls: minimum qualifications, sample lesson expectations, pricing transparency and replacement policies. In other words, the flexibility premium must be justified by robust governance. Our reading on marketplace vetting is especially relevant if your shortlist includes open tutor platforms.

A simple comparison table for procurement teams

Which model is “better” depends on the use case

There is no universal winner. If your school needs large-scale, repeatable maths practice with predictable cost, a fixed-price AI programme can be highly efficient. If you need a broad menu of subjects, high-touch human support and flexible scheduling, a marketplace may fit better. The right choice depends on your intervention aim, your budget pattern and your tolerance for variability. Procurement teams should resist the temptation to choose the model that sounds most innovative rather than the one that best solves the problem.

One useful tactic is to run a pilot against the same success criteria for both models, then compare cost per achieved progress point rather than cost per hour. That reframes the decision around impact. It also gives your finance team a more robust basis for future planning, especially if you are managing multiple intervention strands across a trust.

6. Supplier evaluation: how to score bids without being dazzled

Use weighted criteria that reflect school priorities

Supplier evaluation works best when criteria are transparent and weighted. A practical structure might be: educational effectiveness 35%, safeguarding and compliance 20%, value for money 20%, delivery and reporting 15%, and implementation support 10%. If your trust is highly cost-constrained, you may increase the price weighting, but do not let cost dominate to the point where quality is underweighted. Cheap tutoring that does not change outcomes is not value for money.

Scoring should be done by a panel that includes teaching, safeguarding and procurement perspectives. That reduces the risk of overvaluing a slick platform demo or underestimating operational risk. Ask each supplier to respond to the same set of questions and present the same evidence pack. To make the process more robust, borrow the discipline of structured review used in governance frameworks and the evaluation logic in supplier vetting guides.

What evidence should suppliers provide?

At minimum, require: case studies with comparable cohorts; sample reporting dashboards; safeguarding policies; tutor vetting procedures; pricing breakdowns; implementation timelines; and evidence of impact from independent or school-led evaluation. Be careful with claims that sound impressive but are not comparable to your setting. A provider may show national uplift statistics, but if those are from a very different age group or subject area, they have limited value for your decision.

Schools should also ask for references from settings like theirs. A MAT with 20 primary schools needs evidence from multi-site delivery, not just one strong single-school pilot. A secondary school preparing pupils for GCSE may need proof that the model can track exam-specification progress, not just general confidence gains. This is where procurement becomes an exercise in relevance, not just reputation.

Red flags that should lower the score

Several warning signs should trigger caution. These include unclear pricing, no named safeguarding lead, weak cancellation policy, no comparator data for pupil progress, and reporting that focuses only on usage. Another red flag is a supplier that promises “personalised learning” but cannot explain what input data drives adaptation. If a bid is all confidence and no detail, it is probably not ready for public-sector scrutiny.

It is also worth checking whether the supplier can integrate with your existing school systems and data flows. If the data export is awkward, you will spend time rekeying information and may end up with unreliable monitoring. Good procurement eliminates friction, rather than exporting it to staff.

7. MAT procurement: how trusts can standardise without losing flexibility

Centralise the framework, decentralise the need

MAT procurement works best when central teams set the framework and schools specify the need. The trust can pre-approve a small set of suppliers, define minimum safeguarding and reporting standards, and create a standard evaluation template. Individual schools then select the most appropriate option for their cohort, subject and timetable. This reduces duplicated procurement effort while still allowing local responsiveness.

A common trust mistake is either total centralisation, which stifles school-level nuance, or complete decentralisation, which creates inconsistent quality and uneven pricing. The sweet spot is a trust-wide playbook with local application. For a broader look at how strategy can balance consistency and creativity, see standardising workflows and resilient content strategies, which offer a useful analogy for operational design.

Create a shared evidence base

Trusts should store supplier scorecards, pilot results and progress data in one shared repository. That allows future buying decisions to build on prior evidence rather than starting from scratch each time. It also makes it easier to compare providers across schools using a consistent framework. Over time, you can identify which interventions work best for which pupils, and which providers are strongest on delivery reliability versus academic gain.

This evidence base is also useful in governor reporting and budget setting. If one supplier repeatedly produces better outcomes at similar cost, the trust can negotiate more confidently or expand usage. If another provider has attractive branding but poor retention or inconsistent reports, the central team can step away with evidence, not instinct. That is the essence of strong MAT procurement.

Build a repeatable procurement calendar

Finally, do not leave tutoring procurement until the term has already started. A trust calendar should include needs analysis, market testing, shortlist review, pilot approval, implementation and mid-year evaluation. That makes budget use more predictable and protects curriculum time. It also gives suppliers enough lead time to prepare tutors, align content and complete safeguarding processes.

For busy trusts, the calendar approach can be the difference between strategic intervention and rushed purchasing. It allows you to compare suppliers fairly and avoid emergency buying when exam dates are close. If your team is also managing multiple digital vendors, our guides on procurement discipline and mapping your SaaS attack surface reinforce the same principle: standardise the assessment before you scale the spend.

8. How to test progress reporting and impact claims

Ask for baseline, midpoint and endpoint evidence

Any supplier can claim impact after a pilot. Strong procurement asks how the claim was measured. Did the provider collect baseline data before tutoring started? Was there a midpoint check to adjust instruction? Was the end assessment comparable to the baseline? Without that chain, it is hard to know whether progress resulted from tutoring, maturity, curriculum teaching or exam practice elsewhere.

