The Feature Audit in June is not intended as an interpretation of the features but as a way to measure which of your features are used the most.
Popularity and Frequency results
Your Feature Audit corresponds to two exact numbers from 0% to 100% for each of your feature. One number is the "Popularity", the other one the "Frequency". You get one of each for every feature.
They reflect how often and which portion of your customers interact with a feature. A high popularity and frequency indicates that the feature is used by most of your customers most of the time.
💡 Note: Your overall number of customers evolves on a daily basis so this score is computed dynamically.
Principles used to compute "Popularity"
The popularity is the percentage of your customers that interacted with a feature.
The popularity ratio only applies to customers that can be identified, i.e. your logged-in users
The popularity ratio is calculated on a per week basis to smooth out the effect of weekly seasonality. All products have seasonalities. What we know is that humans mostly tend to have different agendas on weekends.
Popularity starts to be computed when a feature was introduced - in other words when we received a tracking event for the first time.
Total customers who used a feature in a week / Total customers in week
Principles used to compute "Frequency"
The frequency is the percentage of times a customer interacted with a feature out of the times they used all other features.
Frequency starts when a feature was introduced - in other words when we received a tracking event for the first time.
(Total times a feature is used / Total times all features are used) averaged across all your customers
Remove system events
In case you have non user triggered events (ex: a notification sent a user), you can remove them from your feature audit.
To do so head to your event list and mark the event as "automated".
Non user triggered events (also called "automated") won't bias your feature audit anymore.