Observed Signal · Jul 4, 2026 · Organizational Change · Source: DEV Community · Impact: 2/5 · Sentiment: Positive
Measurement & Analytics Market: Cutting QA Reduced Production Incidents — Case Study
A B2B analytics company that ships features 5–6 times daily removed its two-person QA team at the end of Q3 2025 and replaced classic UI-focused testing with a developer-owned quality process plus a new Data Quality team. Backend developers absorbed frontend checks via a Jira "Backend-test" status and took responsibility for verifying numbers on-screen against their outputs, while the Data Quality team built continuous input-monitoring and anomaly dashboards for the data lake. The company runs a ClickHouse-backed data lake (~100TB compressed) with ETL services and stores raw HTML on S3. Comparing two matched release cycles, created production incidents fell from 28→19 in release quarters and 51→26 in the post-release quarters after QA was cut, though the author notes sample size and concurrent Data Quality work limit causal attribution.
Shows an operational alternative to classic QA for data-centric B2B analytics: shifting responsibility to developers and creating a Data Quality function reduced production incidents in this case, which is relevant for analytics/MarTech operators considering testing strategy and data pipeline monitoring.
Key Takeaways & Evidence Grounding
- Company removed its QA/testing department at the end of Q3 2025 and redistributed responsibilities to engineering.
- A dedicated Data Quality team of backend developers was created to monitor ETL inputs, vendor integrations, and data anomalies.
- The product infrastructure includes raw HTML stored on S3 and a ClickHouse cluster; the data lake is ~100TB of compressed text accumulated over three years.
- Release cadence is 5–6 feature deployments per day.
- Created production incidents compared across two cycles: With QA (Q4 2024 → Q1 2025) vs Without QA (Q4 2025 → Q1 2026) — release quarter: 28 vs 19; quarter after release: 51 vs 26.
Connected Companies & Entities
4 Entities mappedAmazon
Global commerce, cloud, advertising and subscription platform company.
“HTML is downloaded and immediately compressed with zstd, then stored on S3 in that form for debugging and research purposes....”
ClickHouse
Open-source analytics database with managed cloud and observability.
“At the transformation stage, the HTML is parsed, broken down into TSV, enriched with internal data, and written directly to ClickHouse....”
Yandex
Russian internet platform with integrated advertising and monetisation stack.
“Post-click analytics data — collected on a schedule from external platforms on the client's behalf: Yandex Webmaster, Yandex Metrika, Yandex...”
Search, video, adtech and cloud giant within Alphabet.
“Post-click analytics data — collected on a schedule from external platforms on the client's behalf: Yandex Webmaster, Yandex Metrika, Yandex...”
Ontology Mapping & Concepts
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.