Track Subcontractor Retention Release Across UK Projects

UK subcontractor retention is typically 3–5% deducted from each interim payment, with half released at practical completion of the package and the second half at the end of the defects liability period (commonly 12 months later). To track subcontractor retention release across projects, build a per-package ledger that captures every retention deduction from interim payment certificates. The first half-release is then tied to the practical-completion certificate for each package; the second half-release is tied to the defects-liability-period end date.

That working answer is now load-bearing for a second job. From financial years beginning on or after 1 April 2025, qualifying construction companies have to publish a twice-yearly disclosure on their retention practices under the Reporting on Payment Practices and Performance (Amendment) Regulations 2025. The same ledger that tells a commercial team what's held and what's overdue is now the data source that has to feed a public statutory disclosure — turning what was always a working-capital nuisance into a workflow with a hard publication deadline and personal director liability attached.

This article serves two faces of the same audience. A subcontractor commercial manager or QS chasing release from one or more main contractors needs the ledger discipline to make overdue retention visible and chaseable. A main-contractor QS or commercial director on the payable side needs the same discipline both to manage their own retention obligations and, above the qualifying-company threshold, to produce the new statutory disclosure. The walk-through below builds one ledger that serves both jobs, then wires it into the disclosure regime so the publication is a query against the data rather than a compliance project mobilised at deadline.

How Retention Is Released — The Mechanic and the Trigger Points

The standard retention deduction is 3–5% of each interim valuation, withheld from the subcontractor by the main contractor at certification. The actual percentage and any overall cap are set by the subcontract; 3% with a cap at half the contract sum's standard limit is common on larger packages, 5% uncapped is common on smaller ones. Each cycle, the deducted amount adds to a cumulative retention balance held against the package over its life — by the time a long-running package reaches practical completion, the cumulative held figure can be a meaningful slice of the contract value sitting outside the subcontractor's working capital.

Release runs in two stages. Half of the cumulative retention — the first moiety — is released against the practical completion event for the package. The remaining half — the second moiety — is released at the end of the defects liability period, on the issue of the certificate of making good defects. This is what the trade calls the half-release and second-half-release, and it is the structure every retention conversation between a subbie and a main contractor's commercial team is built around.

Contract-language convention varies and matters for clarity. Post-2011 JCT contracts use the term Rectification Period rather than Defects Liability Period; the meaning is equivalent and the second-half-release mechanic is identical. NEC contracts use Defects Date and a defects correction period; again the underlying release mechanic holds, even where the labels differ. A practitioner ledger needs to carry the contract-specific term and date for each package rather than collapse them all into one default name. The companion article on the JCT and NEC application for payment versus VAT invoice walks the certification side of the same contract families.

The trigger that the SERP's vendor and template content most consistently fudges is the first-half release. The first moiety should release on the practical completion of the subcontract scope, not on the practical completion of the main contract. Subcontract packages typically complete months — sometimes more than a year — before the main-contract PC; a groundworks package may finish before the superstructure starts, an MEP first-fix package finishes before the second-fix begins. Treating main-contract PC as the trigger systematically delays subbie release by exactly the gap between the two events, which is precisely how chronic retention leakage gets baked into the workflow without anyone arguing about it. The ledger has to record the practical-completion certificate that applies to the package itself, and the date on it, as the trigger.

The defects liability period is conventionally 12 months but is contract-dependent. Six-month periods appear on shorter or simpler packages; 24-month periods appear on engineering-heavy work. Assuming a 12-month default in the ledger is one of the easier ways to compute the wrong release date for a package and either chase too early or, more often, fail to chase at all. Record the contractual period per package alongside the practical-completion date, and the second-half release date computes correctly without thought.

Cash retention is the case this article walks through. An on-demand retention bond from a surety can replace cash retention by contract — releasing the working capital to the subbie in exchange for a bond premium and the surety's underwriting — and is worth raising in negotiation where it fits. The mechanic and ledger structure described here cover the cash case; bonds change the cashflow but not the contract-level structure of practical-completion and defects-liability events.

