Schema drift is the single largest data-pipeline maintenance category — ~31% of maintenance time per Fivetran 2026 benchmark
Summary
Claim: Pipelines break on upstream schema changes — schema drift is the single largest maintenance category, ~31% of maintenance time per Fivetran 2026. (Government feeds also have publication lags — Census ACS lags; BTOS is biweekly; the Bank of Canada notes brief processing delays in Valet.)
Source: Fivetran 2026 benchmark (above); estuary.dev; rudderstack.com (vendor framing, consistent across sources).
Confidence: Industry-consensus (vendor sources align).
Why this matters for Candid: Concrete mechanism behind Fivetran 2026 Enterprise Data Infrastructure Benchmark — data teams spend 53% of engineering time on maintenance; $2.2M/yr/team on pipeline upkeep at enterprise scale — when upstream changes a column type, every dependent pipeline breaks. The SMB-relevant counter-pattern is "use stable APIs (FRED, Valet, GTFS, BoC) and don't over-engineer."
Related entries
Referenced by (3)
- reference Research brief: live data and data-driven tools for SMBs — when it's an edge, when it's overkill (June 2026) relates-to
- rule R5 — Budget for pipeline maintenance from day one; if the client can't commit to upkeep, rent the managed version instead of building one depends-on
- reference Dashboard rot: data sources change schemas, metric definitions drift across departments, organizational attention wanes; custom/embedded builds carry the heaviest ~20-30%/yr maintenance burden relates-to