Attribution models in 2026: what actually survives post-cookie
Last-click is dead, first-touch was always a lie. A practical comparison of multi-touch, incrementality tests, and the resurgence of marketing mix modeling.
Blog
From the blog
Practical notes on analytics engineering, attribution, and building data teams actually like.
Growth
Engineering
Anomaly detection on cloud cost time series without a data science team
How we cut false positives by 70% with seasonal decomposition and a very boring rules layer. Includes the SQL.
Benchmark
The 2026 B2B product analytics benchmark: 312 teams, 5 findings
What separates the top quartile from everyone else (spoiler: it's not tooling budget).
Company
Fluxboard raises Series A to make analytics infrastructure boring
Reliability is a feature. What we're building over the next 18 months.