Attribution is not evidence

Ad platforms are useful, but their reporting incentives are not neutral. Each platform is designed to show the value it believes it contributed, often using its own view of the customer journey.

That can help with optimisation inside a channel. It is weaker when the business needs to decide whether spend created incremental revenue or simply touched demand that already existed.

Incrementality asks a harder question: what would have happened without this spend, region, creative, or intervention? That counterfactual is what budget decisions need.

Geo-lift tests, holdouts, baselines, and Bayesian counterfactual methods help estimate whether a campaign changed outcomes beyond normal movement.

If evidence is strong, the business can scale with more confidence. If evidence is weak, the right action may be to hold, cut, or retest. If the test is underpowered, the next experiment may be the most valuable recommendation.

OpenLift exists to make those decisions clearer. It helps teams move from platform credit to business evidence.

Not every signal deserves the same level of trust. A correlation can be useful for exploration, but a budget change may need stronger proof.

Evidence tiers make that distinction visible inside the recommendation, so operators and leadership can see why the system is confident or cautious.