The Per-Package Retention Ledger

The grain of a working subcontractor retention tracker UK practitioners can actually use is one row per subcontract package per period. A period is the interim valuation cycle — usually monthly. A package that runs eighteen monthly valuations produces eighteen rows in the ledger; a package that runs sixty produces sixty. The ledger is long-format and additive, which is what makes the cross-project consolidation in the next section work. A wide-format retention release spreadsheet construction teams sometimes inherit — one row per package with monthly columns running across the page — collapses under the weight of multiple counterparties and breaks the moment a counterparty appears on more than one project.

The columns the ledger needs to carry, naming what each captures:

• Project identifier — the scheme the package belongs to.
• Subcontract package identifier — the specific package (e.g. groundworks, M&E first fix, dry lining), uniquely keyed.
• Counterparty — the subbie's name on a main contractor's payable ledger; the main contractor's name on a subbie's receivable ledger. The same ledger structure works for both sides; the counterparty column tells you which side the row sits on.
• Period reference — valuation number and the period-end date for that cycle.
• Period gross valuation — the assessed value of work in the period before any deductions.
• Period retention percentage — the contractual retention rate applied in this period.
• Period retention deducted — the cash amount withheld this period.
• Cumulative retention held to date — the running total of retention withheld against this package up to and including this period.
• Practical-completion certificate reference — the identifier of the PC certificate for this package, populated when issued.
• Practical-completion date for the package — the date on the certificate.
• First-half-release amount — the cash amount of the first moiety release.
• First-half-release date — when the first moiety was actually paid.
• Contractual defects liability or Rectification Period length — recorded per package, in months.
• Defects-liability end date — practical-completion date plus the contractual period (or the contract-stated Defects Date for NEC).
• Certificate of making good defects reference — the identifier when issued.
• Second-half-release amount — the cash amount of the second moiety release.
• Second-half-release date — when the second moiety was actually paid.
• Balance outstanding — derived: cumulative retention held minus first-half-release amount minus second-half-release amount.
• Status — derived: held, released, or overdue, computed against the release-date columns and today's date.

The two derived columns are the operational outputs the ledger exists to produce. Balance outstanding tells the commercial team how much money is currently held against this package; the status column tells them whether anything is past its trigger and not yet paid. Without those two derived columns the ledger is a record-keeping exercise; with them, it is the chase list and the cross-project exposure view in one structure.

Make the ledger the single record of fact. Half-released amounts and dates live in the ledger, not on individual certificates filed in package folders. Without that consolidation the held-vs-released subcontractor view across projects cannot be built and the disclosure work later in the article has nothing to aggregate over. The ledger does not have to live in any particular tool — many teams operate it in Excel, some in their accounting or contract-management system, some in a purpose-built workbook a senior surveyor built years ago and never wrote up. The structure is what matters; the rest of the article works regardless of where the rows physically live.

Extracting Retention Data From Interim Certificates and Release Evidence

Each ledger column has a source document. The period gross valuation, the period retention percentage, and the period retention deducted come from the interim payment certificate or the assessed application for payment for that period — the document the contracts call the Notified Sum, on the back of assessing the subcontractor application for payment that produces the retention deduction. The cumulative retention held figure is normally shown on the same certificate but should be recomputed from the deducted history rather than trusted blind; certificates carry transcription errors, and the ledger is more robust when it sums what it has rather than copies a stated total. The first-half-release amount and date come from the practical-completion certificate for the package. The second-half-release amount and date come from the certificate of making good defects issued at the end of the defects liability or Rectification Period.

