Skip to main content

How Moonshot Benchmarks are calculated

Directional ranges + anonymized aggregates - designed for credibility, not hype.

Last updated: Jan 28, 2026 · Methodology v1.0

What we publish

  • Directional benchmark ranges by industry.
  • Anonymized aggregates from MedBay runs (when sample sizes are sufficient).
  • Confidence and threshold gating to avoid misleading small samples.

What we never publish

  • No business names.
  • No addresses.
  • No emails or phone numbers.
  • No free-text submissions.
  • No individual runs.

Thresholds

  • Overall aggregates shown only when total runs >= 200.
  • Per-industry aggregates shown only when runs >= 50.
  • Otherwise show "not enough samples yet".

Data handling & privacy

Benchmarks are calculated from anonymized MedBay runs. We aggregate the data, then publish only summary statistics such as medians or percentiles when possible. We also clamp extreme values to prevent outliers from distorting the result. No individual business data or free-text inputs are ever shown.

Citation

How to cite

Use this citation when referencing Moonshot Benchmarks.

Moonshot Benchmarks (2026). Retrieved from https://www.timelessrobots.com/benchmarks/

FAQ

Are these benchmarks statistically representative?

They are directional rather than statistically representative. The goal is to provide credible ranges and medians without overstating precision.

Why medians instead of averages?

Medians reduce the effect of outliers and better reflect typical outcomes for operational metrics.

How often are benchmarks updated?

Benchmarks are refreshed as new anonymized data is aggregated, with periodic methodology reviews.

Can I contribute data?

Benchmarks update automatically from anonymized MedBay runs. You can contribute by running MedBay with your own inputs.

What happens if sample sizes are small?

We suppress aggregates below the thresholds and show a neutral "not enough samples yet" note.