What is your creative
telling you?
From performance data to your next creative decision.
Interactive prototype · v01
Evidence & missing context +
How this lab thinks Definitions & methodology ↗
Rules compare relationships, not universal industry benchmarks. A change of at least 20% versus a supplied historical rate activates a comparison rule. These are transparent screening heuristics, not causal estimates or significance tests. Hypotheses are ordered for investigation: message alignment, opening, proof, delivery, then buyer quality.
Below 1,000 impressions, the lab withholds diagnosis. Purchase-related rules also require 100 landing-page views. Fewer than 10 purchases adds an uncertainty warning. These thresholds do not guarantee reliability: use a planned controlled test, a mature attribution window and a sample-size calculation for decisions. Traffic and awareness objectives suppress purchase-specific hypotheses.