VercatusSeparating observed changes from causal attribution
SEO/GEO changes should be evaluated as experiments: treated vs matched control page cohorts (same template, similar baseline) analysed with difference-in-differences, synthetic control or Bayesian structural time-series (CausalImpact), because seasonality and algorithm updates confound before/after comparisons. These methods estimate the counterfactual, what the treated pages would have done without the change, from untreated series.
Sources
- 1Brodersen et al. · 2015Inferring causal impact using Bayesian structural time-series models (Annals of Applied Statistics 9, 2015)research.google
- 2
BGrade B: peer-reviewed or a large controlled study. A grade describes the provenance of the sources above; it is not a probability that a change will work. The effect basis says how the size of the effect is known: measured by a study, taken from an industry report, only documented without a size, or not quantified.
Put this evidence to work
Related entries
More in Measurement, or search the whole base.
All Measurement entries