At the period level, the extraction job is the same every cycle. For every subcontract package on every project, the assessed AfP carries a retention-deduction line, and that line has to land in the ledger as a new row keyed to the package and the period. A contractor running ten projects with twenty subcontract packages each is adding two hundred rows per cycle from interim certificates alone — a volume that justifies extracting the line rather than typing it. The same upstream batch typically also carries CIS deductions and gross-of-CIS payment lines that flow into a separate compliance return; one extraction pass over the certificates can populate both ledgers, on the same logic that drives compiling the CIS300 monthly return from the same invoice batch every month.

Practical-completion certificates and certificates of making good defects are issued less frequently than interim certificates, but they are the trigger documents that move retention from the held column to the released column. The ledger has to ingest them the moment they are issued; gaps here are how overdue release goes undetected. A package whose practical completion was certified four months ago but whose half-release has not been recorded in the ledger looks held when it is in fact overdue. The status column in section 3 only does its work if the trigger documents land in the ledger as soon as they are dated.

The change automation makes is in the per-period extraction step. Instead of a commercial assistant manually transcribing the retention-deduction line from each PDF certificate into a spreadsheet at month-end, the certificates flow through a structured extraction step that returns the ledger rows directly, ready to be appended. This is the canonical job our tool is built for: upload the period's batch of interim certificates and assessed AfPs, prompt the AI to extract retention deductions from interim payment certificates automatically — pulling per document the package identifier, the period reference, the gross valuation, the retention percentage, the period retention deducted, and the cumulative retention held — and download a structured Excel or CSV whose rows append directly to the ledger. The same upload-prompt-download flow handles practical-completion and certificate-of-making-good-defects documents when they arrive, populating the trigger and release columns.

A single batch handles up to 6,000 documents, which is more than enough headroom for any contractor's full month of certificates in one job. The accuracy gain matters because the ledger feeds both the cross-project view in the next section and the statutory disclosure later in the workflow; transcription errors compound across the consolidation, and a percentage figure published to the government portal under criminal-liability rules is the wrong place to discover that the source ledger had drift in it.

The Cross-Project View — Where the Ledger Earns Its Keep

The package-level rows are the raw material; the consolidated view is what makes them worth maintaining. Pivot the long-format ledger by counterparty to track retention across construction projects per main contractor (for a subbie) or per subcontractor (for a main contractor). Pivot by project to see exposure per scheme. Pivot by status to see the held / released / overdue split across the whole portfolio. Each pivot is a different question the commercial team needs answered, and each one falls out of the same ledger because the grain is right.

What the consolidated view exposes that the package-level view hides is the scale of the retention book. A contractor running ten projects holds retention in dozens of packages spread across multiple counterparties. Aggregate working-capital lock-up — retention held but not yet due for release — is regularly six and seven figures and almost always materially larger than commercial teams estimate before they build the ledger. The held-vs-released subcontractor split per counterparty is the leverage point: it says where the company's working capital actually sits, and against whom. The overdue subset — retention past its release-trigger date but not paid — is the chase target the commercial team works through every cycle.

The same consolidated view feeds the monthly cost-value reconciliation. Retention is held but not yet earned by the counterparty until released, so it sits as a balance-sheet item rather than as cost in the period. The CVR has to reflect that balance accurately month over month, and the reconciliation falls into place when the source data is the same ledger that the surveying team is already maintaining for chase purposes — the alternative is two separate retention numbers in the business that disagree and have to be reconciled by hand, which is the situation the ledger exists to end. The companion piece on treating retention as a balance-sheet item in the monthly CVR walks the CVR side of the same data flow.

Downstream, the cumulative retention released figure feeds the final account. The closing certificate's adjustment lines include the cumulative retention to be released against the package; the ledger produces this figure directly when reconciliation time comes, which is the natural way of rolling cumulative retention release into the construction final account workbook. Built once at the source, the figure is consistent across the chase pack, the CVR, and the final account — three downstream uses of one underlying data structure.

