-
Notifications
You must be signed in to change notification settings - Fork 76
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Picking label sizes for Chrome-facilitated testing #113
Comments
Before we get into the size of labels, can you give us an example of how the traffic will be labeled? Where in the oRTB spec will you note the relevant details of the experiment? |
A single digit grain should be fine. For consistency Mode A and Mode B should use the same scheme. It is important that Chrome label each Mode B tranche for some time prior to Chrome's cookie suppression so that everyone can capture before and after measurements. This cuts across Sandbox features. Will you at some point have a call(s) to discuss and align, or were you thinkling of covering this in say the biweekly PAAPI (nee FLEDGE) calls? |
@jdwieland8282 for the first part of your question: via #114 we're considering Client Hints as the mechanism, meaning a specific request header that specifies the mode / experiment for that browser. However, we'd like to get feedback on that issue if that would meet your needs, or if not where you'd expect it to appear. |
@dmdabbs good request! I think we should certainly plan some explicit office hours sessions and I'll share the comment with the various people running individual calls for their view on where it makes sense to integrate. |
To have some progressive ramp-up, the first two labels could be for 0.5% of the traffic each, then 1% per label. |
Wrt Mode A it would be good to start with something small like 0.1%, and start quick |
Closing - please refer to the Chrome-facilitated testing article for the size of label groups. |
Chrome’s testing modes will provide labeled traffic that can be used by ad techs to construct different experiment arms. We want to provide flexibility to ad techs to run multiple experimental conditions or to test at different levels of traffic. What size of labels would be most useful to you in each mode? 1%, .5%, .25%, or something else?
See https://developer.chrome.com/docs/privacy-sandbox/chrome-testing/ for context.
The text was updated successfully, but these errors were encountered: