In 2023, a $20 billion Medicaid insurance contract award was protested — not because the winning bid was worse, but because the evaluation process had confirmation bias baked into its design. Evaluators scored suppliers they recognized more favorably on identical proposal content. One evaluator openly stated they would give more points to providers with out-of-state experience — the opposite of incumbency — while another applied the reverse logic. Same criteria, opposite interpretation, neither side aware of the bias driving their scores.
The protest was upheld. The award was at risk. The procurement officials had not intended to be biased. They had simply run an evaluation process with no safeguards against the most predictable error in human judgment.
The original concept: what confirmation bias is
Daniel Kahneman and Amos Tversky documented confirmation bias as a systematic error in how humans process information: we seek evidence that supports our existing beliefs and discount evidence that contradicts them. In Kahneman's framework from Thinking, Fast and Slow, this is System 1 thinking — fast, intuitive, pattern-matching — overriding System 2 analysis without the decision-maker noticing it happened.
The bias is not random. It is predictable. It operates most powerfully when three conditions are present: the decision is complex with multiple variables, the evaluator has prior exposure to one of the options, and the process allows subjective interpretation of criteria. Every supplier selection process hits all three conditions.
How confirmation bias operates in procurement: three stages
Research published in the International Journal of Physical Distribution & Logistics Management found that these biases operate even when procurement professionals believe they are making purely objective assessments. The bias is not a character flaw. It is a cognitive architecture problem. The brain is wired to confirm, not to challenge.
Where the analogy breaks down: procurement-specific complications
In psychology experiments, confirmation bias is studied in controlled settings with clean information. In procurement, the bias compounds across three layers that do not exist in a lab: organizational politics (the stakeholder who already told their boss which supplier they want), commercial pressure (the project deadline that makes re-evaluation expensive), and relationship economics (the supplier who took your team to dinner last month).
Academic debiasing research assumes the decision-maker wants to be objective. In procurement, stakeholders often have career incentives aligned with a specific outcome. A marketing director who championed a particular SaaS platform has reputational skin in the game. Asking them to "consider the opposite" is asking them to consider that they were wrong in front of their boss. The bias is not just cognitive — it is organizational.
What correct execution looks like: four debiasing techniques
Research from Kaufmann, Carter, and Buhrmann (2010) developed a taxonomy of debiasing techniques specifically for supplier selection. Four have the strongest evidence base for procurement applications.
Evaluators receive proposals, read them, discuss together, and reach consensus. Supplier identities are visible. Criteria are interpreted on the spot. Individual scores are not recorded.
Supplier identities are masked. Evaluators score independently using pre-published criteria with written justifications. Enhanced consensus follows: scores can be revised, but individual records are preserved.
Blinded evaluation — removing supplier identities, brand names, and incumbency markers from proposal documents — is the single highest-impact technique. Forbes documented a case where evaluators scored identical proposal content differently based solely on knowing which company submitted it. Blinding eliminates that variable entirely.
Independent individual scoring before group discussion prevents social confirmation from amplifying individual bias. When one evaluator voices a positive impression first, others unconsciously align. Requiring written scores before any discussion breaks this cascade.
The fourth technique is structural: standardized evaluation criteria published before bids arrive, with mandatory fields for every dimension — cost, quality, risk, sustainability, delivery. When evaluators must populate every field, they cannot selectively ignore data that contradicts their preferred outcome.
What this means in practice
Audit your last three supplier selection decisions. Pull the scoring sheets. Check whether evaluators scored independently or in a group discussion. Check whether supplier identities were visible during initial scoring. Check whether criteria were finalized before or after proposals arrived. If any of these checks fail, your process has a structural bias vulnerability — whether or not a bad decision resulted.
Implement one debiasing technique in your next sourcing event. Start with blinded evaluation: have an administrator remove supplier names, logos, and incumbency markers from proposal documents before distributing them to evaluators. This single change requires no new software, no policy approval, and no budget. It takes one person an hour.
Track the result. Compare scores from the blinded round against any previous unblinded evaluation of the same suppliers. The delta between the two — even if small — is the cost of confirmation bias in your organization's procurement decisions. Multiply it by your annual sourcing volume. That number is what you are currently paying for a cognitive bias nobody can see.
Can procurement technology eliminate confirmation bias?
Technology can reduce but not eliminate it. Automated scoring and data-driven shortlists remove some subjectivity, but algorithms can encode existing biases if trained on historical decisions that were themselves biased. The most reliable approach combines technology (automated initial scoring on objective dimensions) with process controls (blinding, independent scoring, written justifications) and human oversight.
How do I convince stakeholders to accept a blinded process?
Frame it as protecting the organization, not as distrusting individuals. The Medicaid contract protest is a useful reference: biased evaluation processes create legal liability. Blinding reduces protest risk and protects the procurement team from accusations of favoritism. Most stakeholders accept blinding when it is presented as risk management, not as a critique of their judgment.
What if a supplier's identity is genuinely relevant to evaluation?
Past performance, financial stability, and capacity are legitimate evaluation dimensions. The key is to evaluate them using objective data — audited financials, verified delivery records, reference checks — not brand recognition or personal relationships. Identity-relevant data can be presented without revealing which company it belongs to during the initial scoring round.
Sources: Kahneman, D., Thinking, Fast and Slow (2011); Kaufmann, Carter & Buhrmann, "Debiasing the supplier selection decision," IJPDLM Vol. 40 No. 10 (2010), emerald.com; Forbes / Gleb Tsipursky, "Prevent Costly Procurement Disasters: 6 Science-Backed Techniques" (March 2023), forbes.com; Payhawk, "Biases in Procurement: How to Make Fairer Decisions," payhawk.com; PMC / NIH, "The power of past performance in multidimensional supplier evaluation" (2024), pmc.ncbi.nlm.nih.gov. Accessed July 2026.