The same consolidated ledger is the data base layer for the statutory disclosure the next section covers. The per-counterparty held-vs-released split — already produced for chase and CVR purposes — is exactly what the regulations require qualifying companies to publish. The work to make the disclosure mechanical has already been done by the time the reporting period closes, provided the ledger discipline has been maintained.

---

The 2025 Statutory Disclosure Regime

The regulatory frame the rest of the article hangs on is the new UK retention reporting regulations 2025. The Reporting on Payment Practices and Performance (Amendment) Regulations 2025 came into force on 1 March 2025 and apply to each financial year of a qualifying company beginning on or after 1 April 2025, requiring qualifying construction companies to report whether their payment practices include the use of retention clauses, the mechanism for releasing retained monies, and the value of retention withheld from suppliers compared with retention withheld by clients. SI 2025/75 amends the underlying Reporting on Payment Practices and Performance Regulations 2017; the construction-specific schedule is what this section deals with.

The qualifying-company test is not a turnover test, and framing it that way is technically wrong even though it appears that way in much of the available commentary. Qualification follows the Companies Act 2006 large-company definition: a company qualifies if it meets any two of three criteria — turnover, balance-sheet total, and average number of employees. The construction-specific reporting threshold maps to that test, with the turnover criterion at £36m for financial years beginning before 6 April 2025 and rising to £54m for financial years beginning on or after that date. The balance-sheet and employee criteria sit alongside, so a company below the turnover threshold can still qualify on the strength of the other two and a company above it might in principle not qualify if it fails the others — though in practice a company over the turnover threshold is almost always over the balance-sheet and employee thresholds too.

The disclosure items the construction-specific schedule requires are in plain language:

• Whether retention clauses are used in qualifying contracts. Stated as on all, on some, or on none of the company's qualifying contracts in the reporting period.
• The standard retention percentage rate the company applies, and any minimum contract value below which retention is waived. A statement of policy — what the contract templates say.
• A description of the contractual mechanism for releasing retained monies. The trigger events (practical completion, certificate of making good defects) and the timing of release against each.
• Whether the company's retention terms with its supply chain are no harsher than the retention terms it accepts from its own clients. The "no harsher than" principle; a yes/no statement with explanation where applicable.
• Two comparative calculations. The value of retention withheld from suppliers and the value of retention withheld by clients, each as an absolute figure and as a percentage of total qualifying contract payments.
• Identifying information of the approving director. The director who signs off the report.

Each report covers a six-month reporting period within the financial year. Publication is on the government's reporting portal within 30 days of the period end.

Failure to publish a required report on time is a criminal offence committed by the company and by every director, on indictment punishable by an unlimited fine. That liability sits with each individual director, not only with the company. It is the reason the disclosure cannot be a manual scramble in the days before deadline; the data has to be a mechanical output of a properly structured ledger, available on demand whenever the reporting window opens.

Mapping the Ledger to the Disclosure Metrics

The disclosure items split cleanly into two kinds. Some describe retention policy — what the company does as a matter of contract template and standing commercial position. Some report retention quantities — how much was withheld over the reporting period and from whom. The policy items are answered from contract templates and policy documents the company already maintains. The quantitative items are aggregations over the per-package ledger built in section 3 and populated as section 4 describes. Both halves resolve to data the company already structures, provided the ledger is in place.

Item by item:

• Whether retention clauses are used in qualifying contracts (all / some / none). Source: the ledger. Count the qualifying-contract packages in the reporting period that carry a non-zero period retention percentage, divided by total qualifying-contract packages in the period. The ratio answers all, some, or none directly.

• Standard retention percentage rate, and any minimum contract value below which retention is waived. Source: contract-template defaults and standing commercial policy. The ledger acts as a cross-check that what is being applied in practice matches the stated policy — if the templates say 3% and the ledger shows packages running at 5%, the policy statement either needs correcting or the contracts being signed do.

• Description of the contractual release mechanism. Source: the contract-template language for the relevant standard form. The ledger evidences the mechanism in operation by recording the release-trigger events against each package — useful as supporting evidence that the stated mechanism is what the company actually runs.

