Intent to Extend Origin Trial: Conversion Measurement API (Attribution Reporting API)

[John Delaney]

Contact emails

john...@chromium.org, cshar...@chromium.org

Explainer

https://github.com/WICG/conversion-measurement-api/blob/main/event_attribution_reporting_clicks.md

Design docs/spec

https://wicg.github.io/conversion-measurement-api/

Summary

This is a new API for measuring conversions (e.g. purchases) and attributing them to clicked ads, without using cross-site persistent identifiers like third party cookies.

Link to “Intent to Prototype” blink-dev discussion

https://groups.google.com/a/chromium.org/d/msg/blink-dev/7B0ldtZR_68/GjLBu0n4DgAJ

 

Link to “Intent to Experiment” blink-dev discussion

https://groups.google.com/a/chromium.org/d/msg/blink-dev/Ib9-tDFitns/Sm9RXTNaBwAJ

Goals for experimentation

For a continuation of the trial, we still hope to see that the measurement data made available through the API produces comparable results to existing click through measurement solutions on the web.

We would like to see that Chrome 94 addresses the reporting issue documented here:

https://groups.google.com/a/chromium.org/g/attribution-reporting-api-dev/c/2KKqWG70w7U/m/7M8fWEbmAgAJ

Experimental timeline

We'd like to extend the origin trial for one additional milestone, continuing through M94. The experiment has been running since M86.

Any risks when the experiment finishes?

This API is only additive and does not affect any existing state. Current conversion measurement solutions will work as expected with and without the API, so we don’t believe there are any risks.

All web-visible names of the API were modified in Chrome 92 which mitigated burn in risk as developers needed to update their implementation to continue experimenting.

We have previously requested an exception to the .5% page load usage limit for Origin Trials. Page load usage is currently around .92% for the feature.

https://chromestatus.com/metrics/feature/timeline/popularity/3365

Reason this experiment is being extended

There are a number of new internal partners experimenting with the API. Chrome 93 has a shorter than normal Stable lifetime, with Stable starting on Aug 31st, and 93 OT tokens expiring on September 14th, only giving two weeks to experiment. We would like to allow a bit longer for experimentation.

We recently received some great feedback from a partner touching on data loss due to network errors: https://github.com/WICG/conversion-measurement-api/issues/201. Chrome 94 includes a fix for this issue, which we would like to verify works.

Ongoing technical constraints

None.

Will this feature be supported on all five Blink platforms supported by Origin Trials (Windows, Mac, Linux, Chrome OS, and Android)?

Yes.

Link to entry on the feature dashboard

https://chromestatus.com/feature/6412002824028160

[Yoav Weiss]

On Sat, Sep 11, 2021 at 12:01 AM John Delaney <john...@chromium.org> wrote:

Contact emails

john...@chromium.org, cshar...@chromium.org

Explainer

https://github.com/WICG/conversion-measurement-api/blob/main/event_attribution_reporting_clicks.md

Design docs/spec

https://wicg.github.io/conversion-measurement-api/

Summary

This is a new API for measuring conversions (e.g. purchases) and attributing them to clicked ads, without using cross-site persistent identifiers like third party cookies.

Link to “Intent to Prototype” blink-dev discussion

https://groups.google.com/a/chromium.org/d/msg/blink-dev/7B0ldtZR_68/GjLBu0n4DgAJ

 

Link to “Intent to Experiment” blink-dev discussion

https://groups.google.com/a/chromium.org/d/msg/blink-dev/Ib9-tDFitns/Sm9RXTNaBwAJ

Can you link to previous extension requests?

 

Goals for experimentation

For a continuation of the trial, we still hope to see that the measurement data made available through the API produces comparable results to existing click through measurement solutions on the web.

We would like to see that Chrome 94 addresses the reporting issue documented here:

https://groups.google.com/a/chromium.org/g/attribution-reporting-api-dev/c/2KKqWG70w7U/m/7M8fWEbmAgAJ

That seems reasonable.

 

Experimental timeline

We'd like to extend the origin trial for one additional milestone, continuing through M94. The experiment has been running since M86.

