task-gdpval-3

PHASE 1: Build the category taxonomy Purpose: Create the controlled three-level spend taxonomy that every downstream PO, scoring input, and strategy recommendation depends on. Rationale: We fix the taxonomy before any spend math because every later number (family totals, addressable spend, subcategory selection) inherits these code mappings; resolving the mapping traps up front prevents misclassified spend from propagating into the scoring and the final recommendation. Steps: 1.1 We inventory the manufacturing procurement folder and separate the classification-framework files from the spend, contract, stakeholder, and deck evidence, so the controlling structure is clear before we map anything. 1.2 We take category_framework.xlsx as the controlling structure and identify the six Level 1 families. [rubric 1.1] 1.3 We confirm Marketing & Communications has four Level 2 subcategories: Events, Agency Fees, Market Research, and Promotions & Advertising. [rubric 1.2] 1.4 We map active spend codes to Level 3 service types and Level 2 subcategories using spend-code descriptions, framework ownership, and the UNSPSC crosswalk, rather than trusting any single field. 1.5 We resolve the mapping traps deliberately: we combine codes 22000005 and 22001115 into one Events subcategory on their shared UNSPSC family 80141600. [rubric 1.3] 1.6 We map code 19001515 to Legal Services on the 190xxxxx cost-centre range rather than its misleading description. [rubric 1.4] 1.7 We map code 20001713 to Management Consulting on the 200xxxxx cost-centre range. [rubric 1.5] 1.8 We keep code 22001105 (satisfaction survey) in Marketing & Communications / Market Research rather than letting it fall to HR. [rubric 1.6] PHASE 2: Build the spend picture Purpose: Apply the taxonomy to the transaction population and create a clean view of spend, contracts, supplier concentration, route-to-market, and timing. Rationale: We clean the data before we total it, because the retired 80-prefix rows and the unstandardised supplier variants would otherwise distort both the family totals and the concentration analysis; and we read contract and timing anomalies as evidence to be explained rather than numbers to be reported blindly. Steps: 2.1 We load spend_raw.xlsx and remap all 8 retired 80-prefix rows (AED 5.1M) using the Finance crosswalk from email_thread_1 before calculating any totals. [rubric 2.1] 2.2 We standardise the supplier-name variants for DWTC, Edelman, DMG, and Horizon before counting suppliers or computing concentration. 2.3 We apply the final taxonomy to every PO and summarise spend by family, subcategory, supplier, department, month, and sourcing route, then total classified spend to AED 263.6M. [rubric 2.2] 2.4 We identify Digital Technology as the largest family at AED 175.8M (66.7% of total). [rubric 2.3] 2.5 We assess Marketing & Communications contract coverage and find a blank-contract-reference rate of 56.2% (27 of 48 POs). [rubric 2.4] 2.6 We conclude those blank references are a system-linkage gap (agreements exist but ERP-Linked = No/Partial), not genuine off-contract spend. [rubric 2.5] 2.7 We identify DMG as the only supplier with no agreement in place, representing AED 6.6M of true uncontracted Events spend. [rubric 2.6] 2.8 We investigate the December concentration and explain it as the Microsoft Enterprise Agreement December true-up (cumulative AED 154.0M across Years 1-3, latest Year 3) rather than an unexplained anomaly. [rubric 2.7] 2.9 We state December spend as AED 154.8M (58.7% of total). [rubric 2.8] PHASE 3: Select the pilot category Purpose: Use the weighted framework plus governance constraints to select the highest-value eligible pilot category. Rationale: We score on addressable rather than raw spend, and we apply governance constraints explicitly, because the highest raw scorer is not eligible once conflicts and locked spend are accounted for; making that reasoning visible is what makes the final selection defensible. Steps: 3.1 We score every Level 2 category on the 11-criterion weighted framework against the cleaned spend picture, scoring spend impact on addressable rather than raw spend. 