Monthly Trend for Core Web Vitals

[vivek shekhar singh]

How can i create a monthly trend for Core Web Vitals? I want to pull historical data from previous months , if i am to use Big Query for that  what should be my SQL query to pull data from last 5 months , if so in almost how many attempts i will exhaust the monthly limit of 1 TB. Is there any plan to add these features in JS CRUX API like by country , by month etc ?

[Rick Viscomi]

The easiest way to get historical trends is to create a CrUX Dashboard. I wrote up a how-to guide at https://web.dev/chrome-ux-report-data-studio-dashboard/. No SQL needed and it gives you trends for all Core Web Vitals metrics as well as secondary metrics like FCP, TTFB, etc. It's only available at the origin-level (per-website) and not for individual pages. I also wrote a little launcher app to jump you straight into a customized dashboard without all the setup.

If you'd prefer to query the origin-level data yourself, you could write a query similar to this one, which I presented in the Mastering CrUX on BigQuery talk last year. Here's a related query that processes CrUX data by country. You could also query the raw histogram data over time if you wanted to do any special processing like non-standard good/poor thresholds or percentiles, but this query will exceed the free 1 TB quota. The `materialized` dataset is much smaller and cheaper to query, similar to the examples in the video linked above, so that's the recommended approach if you did want to extract the data manually.

If you need URL-level data, that would only be possible using the CrUX API, however note that it doesn't support historical data. So you would need to set up a monitoring process similar to the competitive monitoring guide I shared in the other thread, using Apps Script to monitor the CrUX API on a daily basis. Historical data would only be available starting at the point you begin monitoring. Also note that the CrUX API does not currently support country-level data either. If URL-level country and historical data would be useful to you, let me know and I'll make sure to forward it along to the product team. (For anyone else reading, feel free to jump in and add your +1 if you'd also find it useful — that will help send a stronger signal that developers would find the features useful!)

Rick

[vivek shekhar singh]

Thanks Rick . This is super helpful . I see one caveat with the CRUX Dashboard that it shows data which is 1 month behind . For eg i am in August but it shows date till June . Net net can you tell me the materialized dataset ,how many calls ill be able to make in a month considering 1 TB limit.  Yes it would be super helpful if we can add historical data and country-level data to CRUX API 

[Rick Viscomi]

I see one caveat with the CRUX Dashboard that it shows data which is 1 month behind . For eg i am in August but it shows date till June .

Monthly datasets are released on the second Tuesday of each month, so the July dataset will be available next week.

Net net can you tell me the materialized dataset ,how many calls ill be able to make in a month considering 1 TB limit.

BigQuery gives you an estimate of the number of GB each query will process. For example the query I shared earlier will process 12 GB, so you can run it 80 times in a month without going over the free quota.

[vivek shekhar singh]

Can i access 1 TB limit forever once i have an account on Google Cloud Platform or is it limited to only 3 months ? Is there anyway i can access 1 TB of data at least without paying 

[Rick Viscomi]

Per https://cloud.google.com/bigquery/pricing#free-tier the 1 TB free quota resets every month.

[Shawn O'Neill]

"If URL-level country and historical data would be useful to you, let me know and I'll make sure to forward it along to the product team." 

Exactly this, thanks Rick!

[Angie Biron Ren]

"If URL-level country and historical data would be useful to you, let me know and I'll make sure to forward it along to the product team." 

Hi Rick, +100 from me - this would be enormously useful! Currently I'm using "page load time" in GA to get a rough estimate of differences by country, and I see big regional differences but it's not as reliable as the real CrUX data. I'm hesitant to use the web-vitals.js too, for one thing because I'd be adding another resource and for another, I've heard there are many limitations and difficulties to creating a useful report from this. 

So definitely would love url-level and country-level (and url- plus country- level!) data native in CrUX. :)))

[Rick Viscomi]

Thanks Shawn and Angie for the feedback!

[Aymen Loukil]

+100 that's would be useful for monitoring