Measuring Social Impact Across a Supply Chain

Social issues caused 18% of the more than 180,000 material supply chain disruptions tracked globally between January 2023 and October 2025, a significant share of all material disruption events, and the second consecutive year that social factors have accounted for that share. That figure, from Prewave's State of Social Procurement report, is the reason this piece exists. Labour strikes, protests, and worker demonstrations tied to wages and cost-of-living pressure pushed disruption numbers to their highest monthly level in the dataset in March 2025, and trade friction and new tariffs have only sharpened the underlying tension. Companies know social risk is material. Few of them can actually measure it well enough to act on it before it turns into a shutdown, a boycott, or a headline.

What "measuring social impact" shows, and what it doesn't

Three different activities get lumped under the same phrase, and that confusion alone explains a lot of the wasted spend on social measurement programs.

Assessing social risk involves identifying where harm could plausibly happen, before it does. Measuring social performance means tracking how a given supplier actually manages labour conditions, safety, and rights day to day. Measuring social impact means something narrower still: the outcomes for workers and communities that result from all of it. A supplier can score well on performance metrics (audits passed, policies documented) while producing weak impact (wages still below a living-wage benchmark, turnover still high). Treating these three as interchangeable is how a company ends up with a stack of green checkmarks and a workforce that's quietly miserable.

Supply chain sustainability work generally splits into four buckets: environmental impact, labour and human rights, ethics, and sustainable procurement. This piece stays inside the second and fourth, the social side, not the carbon side. That's a deliberate narrowing, because the tools, KPIs, and verification methods for social risk differ enough from environmental metrics that conflating them produces mushy, unusable dashboards.

Scope is the other thing that gets glossed over. EcoVadis data shows 95% of companies have visibility into their tier-1 suppliers, but that visibility only extends to tier 2 or beyond for 42% of companies. So a company measuring "its supply chain" while only looking at tier 1 may, in practice, be measuring less than half of it. And the academic literature hasn't caught up: a 2026 bibliometric review in MDPI found that KPI research on sustainable supply chains still skews heavily toward economic and technological measures, leaving social sustainability, organizational capability, and governance in small and mid-size supplier networks under-studied. The tools lag the problem.

The structural reasons measurement stays difficult even when intent is strong

Intent isn't the bottleneck. Most procurement leaders (a large majority, per procurement survey data) say ESG factors matter in how they choose suppliers. Yet 85% say finding sustainable suppliers is difficult in practice, and only 17% of suppliers report feeling any real pressure from customers to act on sustainability. That's a confidence gap wide enough to drive a container ship through: buyers believe they care, suppliers don't feel the pressure, and somewhere in between, nothing changes.

Part of the gap is just data. Most supply chains run on supplier self-reporting, not independent verification, because there's no shared infrastructure forcing anything else. A supplier fills out a questionnaire, the buyer files it, and everyone moves on until the next audit cycle. Tier opacity compounds the problem: visibility drops off sharply below tier 1, and tier 2 and tier 3, exactly where extraction, raw material processing, and the roughest labour conditions tend to concentrate, are largely unmapped by the companies that ultimately sell the finished product.

Methodology inconsistency turns fragmented data into actively misleading data. The clearest parallel comes from carbon accounting, not social metrics, but the mechanism is identical. An analysis of 871 European CSRD reports from 2025 found that 11% of reports showing a Scope 3 figure reported it as smaller than Scope 1 or 2 combined, a near-impossible outcome given Scope 3 emissions are typically many multiples larger. That's not noise: it's a sign that companies are applying wildly different definitions to the same line item. Social KPIs suffer the same disease: "worker grievance" means something different at every company that reports one, and until definitions standardize, cross-company comparison stays close to meaningless.

Choosing what to measure: a practical set of social KPIs organised by purpose

Ask what decision this number will change, not just what can be measured." A KPI that sits in a sustainability report and never touches a sourcing decision isn't doing any work, no matter how well it's defined.

Risk identification KPIs come first, because they tell a company where to look before anything else. The number of suppliers audited for labour and human rights compliance, the percentage of spend flowing to high-risk geographies or sectors, and the count of confirmed incidents involving child labour or forced labour all fall here. A workable principle: prioritise high-risk-region suppliers for the most frequent audit cycles, rather than letting them cycle through on the same multi-year schedule as lower-risk relationships.

Performance and management-system KPIs come next, and they answer a different question: is the supplier actually set up to catch problems, or did it just pass a snapshot audit? Assessment tools such as ESCP's SEIA rate management system maturity rather than point-in-time compliance, which matters because a supplier can look clean on the day an auditor shows up and still have no real system for catching abuse the rest of the year. Percentage of spend with ethically certified or independently assessed suppliers, plus corrective action closure rates following an assessment, round this category out.

