Data quality is our number one priority. We spent years developing and iterating our proprietary tracking and stitching technologies
We believe we provide the most accurate attribution data available. That being said, accuracy will vary based on the attribution model you are using.
Click-Based Models offer extremely high accuracy/fidelity. These models are deterministic, they are based entirely on hard data points (UTMs, Pixels, Cookies, IP, IDs, etc.). That means that all revenue/transactions attributed are guaranteed.
That also means that these models might not capture the full picture - a lot of view data inherently doesn't exist (we can't read minds yet) therefore a lot of impressions go unattributed. These models will still show some organic, direct/none, unattributed traffic.
View-Based Models aim to fill in this gap. In our Clicks + Views model, we use machine learning to supplement click data and attribute organic traffic. Our goal is to source customers for the campaigns that acquired them. We do this by modeling the entire business - we are able to use trends, signals, and correlation to directionally source revenue and get much closer to reality.
Because we are using soft data, machine learning, and taking more liberties, these models will generally be less accurate. While the revenue isn't guaranteed down to the penny, these models are closer to reality. View-based models help us understand directionality and macro trends in the business.