For schools, the most persuasive reports are those that link diagnostic data to classroom action. If a tutor identifies that a Year 8 pupil is missing fraction equivalence, the report should say so clearly and suggest the next step for the class teacher. That turns tutoring into part of the wider teaching ecosystem rather than a disconnected add-on. Strong reporting should therefore be diagnostic, not just celebratory.

Demand subgroup analysis where relevant

Progress reporting should not stop at averages. Ask for subgroup analysis by key stage, disadvantage status, SEND, EAL and prior attainment where numbers allow. This helps you see whether tutoring is reducing gaps or merely raising the average. If the supplier cannot provide this level of detail, you may miss uneven impact until too late.

Schools that are serious about equity should also ask whether the model is accessible to pupils with additional needs. Can materials be adapted? Are tutors trained in inclusive practice? Can session length or pacing be modified? These questions are important because a tutoring service that works for the median pupil may fail the groups your intervention budget is meant to support.

Use a simple impact review cycle

A practical review cycle is: baseline, six-week check, half-term review, and end-of-block evaluation. At each stage, review attendance, engagement and progress, then decide whether to continue, adapt or stop. This prevents sunk-cost bias and helps you reallocate budget if a programme is underperforming. It also makes your procurement more intelligent over time because decisions are based on evidence, not habit.

Where possible, compare tutoring outcomes with a matched group or previous cohort. Even a simple comparison can reveal whether the intervention is outperforming normal classroom progression. If that comparison is impossible, at least ensure that teacher judgments and standardised assessments are triangulated. The goal is not perfect research design; it is trustworthy decision-making under real-world constraints.

9. A practical tender checklist for schools and MATs

Before you publish the tender

Check that your brief clearly defines the pupil cohort, subject focus, delivery mode, timeframe and success criteria. Confirm who is evaluating bids, how scores are weighted, and what evidence is mandatory. Make sure safeguarding and data protection requirements are explicit, especially where online tutoring or AI-enabled tools are involved. If you want to cross-check your assumptions, it may help to read current market examples alongside AI compliance guidance.

Also ask whether your procurement route needs to be direct award, mini-competition or framework-based. The route matters because it affects speed, auditability and the scope for negotiation. In a MAT, central procurement policy should define thresholds so schools know when they can buy locally and when they must use the trust process.

During supplier evaluation

Require demonstrations that show the actual user journey, not just polished sales slides. Ask the supplier to walk through onboarding, session delivery, reporting and escalation. Then test whether the claims match the documentation. You are looking for operational coherence as much as educational promise.

Where possible, run reference checks with schools of a similar size and context. Ask what went wrong as well as what went right. Good suppliers will be candid about constraints and implementation learning. That honesty is often a better indicator of long-term fit than overconfident marketing language.

After contract award

Do not treat award as the finish line. Build a 30-60-90 day implementation review, then a half-termly performance review. Check whether the agreed KPIs are being met and whether staff are actually using the reports. If the service is not delivering, intervene early rather than waiting until the end of term. Procurement only creates value when contract management is as disciplined as tendering.

This is also where good communication matters. Schools, tutors and leaders need a simple rhythm for updates and issues. If your team struggles with handover or continuity, our article on trusted voice and communication may sound unrelated, but the principle is the same: consistent systems build confidence.

10. The bottom line: buy tutoring like an outcome, not a service

Value for money means outcomes per pound, not pounds per hour

The post-NTP market rewards buyers who are specific, evidence-led and disciplined. If you define the educational problem clearly, write a detailed specification, set measurable KPIs and evaluate suppliers against comparable evidence, you will make better decisions. The strongest suppliers will welcome that scrutiny because it lets their strengths show through. The weakest will usually fade once the questions become concrete.

That is why the comparison between fixed-price AI and per-hour marketplace tutoring should not be framed as technology versus human expertise. It should be framed as which model produces the best educational return for your pupils, in your setting, under your constraints. Sometimes the answer is one model. Sometimes it is a blended strategy. The key is to decide on evidence, not assumptions.

Procurement is now part of intervention strategy

In a world where schools must justify every pound, tutoring procurement is no longer an admin task. It is a strategic lever for attainment, equity and operational resilience. The schools and MATs that do this well will treat procurement as an extension of pedagogy: clear aims, measurable outcomes and regular review. That is how tutoring becomes more than a purchased hour and turns into a meaningful intervention.

For further market context and supplier comparison, revisit our market overview of online tutoring websites, then pair it with practical due diligence from marketplace vetting and procurement pitfall prevention. Taken together, those lessons help schools buy smarter, protect pupils and get better outcomes from every intervention pound.

Pro Tip: When comparing tutoring bids, calculate cost per unit of measured progress, not cost per session. That simple change often reveals which supplier truly delivers value for money.

Related Reading

• 7 Best Online Tutoring Websites For UK Schools: 2026 - Compare current provider models and see how schools are buying tutoring now.
• How to Vet a Marketplace or Directory Before You Spend a Dollar - A useful framework for checking platform credibility and quality assurance.
• Avoiding Procurement Pitfalls in Martech for Secure File Transfer Solutions - Learn how to spot hidden vendor risks before signing a contract.
• State AI Laws vs. Enterprise AI Rollouts: A Compliance Playbook for Dev Teams - Helpful if your tutoring solution includes AI or automated data processing.
• Secure Cloud Data Pipelines: A Practical Cost, Speed, and Reliability Benchmark - A smart read on how to assess operational performance and reliability in digital systems.