 

M86 went to stable almost a year ago. What's the plan beyond M94? Assuming that the bug is fixed, will the experiment end there? Would you be able to share learnings from the trial? Are you planning to ship the API in its current form?

Any risks when the experiment finishes?

This API is only additive and does not affect any existing state. Current conversion measurement solutions will work as expected with and without the API, so we don’t believe there are any risks.

All web-visible names of the API were modified in Chrome 92 which mitigated burn in risk as developers needed to update their implementation to continue experimenting.

We have previously requested an exception to the .5% page load usage limit for Origin Trials. Page load usage is currently around .92% for the feature.

https://chromestatus.com/metrics/feature/timeline/popularity/3365

Reason this experiment is being extended

There are a number of new internal partners experimenting with the API. Chrome 93 has a shorter than normal Stable lifetime, with Stable starting on Aug 31st, and 93 OT tokens expiring on September 14th, only giving two weeks to experiment. We would like to allow a bit longer for experimentation.

We recently received some great feedback from a partner touching on data loss due to network errors: https://github.com/WICG/conversion-measurement-api/issues/201. Chrome 94 includes a fix for this issue, which we would like to verify works.

Ongoing technical constraints

None.

Will this feature be supported on all five Blink platforms supported by Origin Trials (Windows, Mac, Linux, Chrome OS, and Android)?

Yes.

Link to entry on the feature dashboard

https://chromestatus.com/feature/6412002824028160

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/b863ba43-5b38-40d2-95d2-12f64a10fca7n%40chromium.org.

[John Delaney]

Can you link to previous extension requests?

First extension: https://groups.google.com/a/chromium.org/g/blink-dev/c/C0P7ePjITJQ/m/ZYFmn30SDQAJ 

Second extension: https://groups.google.com/a/chromium.org/g/blink-dev/c/xCWP1ltlAgw/m/acwN0GIRBQAJ

M86 went to stable almost a year ago. What's the plan beyond M94? Assuming that the bug is fixed, will the experiment end there? Would you be able to share learnings from the trial? Are you planning to ship the API in its current form?

We intend to end the experiment in 95. We are currently working on a number of significant extensions/changes to the API including: aggregate reports [1], support for views [2], and others. These support use-cases which are not covered in the currently implemented "Event-level Click" API. We plan to start a "V2" experiment when this new functionality is available which we believe will produce new, different insights from developers. We are not currently planning to ship Attribution Reporting without first experimenting with those extensions.

We can certainly share learnings with respect to the "Event-level Click" API when this initial experiment concludes. 

[1] https://groups.google.com/a/chromium.org/g/blink-dev/c/2zA5-TuVSkA/m/GVqUUkhfAwAJ

[2] https://github.com/WICG/conversion-measurement-api/blob/main/event_attribution_reporting_views.md

[Alex Russell]

Discussed at today's API OWNERS, the relatively high use plus the length of the ongoing trial adds a level of concern. Given that this is an x-origin OT, relatively high use might not be the whole picture, so it would be great to know in addition how many individual teams have signed up to use the OT. Is dozens? Hundreds? Thousands?

Thanks.

Best Regards,

Alex

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

[John Delaney]

Hi Alex,

Given that this is an x-origin OT, relatively high use might not be the whole picture, so it would be great to know in addition how many individual teams have signed up to use the OT. Is dozens? Hundreds? Thousands?

There are O(10) teams experimenting with the API.

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

[Yoav Weiss]

O(10) sounds manageable in case we want to change course, so that reduces the burn-in risk.

Since this is an atypical OT extension, approving it would require 3 LGTMs.

LGTM1 to extend experimentation till M95.

[Chris Harrelson]

LGTM2

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAL5BFfVMWo2TOFBnQpTp4yLcEpLRhCwu0iyJH8dqS66-__RsmA%40mail.gmail.com.

[Alex Russell]

LGTM3 to extend to 95

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/b863ba43-5b38-40d2-95d2-12f64a10fca7n%40chromium.org.

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.