CrUX History API or BigQuery? Pick the Right Core Web Vitals Monitoring Surface
CrUX History API or BigQuery for Core Web Vitals monitoring? See which surface fits your monitoring question before you build reports on the wrong one.

Your team wants to watch Core Web Vitals over time, and two Google surfaces keep coming up: the CrUX History API and CrUX on BigQuery.
They draw from the same Chrome user experience dataset, but they answer different questions.
Pick the wrong one and your monitoring reports will quietly miss what you actually need to see.
The decision comes down to three documented differences: resolution, update cadence and metric detail.
This article walks through each one, adds the eligibility check you should run first, and shows how the choice plays out in a concrete situation.
What each surface actually is
The CrUX History API provides programmatic access to CrUX historical data by page or origin, filterable by form factor and metrics.
It serves Web Vitals metrics at both origin and page level, and it updates weekly.
The values come from the past 40 weekly collection periods, each covering 28 days as a rolling window.
CrUX on BigQuery is a publicly accessible database of origin-level data.
You can query any origin for which data is collected, analyze any metric CrUX supports and filter by all available dimensions.
Full metric histograms are stored in the tables, and the data goes back to 2017.
It updates monthly, with each month's release landing on the second Tuesday after the collection period.
Both surfaces draw on the same underlying dataset, so the choice is not about data quality.
It is about which shape of the data matches your monitoring question.
Decision one: do you need page-level trends?
This is the sharpest dividing line.
The History API serves Web Vitals metrics both by origin and at page level.
BigQuery tables contain origin-level data only; page-level data is not available in BigQuery.
If your monitoring question is about a specific landing page, a template or a section of the site, BigQuery cannot answer it.
As an illustrative consequence of origin-level aggregation, a regression on one important page could hide inside a stable origin average.
If your question is about the whole site, or about comparing across dimensions, BigQuery is the stronger fit.
You can compare origins by country, form factor and collection month where those tables are available.
The tools documentation notes that the History API has no country dimension, while BigQuery supports all available dimensions.
Decision two: how fresh does the data need to be?
The History API updates weekly.
Historical data is released every Monday and contains the 40 most recent 28-day collection periods ending on Saturdays.
BigQuery updates monthly, released on the second Tuesday after each monthly collection period, covering the last 28 days of that period.
Map your review rhythm to these schedules.
If your team reviews performance trends weekly and wants to catch a shift within a few weeks, the History API's Monday releases fit that rhythm.
If your monitoring is a monthly or quarterly review, BigQuery's monthly cadence is sufficient, and you gain the dimensional analysis in exchange.
Neither surface gives you real-time data.
Both report aggregated user experiences on their respective release schedules, so neither can verify a fix immediately after deployment.
Decision three: which metrics and how much detail?
The History API serves a subset of key metrics.
The web vital metrics are available in all CrUX tools, so LCP, INP and CLS trends are covered either way.
BigQuery stores all metrics except LCP resource types and subparts, with full metric histograms.
That matters if you want to visualize the full distribution of performance rather than a summary value.
One caveat from the documentation: percentiles in BigQuery are interpreted from coarse histogram data, which results in approximate values.
If your reporting depends on precise percentile figures, verify how much that approximation matters for your thresholds.
BigQuery also holds data since 2017, which supports long-run historical analysis.
The History API covers the previous 40 weeks.
If your question involves comparing today against conditions from years ago, BigQuery is the only one of the two that can do it.
Check eligibility before you build
Before committing to either surface, confirm the pages or origin you want to monitor actually appear in CrUX.
Pages and origins must be publicly discoverable and sufficiently popular to be included.
Discoverability follows the same indexability criteria as search engines.
A page that is non-200 after redirects, blocked by noindex, or otherwise not indexable will not qualify.
The popularity threshold uses a minimum number of visitors whose exact value Google does not disclose.
There is one nuance worth knowing.
Page-level data in PageSpeed Insights and the CrUX API requires the user, origin and page criteria to all be satisfied.
Experiences meeting the user and origin criteria but not the page criteria aren't included in origin-level CrUX data, per the methodology documentation.
If you plan page-level monitoring on the History API, verify that the specific pages you care about actually have data before building reports around them.
Also note that pages and origins are automatically added or removed as eligibility changes, and you cannot manually submit a page or origin for inclusion.
A page that drops out of the dataset will simply stop appearing in your monitoring, so build that possibility into how you interpret gaps.
A worked example, clearly illustrative
Suppose a marketing team wants to know whether a specific landing page's LCP trend is improving after a template change.
Mapping that question: it needs page-level resolution, which points to the History API, and weekly cadence fits their review rhythm.
The team would first confirm the page has page-level CrUX data at all, since eligibility is not guaranteed.
Now suppose a different team wants to compare origin-level performance distributions across countries and form factors over several years.
That maps to BigQuery: origin-level only, all dimensions, full histograms and data since 2017.
The monthly cadence suits a quarterly review.
These are illustrative mappings, not measured results.
The point is that the question itself determines the surface, and you can run the same mapping against your own monitoring requirements.
A short decision checklist
- Does your question concern specific pages? If yes, the History API is the only one of the two that serves page-level data. If no, continue.
- Do you need to compare by country or form factor? If yes, BigQuery supports all available dimensions; the History API has no country dimension. If no, continue.
- Do you need full metric histograms or history back to 2017? If yes, BigQuery stores them, with approximate percentiles from coarse histograms. If no, the History API's subset of key metrics over 40 weeks may be enough.
- Does your review rhythm need weekly updates? If yes, the History API releases every Monday. If monthly works, BigQuery's second-Tuesday release fits.
- Have you verified the pages or origin meet CrUX eligibility? If not, check discoverability and popularity before building anything on either surface.
These questions can point in different directions, which is expected.
A team needing page-level trends and dimensional analysis may end up using both surfaces for different questions rather than forcing one to do a job it was not built for.
Where this fits in your monitoring workflow
Whichever surface you choose, the output still has to land somewhere your team acts on it.
If you are already consolidating reporting across tools, it may help to think about how CrUX data sits alongside your other monitoring surfaces.
For a related decision about consolidating SEO reporting, see Semrush SEO Dashboard or Individual Reports? Decide Before You Commit.
If your monitoring extends to content performance, Blog Analyze Tab or Content Performance Tab? Pick the Right First Check covers a similar first-check decision.
And if part of your monitoring involves user-facing automation, When Should a Chatbot Appear on a Page? Pick the Right Trigger addresses the page-level trigger question.
The core takeaway: match the surface to the question.
Page-level trends and weekly freshness point to the History API.
Origin-level analysis across dimensions, full histograms and long history point to BigQuery.
Verify eligibility first so you do not build reports around pages that have no CrUX data today, and expect gaps if eligibility changes later.
Source references: developer.chrome.com; developer.chrome.com; developer.chrome.com.
How Meshline can help. Connect automation, Organic Marketing (demand generation), and customer lifecycle management (Revenue Intelligence).
Bring topic planning, content publishing and performance feedback into the conversation about your workflow. Book a Meshline demo.