1. Introduction
Procurement organisations are perpetually caught between two gravitational pulls. On one side, consolidation: fewer suppliers mean greater leverage, lower transaction costs, deeper relationships, and better data. On the other side, diversification: more suppliers reduce dependency, create competitive pressure, and hedge against disruption. The COVID-19 pandemic, the semiconductor shortage of 2020–2022, and the Red Sea shipping disruptions of 2024 forced many organisations to reexamine supplier count decisions that had been made, largely implicitly, during decades of relative stability.
The question "how many suppliers per category?" does not have a universal answer. But it does have a structured answer — one that varies systematically with the nature of the category, the characteristics of the supply market, and the strategic posture of the buying organisation.
This article reviews the three principal sourcing configurations — single-source, dual-source, and multi-source — and proposes a practical decision framework grounded in four variables: risk, volume, complexity, and switching cost.
2. Definitions and Typology
Single-sourcing refers to the deliberate allocation of 100% of a given category's volume to one supplier. It differs from sole-sourcing, where only one supplier exists in the market; single-sourcing is a choice, not a constraint.
Dual-sourcing divides volume between exactly two suppliers. The split may be symmetric (50/50) or asymmetric (e.g., 70/30 primary-to-secondary), with the secondary supplier maintained primarily as a contingency or competitive check.
Multi-sourcing distributes volume across three or more suppliers. This model dominates in categories characterised by high spend, commodity-like supply, or significant supply risk, and is common in raw materials procurement, logistics, and IT services.
Each model represents a distinct position on the consolidation–diversification continuum (Kraljic, 1983; Treleven & Schweikhart, 1988).
3. The Academic Backdrop
The theoretical foundation for supplier count decisions traces to Kraljic's (1983) seminal portfolio matrix, which classifies spend categories along two axes — profit impact and supply risk — and prescribes different supplier relationship strategies for each quadrant. While Kraljic did not explicitly address supplier count, subsequent literature has extended his framework in that direction.
Treleven and Schweikhart (1988) provided an early formal treatment of single vs multiple sourcing, modelling the trade-off between volume-based cost savings and supply interruption risk. They found that single-sourcing is optimal under low probability of supply disruption and high economies of scale, while multiple sourcing dominates when disruption risk is high and scale economies are modest.
Tang (2006) advanced the conversation with a broader framework for supply chain risk management, arguing that supply base diversification is one of several "robust strategies" for managing uncertainty. His analysis suggests that the value of a second or third supplier is not linear: the marginal risk reduction of adding a supplier diminishes rapidly beyond two or three sources, making extreme multi-sourcing often wasteful.
Berger, Gerstenfeld, and Zeng (2004) formalised the decision using decision-tree modelling, concluding that the optimal number of suppliers depends on the probability and cost of supply failure, the volume discount structure, and the administrative cost of maintaining supplier relationships — findings that directly inform the framework presented here.
More recently, Choi and Krause (2006) warned of what they termed supply base complexity: as the number of suppliers increases, the cognitive and administrative burden on procurement grows non-linearly, reducing the organisation's ability to manage any individual relationship well. Their empirical work found that buyers frequently underestimate this cost.
On the operations side, Dyer and Singh (1998) and Liker and Choi (2004) demonstrated that deep, trust-based supplier relationships — which require concentrated, single or dual-source arrangements — generate proprietary advantages in innovation, quality, and responsiveness that arm's-length multi-source relationships rarely achieve.
4. The Four Decision Variables
The framework proposed here centres on four variables, each of which pushes the optimal supplier count in a predictable direction.
4.1 Supply Risk
Supply risk encompasses the probability and severity of supply disruption, including supplier financial instability, geographic concentration, natural disaster exposure, geopolitical dependency, and single-point-of-failure characteristics. It also includes demand-side concentration risk: a supplier that serves many buyers may deprioritise any one of them during scarcity.
Direction of effect: Higher supply risk increases the optimal supplier count. When the consequences of a supply failure are severe — measured either in revenue loss, production downtime, or customer impact — the insurance value of a backup supplier becomes substantial. The 2021 automotive semiconductor shortage, during which companies like Toyota fared better than peers partly due to its historically unusual practice of holding multi-month component inventories and qualifying secondary suppliers, is often cited as evidence (Holweg & Frits, 2004; The Economist, 2021).
Practitioners often assess supply risk using the SCRM (Supply Chain Risk Management) matrix or the Probability × Impact methodology described by Zsidisin and Ritchie (2009).