Outcome and impact KPIs are the hardest to build and the most important to have. Total social investment as a share of profits gives a hard number to track over time. Worker wellbeing indicators, wages measured against local living-wage benchmarks, injury rates, and turnover, get closer to lived experience than any policy document can. Community benefit indicators for social enterprise suppliers close the loop, connecting procurement spend to measurable change outside the factory gate.

Setting the scope of assessment: tiering, geography, and materiality

Start at tier 1, because that's where visibility already exists, but don't stop there out of habit. Spend concentration and risk concentration should decide which tier-2 and tier-3 relationships get assessment budget, because that's where the real exposure usually sits. Significant social risk in global supply chains concentrates in extraction, agriculture, and manufacturing tiers that the buyer never signs a contract with directly, several steps removed from the brand name on the box.

Geography works as a fast materiality filter. High-risk country lists published by international bodies and national regulators exist precisely so companies don't need to map an entire supply chain before deciding where to spend limited assessment resources. Category matters just as much: electronics assembly, apparel manufacturing, food production, and construction each carry a structurally different social risk profile, and a single KPI set applied across all four will underweight the risks specific to each.

Regulation sets a floor, not a target. One major sustainability reporting law directly affects more than 50,000 companies, with thousands more touched indirectly through supply relationships. For companies inside its scope, CSRD defines the minimum disclosure required by law. Materiality-based scoping, built around where the actual risk and spend concentrate, should sit above that floor, because compliance minimums were never designed to produce the most useful data, only the most legally defensible one.

Verification approaches: what remote assessment, audits, and network analysis each can and can't do

On-site audits carry the most credibility on paper, and the highest cost in practice. They're expensive to run at scale, and they suffer from a timing problem: conditions on the ground can shift in the weeks after an auditor leaves, and suppliers given advance notice have every incentive to stage conditions for the visit itself. An audit is a photograph, not a video.

Remote, system-based assessment tries to solve for that by scoring management maturity instead of a single moment in time. ESCP's SEIA model is the clearest example: its 2026 Insights Report covers 516 completed assessments across 37 countries, and every single company that went through reassessment showed a score improvement the second time around. That 100% figure suggests the model is generating genuine behavioural change.

Network-level analysis takes a different angle entirely, treating social harm as something that spreads through trade relationships rather than sitting isolated at a single supplier. A PageRank-style framework introduced in a University of Exeter paper models harm propagating outward through a network, which makes it useful for catching accumulated risk and greenwashing patterns that a one-off audit, focused on a single node, would never surface.

AI is showing up everywhere in procurement risk workflows now, scanning for compliance issues and flagging exposure automatically. None of that changes the underlying constraint: an AI system is only as reliable as the supplier data feeding it, and most of that data has never been independently verified by anyone. Faster processing of bad data is still bad data, just delivered with more confidence.

Annual audit cycles also move too slowly for how fast social risk actually shifts. March 2025's spike in disruption events emerged far faster than an annual audit cycle could detect. Continuous monitoring tools that track strikes, protests, and regulatory action in near real time won't replace periodic assessment, but they close the gap between when a risk emerges and when a company finds out about it.

Worker voice belongs in the verification mix as its own layer, separate from everything above. Management scores and audit findings capture what management chooses to report. Worker surveys, grievance mechanisms, and independent worker interviews capture something else: the daylight between the policy on paper and what people actually experience on the floor. Skipping that layer means a program can look complete while missing the exact thing it exists to catch.

Turning assessment data into decisions: from measurement to action

Only 13% of businesses fully embed resilience KPIs into their strategy, yet that small group recovers faster from disruption, works better with suppliers, and holds more stakeholder trust than the rest. The remaining majority aren't necessarily failing to collect data; they're failing to route it anywhere that matters. They're failing to route it anywhere that matters.

Three decisions should sit downstream of any measurement program. Supplier development is the first: assessment findings should point toward capacity-building support. SEIA's corrective action and coaching model works this way by design, turning an assessment into an improvement plan rather than a verdict. Sourcing decisions come second, feeding social risk scores directly into supplier selection, spend concentration limits, and preferred-supplier lists. This is where measurement finally touches the actual leverage a buyer holds. Portfolio-level prioritization is the third and broadest: aggregating assessment data across suppliers to spot systemic risk in a geography, a category, or a tier, shifting the response from fixing one supplier at a time to redesigning the structure that keeps producing the same problem.

Social procurement can also work as a positive lever, not just a defensive one. The Rise Ahead Pledge counts 25 signatories that have collectively put USD 525 million into social innovation since it launched, with social procurement as the single largest category of that spend at 36% of the total. That's procurement functioning as an impact mechanism in its own right, not merely a risk filter sitting upstream of a purchase order.

The last link in the chain is assurance. A growing share of large corporations now get independent assurance on their sustainability disclosures, and assurance only works if the underlying data can survive scrutiny. Measurement data that's been independently verified is what makes that assurance possible in the first place, closing the loop between the numbers a company uses to make internal decisions and the numbers it puts in front of regulators, investors, and the public.