3.2 We remove the locked AED 154.0M Microsoft EA from Digital Technology, leaving AED 21.8M addressable (8.3%). [rubric 3.1] 3.3 We exclude Management Consulting for conflict of interest, because the incumbent consulting firms also advise the procurement transformation. [rubric 3.2] 3.4 We exclude Legal Services on the General Counsel's three reasons: confidential active matters, panel agreements running another year, and a review that signals instability. [rubric 3.3] 3.5 We note that Management Consulting is the highest raw scorer before constraints are applied. [rubric 3.4] 3.6 We select Marketing & Communications as the highest-scoring category still available after constraints. [rubric 3.5] PHASE 4: Select the pilot subcategory Purpose: Identify the best Marketing & Communications subcategory for the first category strategy. Rationale: We choose the subcategory on a combination of corrected spend size, sourcing competitiveness, and stakeholder evidence, separating the fixed venue from the contestable services so the strategy targets spend we can actually influence. Steps: 4.1 We compute corrected M&C subcategory totals and identify Events as the largest at AED 27.1M (Agency Fees 12.2M, Market Research 9.4M, Promotions & Advertising 2.2M). [rubric 4.1] 4.2 We use the teams_chat evidence to classify the blank-route Events rows as non-competitive. [rubric 4.3] 4.3 We calculate the Events non-competitive sourcing rate as 100.0%. [rubric 4.2] 4.4 We use interview evidence to confirm Events is CEO-facing, process-poor, and the largest budget-visibility problem, and we identify the main conference venue as fixed and government-hosted. [rubric 4.4] 4.5 We separate that fixed venue from the contestable execution services (stand build, AV, hospitality, logistics). [rubric 4.5] 4.6 We select Events as the pilot subcategory and position Agency Fees as a Wave 1 follow-on. [rubric 4.6] PHASE 5: Build the Events requirements brief Purpose: Turn the selected subcategory into a structured requirements brief with demand tiers, ownership, costs, and open governance issues. Rationale: We build the brief from the annual plan and stakeholder evidence, surfacing the budget-visibility gap and the governance breach explicitly, because these are the open issues that must be resolved before the category strategy can move forward. Steps: 5.1 We read the annual plan and identify the 12 FY26 demand items with TBC (unconfirmed) budgets. [rubric 5.1] 5.2 We use the Finance email template in email_thread_6 to request the missing budgets from the Finance Business Partner. [rubric 5.2] 5.3 We classify planned events into three demand tiers: Strategic Fixed, Planned Flexible, and Emerging/TBC. [rubric 5.3] 5.4 We flag the Global Manufacturing Summit supplier switch from DMG to DXBLive (about AED 3M) as a no-contract / no-procurement-process governance breach. [rubric 5.4] PHASE 6: Assemble and QA the strategy deck Purpose: Populate the provided nine-slide skeleton into a leadership-ready strategy deck and verify every headline figure ties back to the analysis before release. Rationale: We build the leadership deck last, drawing every slide from the evidence assembled in the prior phases, because the strategy's credibility rests on each headline figure tying back to the spend picture and the scoring answer; the final QA pass is what guarantees that internal consistency. Steps: 6.1 We populate all nine skeleton slides (Title, Executive summary, Why this pilot, Agenda, Spend picture, Internal requirements, Strategy, Initiatives, Roadmap) and remove every bracketed placeholder. [rubric 6.1] 6.2 We give the Executive summary slide an action title and lead with the AED 27.1M Events opportunity as the headline figure. [rubric 6.2] 6.3 We build the Spend-picture slide with a chart/data visual and a data-quality note covering the 80-prefix remap and the supplier-variant merge. [rubric 6.3] 6.4 We tie each priority initiative on the Initiatives slide to the specific finding that drives it. [rubric 6.4] 6.5 We sequence the Roadmap across three horizons (0-3 / 3-6 / 6-12 months) and list three next steps with named owners and dates. [rubric 6.5] 6.6 We QA the deck so every headline figure is internally consistent with the analysis (AED 263.6M total; Digital Technology 175.8M/66.7%; December 154.8M/58.7%; Events 27.1M; 100% non-competitive), then release the populated deck. [rubric 6.6, output: Output/3_Strategy_Deck_populated.pptx]