How does anyone actually know if an ad worked?

Marketers talk about attribution, lift, and mix models like everyone already knows what they mean. This is a place to slow down and find out — starting from a plain question someone might actually ask, and working through, step by step, how you'd go about answering it. Everything here uses made-up data, so there's nothing to get wrong.

Explore Solutions Open the Lab

Synthetic data only — no real customer data, no uploads, no auth

The learning journey

Four categories, one shared pipeline

Every solution — from onboarding first-party data to modelling the full marketing mix — follows the same path: business problem, methodology, synthetic data, model, and a business answer.

Start with a customer question

Customers describe problems in plain English, not methodology terms. Pick one to see how it's translated into a technical solution.

“Use our customer data for advertising”“Understand whether advertising caused incremental conversions”“Measure whether advertising drove physical store visits”“Understand which marketing channels drive revenue”“Find more people like our best customers”“Determine which customers are worth acquiring, not just cheapest to acquire”

How does anyone actually know if an ad worked?

The learning journey

Four categories, one shared pipeline

Data & Activation

Audience

Measurement

Investment & Strategy

Start with a customer question