Inquiry on Replacing ECT with RTT for Data Filtering

[JeongHyun Yoon]

Hi CrUX Team,

I have been using ect (Effective Connection Type) as a filtering condition in my queries, but with the recent addition of round_trip_time (RTT) data in the CrUX dataset, I am considering switching to RTT-based filtering instead.

Currently, my query looks like this:

SELECT 

largest_contentful_paint 

  FROM `chrome-ux-report.all.202412` 

  WHERE origin = 'https://www.google.com' 

  AND form_factor.name = 'phone' 

  AND ect.name = '4G'

I would like to modify this query to filter based on round_trip_time instead of ect. Specifically, I want to retrieve largest_contentful_paint values only when the RTT is 270ms or lower.

Could you advise on the best way to achieve this? Additionally, are there any notable differences or limitations in using RTT compared to ECT for filtering?

thanks

[Barry Pollard]

It is not possible to filter data on RTT like it was done with ECT. This was a conscious decision made to increase coverage.

This was covered in the launch blog post: https://developer.chrome.com/blog/crux-2025-02?hl=en#removal_of_the_ect_dimension

An important difference is that ECT was a CrUX dimension—meaning the other metrics could be segmented by ECT. RTT is a CrUX metric instead of a dimension, so it's not possible to view LCP by RTT for example, but only to see the RTTs by the other dimensions (device type and country).

 

This may sound more limiting, but the move from dimension to metric actually unlocks more data in CrUX. This is because CrUX has certain minimum thresholds before we are able to show data. We already made dimensions optional in 2022 meaning we removed ECT, or device where necessary to report at a higher level, but metrics that were not on most page loads (Interaction to Next Paint (INP), different navigation types, and now LCP image subparts) were frequently not available for origins in BigQuery.

 

By reducing the number of dimensions, the data is less segmented, so the number of origins meeting these minimum requirements is increased. In January, we report INP for 68.1% of origins, whereas for the December dataset, we reported INP only for 64.5% of origins. The mechanism applies also to Navigation Types, the LCP Subparts and Resource Types in this release—they all benefit from the removal of the ECT dimension. In the CrUX APIs, the increased coverage has taken effect from the beginning of February.

To split user data by RTT (or ECT), you must collect this data yourself through a RUM solution.

--
You received this message because you are subscribed to the Google Groups "Chrome UX Report (Discussions)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chrome-ux-repo...@chromium.org.
To view this discussion visit https://groups.google.com/a/chromium.org/d/msgid/chrome-ux-report/d08429aa-7e42-4391-b12a-2d0531b6f222n%40chromium.org.

[JeongHyun Yoon]

Thank you for your prompt response and clarification. I appreciate your time and effort in addressing my inquiry.

Best regards,

2025년 3월 3일 월요일 오후 11시 19분 9초 UTC+9에 barryp...@google.com님이 작성: