Centralized vs Decentralized Procurement: How Multi-Project GCs Should Decide

The Decentralized Default

Most construction companies start decentralized. The first project needs rebar — the site PM calls the local supplier they have worked with before. It works. Second project, same thing. By the time a contractor is running five projects simultaneously, they have five separate vendor relationships, five different price lists for the same materials, and no easy way to see total spend across the portfolio.

This is not a failure of intent. Decentralization has real advantages: faster response times, local supplier relationships, and site teams who feel accountable for what they buy. A PM who sources directly has more skin in the game — they will push for better delivery timing because it is their project schedule at stake.

The cracks show at scale. A SAR 600M GCC contractor running six active projects, with structural steel pricing varying by just 6% between the best and worst site-level deal, loses SAR 1.4M annually — not through waste or fraud, but through the absence of consolidated buying power.

What Centralized Procurement Actually Means

Centralized procurement is often misunderstood as "head office decides everything." Done right, it means something more specific:

• The central team negotiates framework contracts and rate cards with approved vendors
• Projects raise requisitions against those contracts — they still control timing and quantity
• New vendor relationships and above-threshold purchases require central approval
• Spend data flows to one place, enabling portfolio-level analysis

What does not change: site teams still plan their own demand, initiate requisitions, and confirm delivery. The difference is that they are working from a pre-negotiated menu rather than going back to the market fresh on every order.

The Real Trade-Off — With Numbers

The debate comes down to three variables: speed, leverage, and visibility.

Speed

Site-level buying is faster for small, urgent purchases. A site team can call a local concrete supplier and have a truck on site the same day. A central procurement cycle might take three to five business days if approvals queue up. For urgent small orders, that lag is a real operational constraint.

Leverage

A SAR 500M contractor buying 8,000 tonnes of rebar across six projects should be one buyer, not six. That volume commands different pricing. If each project buys independently, the leverage disappears. For a contractor with SAR 200M+ in annual procurement spend, even a 3% improvement on top-10 materials represents SAR 6M — what a procurement function costs to run for several years.

Visibility

When every project manages its own vendor relationships, spend analytics become impossible. You cannot run a vendor performance review if you do not know that the same supplier is delivering to three of your sites at different service levels. You cannot negotiate next year's rates without consolidated volume data.

The Hybrid Model Most Mature GCs Operate

Experienced multi-project contractors do not choose a binary. They build a hybrid: strategic decisions centralized, tactical execution decentralized.

Central team owns

• Vendor prequalification and the approved vendor list
• Framework contracts and rate card negotiations for the top 20 materials by spend
• Any purchase above SAR 75K or involving a new vendor
• Monthly cross-site spend analysis and vendor performance scoring
• Annual vendor review and contract renewal decisions

Site teams own

• Demand planning and requisition timing
• Day-to-day supplier coordination — delivery scheduling, site access
• GRN confirmation: verifying quantities and quality at the point of receipt
• Local sourcing up to SAR 75K from the approved vendor list, within the approved budget code

The threshold is the key design decision. Set it too low — SAR 10K — and you create bottlenecks where every minor purchase queues a central approval. Set it too high — SAR 500K — and you have effectively decentralized everything that matters. Most GCC GCs in this model set site autonomy at SAR 50K–100K, calibrated to their average order size and team capacity.

The Infrastructure Requirement: One Shared System

A hybrid model only functions if data flows both ways. Site requisitions must enter the same system the central team sees. POs issued by either party must populate the same committed cost register. GRNs confirmed on site must close out the same PO, triggering the same accounts payable process.

Without that, the hybrid becomes theater. The central team negotiates good rates but has no visibility into whether sites are actually using them. You cannot measure framework contract compliance if contract coverage and actual spend live in separate places.

The approved vendor list is the critical connective tissue. Not a static spreadsheet, but a live register with:

• Prequalification status and expiry dates
• Commodity categories each vendor is approved for
• Live performance scores fed by delivery records, NCRs, and invoice exceptions
• Rate cards updated per negotiation cycle

When a site PM raises a requisition and selects a vendor from this list, they are executing within a negotiated framework — not making a fresh procurement decision. The system enforces the structure without requiring manual policing.

Red Flags That the Current Approach Is Not Working

If you are unsure which model you are actually operating, these signals indicate the drift has become costly:

• Vendor count exploding: 300+ active vendors across the portfolio with no visibility into how many supply the same commodities to multiple sites at different rates
• Price variance on common materials: The same product costs meaningfully more on one project than another and no one has flagged it
• Finance reconciliation overhead: Accounts payable is matching invoices to multiple contact names for the same supplier because each site called them independently
• No leverage at renewal: When steel prices soften, you cannot approach your main rebar supplier with consolidated volume data because the relationship is fragmented across six site PMs
• New vendors appearing regularly: Projects are bringing in unregistered suppliers because the approved list does not have anyone fast enough — a bottleneck problem the structure itself needs to solve, not a reason to abandon governance

Starting Steps for GCs Moving Toward a Structured Model

The shift does not require a system replacement. Start with the data you have.

1. Build a cross-site vendor spend report

Who supplied what, to which projects, at what total value, in the last 12 months. This baseline makes every future vendor conversation more credible — and makes the case internally for the structural change.

2. Identify your top 15 materials by spend

Concrete, rebar, formwork, aggregates, and MEP materials typically account for 60–70% of direct material cost. These are the commodities where centralized rate cards have the biggest financial impact.

3. Run a price variance audit

For each of those 15 materials, pull the prices paid across all projects in the last six months. The spread quantifies the opportunity — and is usually the number that gets a CFO's attention.

4. Set an approved vendor threshold

Define the value above which central approval is required, and build this into your requisition workflow so it is enforced automatically — not by policy reminders that get ignored under schedule pressure.

5. Consolidate the vendor register

Merge project-level vendor lists into a single register. Flag duplicates — same supplier, different contact names. Begin formal prequalification on the vendors with the highest spend across the portfolio.

The transition works best at contract boundaries. Enforce the new model on your next contract start. Carry forward legacy purchasing on existing projects but do not expand the vendor list outside the approved register.

Takeaways

Centralized and decentralized procurement are not philosophies — they are tools with different strengths at different scales. A two-project contractor buying locally is making a sensible decision. A twelve-project GC doing the same thing is leaving SAR millions of leverage unrealized and making vendor governance nearly impossible.

For GCC contractors above SAR 150M in annual procurement spend, the practical answer is a hybrid: central framework contracts and vendor governance, site-level execution within approved limits. That structure requires a shared system where every requisition, PO, and GRN flows to a single view — because without the data, the governance is just a policy document nobody reads.