60 SaaS APIs graded on data engineering friendliness: can you actually get your own data out?
Each dimension is scored 0–5 against the current official docs (cited by URL in every scorecard) and the production connector code of ingestr; when docs and code disagree, the code wins. Scores are weighted by the source's data volume class, judged by the highest-volume table users actually sync (weights shown per dimension above as small / medium / large): missing incrementality on a 10k-row config object is a tax, on billions of ad rows it is disqualifying. The flagship table sets the ceiling: pain on the endpoint everyone needs is never averaged away by nice minor endpoints. Total = Σ(score × weight) ÷ 5. Grades: A 90+ · B 75–89 · C 60–74 · D 40–59 · F <40. Full anchors and rules: RUBRIC.md.