4.2 Spend Volume
Volume (or spend) determines the absolute magnitude of cost leverage available through supplier consolidation. A high-spend category offers large absolute savings from volume discounts, preferential terms, and dedicated capacity — benefits that accrue disproportionately to the single-source buyer. A low-spend category may not justify the administrative overhead of managing multiple suppliers at all.
Direction of effect: Higher spend volume pulls toward consolidation (single or dual source) to capture leverage, unless risk is also high — in which case the cost of diversification is justified by the exposure it hedges. Conversely, low-spend categories should be consolidated aggressively: maintaining multiple suppliers for a €50,000/year category is almost never economically rational.
Spend volume interacts with market concentration. In categories where a dominant supplier controls a large share of global capacity (e.g., rare earth metals, certain specialty chemicals, ASML lithography equipment), single-sourcing may be unavoidable even at high spend levels, making risk mitigation through inventory or design substitution the only available lever.
4.3 Product/Service Complexity
Complexity captures how difficult it is to specify, develop, and qualify a supplier's output. It includes technical complexity (tolerances, certifications, proprietary processes), relational complexity (customisation, co-development, tacit knowledge), and integration complexity (how tightly the supplier's output is embedded in the buyer's operations or product).
Direction of effect: Higher complexity pushes toward fewer suppliers. Qualifying a new supplier for a complex aerospace subassembly may take two to four years and cost millions of euros in engineering and certification work. The prospect of replicating this across multiple suppliers simultaneously is often economically prohibitive. Complex categories therefore tend toward single or dual sourcing, with the second source maintained as a qualified contingency rather than an active volume recipient.
Low-complexity categories — MRO supplies, standard packaging, commodity chemicals, generic IT hardware — carry no meaningful qualification burden and can be multi-sourced with little friction.
4.4 Switching Cost
Switching cost is the fully loaded cost of replacing an incumbent supplier: not merely the cost of qualifying a new source, but the transition costs, productivity loss during changeover, risk of quality or delivery degradation, and any contractual penalties. It is closely related to complexity but distinct from it: a supplier may be technically simple to replace yet still costly to switch due to ERP integration, just-in-time scheduling dependencies, or long contractual lock-in.
Direction of effect: High switching cost reduces the practical benefit of multi-sourcing. If moving volume between suppliers takes 12–18 months and requires significant operational disruption, the theoretical resilience of a multi-source strategy may be illusory — the backup supplier cannot be activated quickly enough to prevent damage in most realistic disruption scenarios.
High switching cost also reduces the effectiveness of using multiple suppliers as competitive leverage: a buyer who cannot credibly threaten to switch cannot negotiate from strength, regardless of how many suppliers are nominally on the approved vendor list.
5. The Decision Matrix
The four variables can be operationalised into a scoring matrix. Each variable is rated on a three-point scale (Low / Medium / High), and the resulting profile is matched to a recommended sourcing configuration.
┌───────────────────┬──────────────────────────────────────────────────────────┐
│ VARIABLE │ LOW MEDIUM HIGH │
├───────────────────┼──────────────────────────────────────────────────────────┤
│ Supply Risk │ → Single → Dual → Multi │
│ Spend Volume │ → Single/Dual → Dual → Single/Dual* │
│ Complexity │ → Multi → Dual → Single │
│ Switching Cost │ → Multi → Dual → Single │
└───────────────────┴──────────────────────────────────────────────────────────┘
* High volume + high risk = Dual, not SingleThe aggregate signal across the four dimensions determines the recommended configuration. A category scoring High Risk / High Volume / Low Complexity / Low Switching Cost — a classic commodity with strategic exposure — warrants multi-sourcing. A category scoring Low Risk / High Volume / High Complexity / High Switching Cost — such as a proprietary engineered component from a financially stable, geographically diversified supplier — warrants single-sourcing with a robust contingency plan.
6. Practical Guidance by Configuration
6.1 When to Single-Source
Single-sourcing is appropriate when:
The supplier has a proprietary capability that cannot be replicated without prohibitive investment (technology, regulatory approval, geographic access).
The complexity of qualification makes duplication impractical within the planning horizon.
The supply risk is demonstrably low (financially robust supplier, diversified supply chain, long track record).
Deep co-development, joint IP creation, or lean/JIT integration is required.
When single-sourcing, risk mitigation must be achieved through other means: supply assurance agreements with volume and delivery guarantees, buffer inventory strategies, "design-to-substitute" product architecture that enables emergency substitution, and rigorous supplier financial monitoring (Cousins et al., 2008).
