You're asking about real-world benchmarks. No one will give you an accurate comparison because they don't factor in the real cost.
That "cleaner, more actionable list" is an illusion if it ignores your monorepo's dependency overrides. You're just trading false positives for hidden blind spots.
Your real benchmark is operational overhead. Are you ready to feed Mend build artifacts forever? Can you afford the person-hours to learn Snyk's quirks? The pricing for both starts long before the invoice.
show me the bill
Exactly. That "operational overhead" you mention is just vendor code for "you're now the product manager for their data pipeline."
The real kicker is when that overhead shifts. Say you've built that perfect feed for Mend, and then they deprecate an API endpoint in their next integration update. Your "accurate" list is now broken until you re-engineer the pipeline. Suddenly you're not just maintaining your monorepo, you're maintaining their data ingestion spec.
Snyk's quirks are at least a known, static cost. You learn them once.
Trust but verify.
Your real-world benchmark is the one you already have: the false positives from your current scanner. Take that list and compare how Snyk and Mend would have handled those same dependencies.
We tried this with a complex AWS Lambda monorepo last year. We found Snyk's default reports were quieter, but they were missing some nested dependencies in our layer builds. Mend's initial report was noisy, but flagged issues Snyk missed because it could hook into our SAM build process.
The actionable list came down to which tool we could reliably map to our actual deployed artifacts. For us, that was Snyk with some custom CLI scripting, because our build process was too fluid for Mend's required pipeline integration. Your mileage will depend entirely on that mapping.
terraform and chill