Specialised equity-crowdfunding platforms produce two signals about the same venture: a pre-campaign expert certification issued by the platform’s review panel and within-campaign traction generated by early investors. Using 99 rounds closed between 2016 and 2023 on Capital Cell, a life-science equity-crowdfunding platform, with outcomes recorded at the end of 2025, we study the predictive content of their disagreement, which we term signal discordance. The two evaluators rank ventures in systematically opposite ways ( , ). The expert rank is ordinally associated with long-run venture quality ( , ), whereas the traction rank is not ( , ). The expert association appears concentrated where the evaluators conflict ( in the discordant half against 0.119 in the concordant half), but the difference is not statistically significant (bootstrap 95% interval ), so we present this localisation as suggestive. On a survival outcome built from venture status alone, which does not depend on a later priced round, the difference is significant at the 5% level, although it rests on 15 write-offs. A soft quantum kernel with discordance-dependent softness, benchmarked against classical baselines and replicated on the Quantum Blue QPU at the Barcelona Supercomputing Center, is used as a methodological tool, not as a source of predictive advantage: long-run venture quality is close to unpredictable from campaign-time information by any method. Because the sample comes from a single doubly-screened platform, all results are predictive rather than causal.
When expert screening and crowd validation disagree: signal discordance and venture quality in life-science equity crowdfunding / Sáez-Ortuño, L., Sagarra, M., Forgas-Coll, S., Isgrò, V., Ferrara, M.. - In: SMALL BUSINESS ECONOMICS. - ISSN 0921-898X. - (2026), pp. 1-33. [10.1007/s11187-026-01293-6]
When expert screening and crowd validation disagree: signal discordance and venture quality in life-science equity crowdfunding
Ferrara, MassimilianoConceptualization
2026-01-01
Abstract
Specialised equity-crowdfunding platforms produce two signals about the same venture: a pre-campaign expert certification issued by the platform’s review panel and within-campaign traction generated by early investors. Using 99 rounds closed between 2016 and 2023 on Capital Cell, a life-science equity-crowdfunding platform, with outcomes recorded at the end of 2025, we study the predictive content of their disagreement, which we term signal discordance. The two evaluators rank ventures in systematically opposite ways ( , ). The expert rank is ordinally associated with long-run venture quality ( , ), whereas the traction rank is not ( , ). The expert association appears concentrated where the evaluators conflict ( in the discordant half against 0.119 in the concordant half), but the difference is not statistically significant (bootstrap 95% interval ), so we present this localisation as suggestive. On a survival outcome built from venture status alone, which does not depend on a later priced round, the difference is significant at the 5% level, although it rests on 15 write-offs. A soft quantum kernel with discordance-dependent softness, benchmarked against classical baselines and replicated on the Quantum Blue QPU at the Barcelona Supercomputing Center, is used as a methodological tool, not as a source of predictive advantage: long-run venture quality is close to unpredictable from campaign-time information by any method. Because the sample comes from a single doubly-screened platform, all results are predictive rather than causal.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