6.2 When to Dual-Source
Dual-sourcing is the most versatile configuration and the default choice for many strategic categories. The 70/30 asymmetric split is the most common real-world pattern: it preserves most of the volume leverage of single-sourcing while maintaining a qualified, commercially active secondary source that can scale rapidly in disruption scenarios.
The secondary supplier must receive sufficient volume to remain technically current and operationally capable. A supplier allocated 5% of volume may not invest in the relationship, may lose process familiarity, and may be unable to absorb a volume surge when needed. Research suggests that the secondary supplier's share should not fall below 15–20% to maintain meaningful contingency value (Berger et al., 2004).
Dual-sourcing is also effective as a commercial lever: the existence of a qualified second source creates competitive tension that single-sourcing eliminates.
6.3 When to Multi-Source
Multi-sourcing is appropriate in categories where:
Supply risk is high and geographically concentrated (e.g., rare earth elements from China, specialty chemicals from a single industrial zone).
Products or services are commodity-like, with minimal differentiation between suppliers and negligible switching cost.
The market is fragmented and no single supplier can satisfy total demand.
The organisation's procurement capability can handle the relationship management overhead.
The principal risk of multi-sourcing is relationship thinning: spreading volume across many suppliers reduces leverage and commitment from each, and can prevent the deep partnerships that drive supplier innovation and preferential treatment. Choi and Krause (2006) recommend actively managing the supply base down to the smallest number that still achieves the risk and competition objectives.
7. Dynamic Considerations
The optimal sourcing configuration is not static. Several factors make it appropriate to revisit supplier count over time:
Market maturation. A category that begins as highly complex (requiring single-sourcing) may become more standardised as technology matures, enabling gradual expansion to dual or multi-source.
Geopolitical shifts. The reshoring and friend-shoring trends accelerated after 2020 have prompted many organisations to add regional backup suppliers to categories previously single-sourced from Asia (Simchi-Levi & Haren, 2022).
Supplier performance deterioration. A single-source supplier experiencing quality or financial stress is the clearest trigger for urgent dual-source qualification.
Spend growth. As a category's spend volume grows, the absolute cost of supply failure grows with it, potentially warranting a shift from single to dual source even without any change in the underlying risk profile.
The framework should be applied as an annual or biennial portfolio review process, not a one-time decision.
8. Integration with the Kraljic Matrix
The framework maps intuitively onto Kraljic's (1983) four quadrants:
Routine/Non-critical (low risk, low impact): Single or dual source; consolidation priority.
Leverage (low risk, high impact): Single or dual source; volume leverage maximised.
Bottleneck (high risk, low impact): Dual or multi-source; risk reduction priority even at cost.
Strategic (high risk, high impact): Dual source with deep partnership; balance leverage and resilience.
The key contribution of the present framework is to disaggregate supply risk into its constituent drivers, and to add complexity and switching cost as explicit variables that Kraljic's two-axis model does not capture.
9. Common Pitfalls
The "preferred supplier" trap. Many organisations nominally maintain two or three suppliers per category, but direct 90%+ of volume to one. The secondary suppliers atrophy, lose capability, and provide no real resilience. Dual-sourcing requires active volume management.
Qualification without activation. Approved Vendor Lists frequently contain suppliers that have never received a commercial order. A supplier that is qualified but dormant provides false comfort; it will not be able to ramp quickly in an emergency.
Ignoring total cost. Multi-sourcing appears to reduce cost through competitive tension, but the total cost — including supplier management, quality harmonisation, logistics complexity, and systems integration — often exceeds the savings. Procurement organisations should calculate the total cost of sourcing configuration, not just unit price differentials.
Static decisions. Category sourcing strategies are often set once, at contract signature, and never revisited. The dynamic considerations outlined above make this a significant governance risk.
10. Conclusion
The question of how many suppliers to maintain per category is not a binary choice but a continuous decision problem shaped by four key variables: supply risk, spend volume, product/service complexity, and switching cost. Single-sourcing maximises leverage and enables deep partnership at the cost of resilience. Multi-sourcing maximises resilience and competitive tension at the cost of leverage and relationship depth. Dual-sourcing, in its asymmetric 70/30 form, is the most versatile configuration for strategic categories, offering a practical balance between both objectives.
The framework presented here provides procurement professionals with a structured, repeatable methodology for making — and periodically revisiting — this decision across a category portfolio. In an era of heightened supply chain volatility, the supplier count decision has moved from a tactical footnote to a strategic imperative.