• "No harsher than" disclosure. Source: a comparison of the standard retention terms the company accepts from its own clients (its receivable side) against the standard retention terms it imposes on its supply chain (its payable side). The ledger holds both sides under one structure, which is what makes the comparison computable rather than an exercise in collating disparate contract files.

• Value of retention withheld from suppliers — absolute and percentage. Source: the ledger. Absolute is the sum of the period-retention-deducted column over the reporting period for all subcontractor packages where the company is the main contractor. Percentage is that figure divided by total qualifying contract payments to those suppliers in the same period.

• Value of retention withheld by clients — absolute and percentage. Source: the ledger. Absolute is the sum of the period-retention-deducted column over the reporting period for the company's own contracts where the client is withholding (the receivable side). Percentage is that figure divided by total qualifying contract payments received from those clients in the same period.

• Approving director identifying information. Source: company records, applied at sign-off. Not a ledger item.

The implication is direct. With the ledger structured as section 3 specifies and populated as section 4 specifies, every quantitative item on the disclosure is a SUMIFS query against the ledger. The work that has to happen is the ledger work, done monthly as part of the normal commercial cycle. The disclosure is the output, computed twice yearly when the reporting period closes. The regulation becomes a report on top of an existing data structure rather than a project mobilised at deadline against scattered sources.

Two Workflows From One Ledger

On the subcontractor side, the chase pack is a per-counterparty filter of the ledger to packages where the status column reads overdue — packages whose practical-completion or defects-liability trigger date has passed without the corresponding release recorded. The pack assembles the package reference, the trigger event and its certified date, the half-release amount due, the days overdue, and the supporting certificate references. Sent to the main contractor's commercial team, that is the evidence base for chase calls, escalation up the contractor's commercial chain, and where required, formal recovery action. The discipline is that the pack is generated from the ledger, not assembled from memory and folder hunting; what makes it credible to the receiving team is the certified-document references on every line.

There is a second use of the disclosure regime that current SERP coverage misses entirely. Subcontractors whose main-contractor counterparties qualify under the regime can read those companies' published reports and use the disclosed "retention withheld from suppliers" figures as evidence in retention-clause negotiations on new contracts. A subbie pricing a new subcontract who can see that the prospective main contractor's published average retention rate against its supply chain was 4.8% over the last reporting period has direct, citable evidence to push back against a 10% retention term in the proposed subcontract. The regime makes that comparator public; the subbie's own ledger discipline is what lets them know what to do with it. This dynamic is going to take a few reporting cycles to settle in commercial practice, but the data is published the moment the regime is in force.

On the main-contractor side, the disclosure publication is the period-end aggregation already mapped in the previous section. Internal sign-off requires the approving director on record, the data extracted from the ledger by SUMIFS over the reporting period, and the publication on the government portal within 30 days of the period end. The work that makes that sequence mechanical is monthly ledger discipline maintained across the whole financial year — not a project mobilised when the reporting deadline appears on the legal team's calendar. The point of the architecture in this article is that the disclosure is a query, not a project.

The policy direction is worth one sentence and no more. Government consultation on banning retention outright is live at the time of writing, so the structural-data argument is "do this now under current rules and structure the data so it serves whichever regime survives". A ledger built to the disclosure-aware structure described here holds its value if retention is restricted further, replaced by a different security mechanism, or banned outright; it does not become wasted work. The same structure also extends naturally into wider construction invoice processing automation — the same per-document extraction discipline that populates the retention ledger feeds the commercial team's other working data needs.

Subbies: load the existing certificates and AfPs into the ledger this month so the overdue subset becomes visible — most teams find the chase value in the first cycle they run end-to-end. Main contractors above the qualifying threshold: structure the ledger now so the next reporting period closes mechanically rather than as a scramble at deadline. The hard work is the structure; once the structure is in place, both workflows run from it.