9 September 2026. Root's bounded acquisition after the supplier-platform forest. No final evaluation output was located in the scoped public search. Requirements, announced funding, unpublished deliverables, disclosed copies and observed performance remain separate.
Primary retained original: ../supplier-platform-wave-2026-09-09/cddp-foi-bundle.pdf, SHA256 71bd6a522a765d53f9aa79b239134226be64891ea3c67711abc9d788fb87748f. Source: http[local research file] . This run read PDF pages 18,34–35,62–68,88–100 and visually inspected pages 18,63,93,95,97,100,193–194. Local PNGs are derivatives of that original. Page18 is almost fully redacted; no hidden text was reconstructed. The prior run read the contract/IP pages separately.
Official release: http[local research file] . This run read the complete exemption account, not only its document inventory. The public page contains an unreconciled date discrepancy: publication 28 June 2024, request 25 August 2023, response field 25 September 2024. Original bundle filename contains FOI2023 while release path/reference says FOI2024. Preserve those labels without silently fixing the response year.
The October 2021 business case, PDF62–68 (printed61–67), envisages process and impact evaluation, consultation with an independent evaluator and HMT, and sharing supporting evidence with Treasury teams. Process evaluation would examine cross-government coordination and working practices, with KPIs developed after an initial systems audit. Impact evaluation would assess whether improved data products enhanced government understanding and operational response. Specific criteria were still to be agreed. Publicly legible criteria concern funded delivery, government data sharing/analysis, stakeholder benefit and lessons. Detailed portions of both evaluation strands are visibly redacted; absence of a readable accuracy, rights or false-positive metric is not proof that no such metric was used.
The expert-evaluator specification, contract 102384, PDF88–92, asks for independent end-to-end monitoring, value-for-money assessment, assessment of digital practice and the effectiveness of correctly identifying misinformation/disinformation trends. This correctness requirement matters: the available specification is not limited to efficiency. The contract also requires reporting that senior civil servants across Whitehall can use, with successes, failures and long-term recommendations. A final report is an expressly commissioned deliverable. Experimental or quasi-experimental evaluation is conditional; interest in possible A/B testing is not evidence an experiment was conducted, let alone a population manipulation experiment.
The provisional schedule contains a feasibility report to HMT in February 2022, a phase-two decision in March 2022, interim evaluation in August 2022 and final evaluation gateway in February 2023 before phase end in March 2023. Earlier rows carry apparent year inconsistencies (November–February 2021 and December/January 2021); these remain as printed. PDF92 expressly makes subsequent-phase funding conditional on assessment of the discovery feasibility review and approval by HMT officials. This locates an intended funding gate. It does not identify the actual official, decision minute, evidence considered or whether the gate was exercised as planned.
The internal DCMS team administers contracts, supplies governance secretariat and leads cross-government engagement. The separately contracted digital project-management team leads discovery/delivery, while the evaluator examines it. These are specified roles, not names of the actual committee members. PDF62 and68 direct evaluation evidence and financial/performance information toward the Shared Outcomes Fund team, HMT spending team and other government departments.
The release inventory identifies document6 as the DCMS/Oliver Wyman evaluation-and-monitoring contract. PDF95 is order102384, but the supplier name, address and registration number are visibly redacted; the brand identification comes from the department's inventory, not a recovered supplier field. Exact legal-entity registration remains unresolved. No related corporate ownership claim is made.
PDF93's pricing table is visible in the rendered page despite failing text extraction: Phase I GBP108,400 excluding VAT; Phase II GBP547,600; total fixed tender price GBP656,000 excluding VAT. The same GBP656,000 appears in PDF97 as the estimated first-year charges used for liability calculation. Payment is in arrears after relevant deliverables satisfy DCMS. This is a disclosed contractual/tender price, not proof of payment. Tables B and C below contain blacked-out fields despite the inventory calling the pricing document disclosed; do not say all prices or rate cards were disclosed.
The order is dated 14 January 2022 (PDF96), while PDF97 visibly prints a 20 December 2022 start, three-month initial period and 31 March 2023 expiry, alongside a phase-one extension reservation. Those terms differ from the specification's December2021-to-March2023 project schedule. A clerical mistake, later amendment or different phase might explain the difference; none is established. PDF100 signature/name/role/date fields are visibly empty. The released copy does not independently establish execution, and it does not prove no executed contract exists.
PDF193–194 visibly redact the table of programme outputs/success measures and detailed milestone descriptions/acceptance criteria. The dates 15 July 2022 and 30 September 2022 remain visible. This confirms that the missing substantive cells are redactions, not merely an OCR failure. Conversely PDF93 demonstrates why failed text extraction alone must not be called redaction. The contract variation's reported MVP completion is stronger than a plan, but still not an independent assessment of accuracy, fairness, effectiveness or downstream platform action.
The department's release account applies personal-data, commercial, policy-development, security and public-affairs exemptions to specified parts. It says section36 was applied across every document except the pricing schedule and defends withholding operational capabilities/vulnerabilities. Discovery Report and Product Operating Model are withheld in full under the described security/public-affairs basis after removal would make the remainder illegible. These are the department's reasons; no ICO or tribunal review of this particular release was acquired. The earlier ICO Crisp-report ruling is a different record and does not adjudicate these CDDP redactions. Neither-confirm-nor-deny wording concerning additional security-related records proves neither their existence nor absence.
A separate Treasury Shared Outcomes Fund Round2 summary publicly lists CDDP funding of GBP8.4 million and participants DCMS, FCDO, Home Office, Cabinet Office, Ministry of Defence, GCHQ and NCSC. This is announced pilot funding and participation, not a ledger of money received by Faculty or Oliver Wyman, nor proof every participant supplied intelligence data or controlled the platform. Do not add its GBP8.4m to supplier prices, which may be components or differently scoped amounts.
Retained original shared-outcomes-round2.pdf: 230,526 bytes, SHA256 c91b2c59cfcafc70bb10f4aab856d4d15c92166957f587599d47a94a671598bc. Read PDF9/printed6, CDDP entry. Source: http[local research file] . Public reader failed once; ordinary direct download succeeded. Do not confuse adjacent Global Supply Chain Intelligence Pilot and its GBP4.7m entry with CDDP.
Scoped exact-title searches combined CDDP, final report, evaluation, Oliver Wyman and contract102384 across government/parliament/supplier sources. They returned the original commissioning material but no actual final evaluation, agreed KPI set, Treasury gate decision, or completed acceptance certificate. This is an acquisition gap, not evidence that evaluation never occurred or was adverse. No new request was sent.
Most revealing likely records are the initial audit/KPI agreement; discovery feasibility report sent to HMT and funding decision; interim/final evaluation report and its supporting evidence; milestone acceptance record; and the log translating analyst findings into subsequent action. Commissioned documents have firmer existence support than guessed internal communications; delivered versions remain unconfirmed.
A separate official source surfaced: the Online Safety Data Initiative (OSDI) appointed PA Consulting to evaluate a Faculty/PUBLIC delivery consortium. OSDI is not CDDP, its evaluator is not Oliver Wyman, and its results cannot stand in for CDDP's. It offers a useful next comparison: how the same public funding model and overlapping suppliers handle classification, independent assessment and publication in a less opaque project. Source lead: http[local research file] . Only its indexed official opening was read in this run; further acquisition follows a forest review.