Verifiability Is a Necessary Condition, Not a Sufficient One
The digital credentials movement spent the better part of the 2010s solving an important but narrow problem: fake certificates. A paper certificate is trivially forgeable. A PDF badge is slightly harder but still vulnerable to simple image editing. Open Badges 2.0, and later Open Badges 3.0, addressed this by embedding cryptographic assertions in the badge metadata — anyone can verify a badge by calling the issuer's endpoint and checking the assertion payload against the earner's record.
That is genuine progress. But the field has been so busy celebrating verifiability that it hasn't asked the harder question: what, exactly, is being verified? A verifiable badge attached to a ten-minute module completion is still a completion badge. Cryptographic signing does not retroactively add rigor to the underlying activity. The authenticity problem and the meaning problem are distinct, and most platforms have addressed only the first.
What Credential Inflation Actually Looks Like
Credential inflation in the digital badge context operates differently than degree inflation but through a similar mechanism: when the supply of a credential outpaces the signal it was supposed to carry, the marginal hiring value of holding that credential drops toward zero.
Consider the trajectory of a technology skills badge from a major learning platform over a five-year window. In year one, the badge is awarded to learners who complete a structured course with a scored assessment — perhaps 30% of course starters earn it. By year three, completion-rate pressure leads the platform to soften the assessment. By year five, the badge is effectively awarded for 80% completion of the video content. The hiring managers who initially flagged the badge as a signal have learned to ignore it. The badge still verifies — it's still cryptographically authentic — but it no longer predicts anything.
This isn't a hypothetical. Any L&D practitioner who has spent time reviewing badge portfolios on professional profiles has encountered this pattern. The badges multiply; the interpretive signal diminishes. From a measurement perspective, this is a validity problem, not a technology problem. The badge metadata doesn't contain the information a verifier needs to evaluate the credential: what was the assessment, what was the passing standard, how was the passing standard validated, and how many people failed?
The Metadata That Actually Makes a Credential Defensible
Open Badges 3.0 defines a credential schema that supports richer assertions than its predecessors. A compliant badge payload can include the assessment criteria, the evidence of achievement, the alignment to competency frameworks (such as IMS Global's Credential Engine or CASE framework nodes), and the issuer's verification endpoint. In principle, this is the apparatus needed to make a badge interpretable.
In practice, most issuers populate only the minimum required fields: earner name, badge name, issue date, issuer URL. The richer metadata fields — criteria, evidence, alignment — are either left empty or filled with boilerplate. A hiring manager or academic reviewer pulling the badge's JSON-LD assertion via the verification URL sees only that the learner completed something, on a specific date, at a specific organization. That's a participation record, not a competency claim.
What a defensible credential requires at minimum:
- A description of the assessment or demonstration standard used to award the credential
- A passing threshold stated in terms of the underlying measurement model — a raw percentage or a scaled score with a defined cut point
- A pass rate or difficulty indicator that contextualizes the threshold (a 90% pass rate means something very different from a 40% pass rate)
- A validity period where applicable, especially for competencies that decay
The Comprehensive Learner Record (CLR) standard from 1EdTech (formerly IMS Global) goes further by linking multiple credentials and learning experiences into a verifiable narrative — not just "this person passed this assessment" but "this person demonstrated these competencies across these contexts." CLR adoption is still limited to early-moving institutions, but the directional pressure is clear: the market is demanding more interpretive signal, not less.
The Hiring Manager's Problem
Think about the decision a hiring manager faces when reviewing a candidate with fifteen digital badges. To extract value from that portfolio, they need to evaluate each badge's issuer reputation, understand the underlying assessment standard, and adjust for the possibility that any given badge is a completion certificate rather than a mastery certificate. The cognitive cost is high. The default response is to ignore most of the portfolio and fall back on the degree, the employer brand, and the personal interview — exactly the signals that the open credentials movement was supposed to augment.
The result is a credentialing ecosystem that costs learners real time and often real money to populate, but delivers limited incremental signal to credential consumers. This isn't saying the digital credentials movement was a mistake — the infrastructure built over the past decade is genuinely valuable. What we're saying is that infrastructure without rigor produces no durable signal. A distributed ledger full of completion badges is not a trustworthy credentialing system; it's a structured archive of learning events.
What Rigor Requires in Practice
Building a credential that a hiring manager can interpret requires decisions at the assessment design level, not the badge metadata level. Specifically:
The assessment must be scored against a defined standard, not a curve. Criterion-referenced assessment — where the passing threshold is defined by what the learner can do, not how they rank relative to peers — is the appropriate model for competency credentials. Norm-referenced scoring (percentile ranks) is useful for selection; it is not appropriate for credentialing mastery.
The passing standard must be validated. A cut score of 70% is not a defensible standard unless someone has conducted a systematic review of what 70% performance actually implies about competence. Standard-setting procedures — modified Angoff, bookmark method, or body-of-work review for performance-based assessments — exist precisely to anchor cut scores to operational competency definitions rather than arbitrary thresholds.
The assessment must be protected against gaming. A credential earned through an open-book, untimed assessment of factual recall is measuring something different from a credential earned through an adaptive instrument under controlled conditions. This doesn't mean every credential needs high-stakes proctoring — context matters — but the assessment design choices need to be disclosed in the credential metadata so the consumer can make an informed judgment.
The technology stack for digital credentials is sufficiently mature. The remaining obstacle is institutional will to build assessment infrastructure that earns the credential metadata rather than just issuing it. Organizations that make that investment are positioned to have their credentials noticed in a noisy marketplace. Those that don't are adding to the noise.
