You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current explainer for the event-level API discusses 3 bits of conversion data, with 5% noise applied — that is, with 5% chance, we send a random 3 bits, and the other 95% of the time we send the real conversion data. This is pretty inflexible.
Instead, we could consider a smooth trade-off between the number of unique ids on the conversion data(3 bits has 8 ids) and the amount of noise(here uniform 5%). Ad-tech can have the flexibility to choose the maximum number of ids and browsers would add noise accordingly to maintain similar levels of privacy.
The text was updated successfully, but these errors were encountered:
It would be great for us to have this even if the noise ratio is increased since the current 3 bits are too small for us to use for the ad measurements.
We need at least 6 bits for the conversion data.
The current explainer for the event-level API discusses 3 bits of conversion data, with 5% noise applied — that is, with 5% chance, we send a random 3 bits, and the other 95% of the time we send the real conversion data. This is pretty inflexible.
Instead, we could consider a smooth trade-off between the number of unique ids on the conversion data(3 bits has 8 ids) and the amount of noise(here uniform 5%). Ad-tech can have the flexibility to choose the maximum number of ids and browsers would add noise accordingly to maintain similar levels of privacy.
The text was updated successfully, but these errors were encountered: