At Pencil, we use advanced machine-learning models to provide predictions for your ads. This guide explains, in simple terms, how those predictions are calculated.
The Media Performance Score predicts how likely your ad is to perform well compared to others you have run. It is a number between zero and one hundred, representing a percentile ranking based on the estimated click-through rate (CTR). For example, a score of eighty-five means the ad is predicted to achieve a higher CTR than eighty-five percent of the Meta ads in your connected ad accounts.
The score learns only from your own ad data. To score an ad, Pencil finds the most similar ads already run from the Meta ad accounts connected to your workspace, then uses their real click-through rates to predict how the new ad will perform. Your score never draws on other customers' ads, and your ads are never used to score anyone else's work.
The score requires a minimum amount of historical ad performance data before it will appear. If your workspace does not yet meet this threshold, you will see "Insufficient data" in place of a score. For full detail on what the score shows, why you might see "Insufficient data", how to enable the score, and what improvements are on the way, see What does the Media Performance Score show?
The Media Performance Score is one component of Pencil's broader scoring framework. You can also create custom scores tailored to your brand. See How Can I Set Up Custom Scores in Pencil? for more.