Why do small websites have no CrUX data, and what should a plan do?
Part of Fair publisher planning
The short answer
Small websites can have no Chrome UX Report record because Google includes only publicly discoverable origins that pass a popularity threshold it does not disclose, and only some Chrome users contribute. A missing record says the site did not qualify for the dataset. It says nothing about whether real readers exist; a plan should label it missing, never zero.
Key takeaways
- CrUX leaves out origins below an undisclosed popularity threshold, and country datasets apply that threshold again inside each country.
- Visits from Chrome on iOS, Android apps using WebView and other Chromium browsers do not enter the data, so some readerships are partly invisible to it.
- A missing record is a fact about the dataset, so a plan should give it its own label instead of a blank cell or a zero.
- Small sites can be supported with other evidence: sampled articles, publishing history, the selling path in ads.txt and sellers.json, and aggregated figures from the publisher.
- The fair rule is to rank by fit and quality first, then show popularity evidence, or its absence, beside each site.
Why do some websites have no CrUX data?
The Chrome UX Report is built from real Chrome visits, but not every site makes it in. The CrUX methodology (opens in a new tab) sets two conditions for every origin and page: it must be publicly discoverable and sufficiently popular. On the second condition Google writes that "An exact number is not disclosed", and that "The minimum number is the same for pages and origins." Pages and origins that miss the threshold "are not included in the CrUX dataset."
Discoverability follows search indexing rules. A page served with a status code other than 200 after redirects, or marked noindex in a header or a robots meta tag, cannot qualify. A third rule applies to whole origins: origins that have more than 20 percent of their total traffic excluded because of ineligible combinations of dimensions are excluded entirely from the dataset.
| Reason in Google's methodology | What it means for a publisher | What it does not mean |
|---|---|---|
| Below the undisclosed popularity threshold | Too few eligible Chrome visits in the period to be published | That the site has no readers, or fewer readers than ranked sites |
| Not publicly discoverable (non 200 status, noindex) | The address checked may redirect, error or be set to noindex | That the publisher's articles are hidden from readers |
| More than 20 percent of traffic in ineligible dimension combinations | The whole origin is dropped, not only the ineligible part | That the site's traffic is invalid or suspicious |
| Readers on browsers that do not contribute | Visits from Chrome on iOS, Android WebView and other Chromium browsers are not counted | That those readers do not exist or do not see ads |
| Wrong origin queried | The lookup used the wrong host or protocol for the site | Anything about the site at all, until the lookup is fixed |
Does missing CrUX data mean a site has no audience?
No. A missing record means the origin did not qualify for a dataset with fixed rules about who contributes. Only users who have usage statistic reporting turned on, sync their browser history and have no sync passphrase are eligible, on desktop Chrome or Chrome for Android. Chrome on iOS, Android apps using WebView and other Chromium browsers such as Microsoft Edge are listed as excluded. A site's readers can be real, loyal and in the right market while some or all of their visits never reach CrUX.
Country data narrows the picture further. The CrUX dimensions page (opens in a new tab) says country values are inferred from users' IP addresses and that country level datasets apply "the standard eligibility requirements" at a country level. A publisher can therefore clear the threshold worldwide and still have no record in the Canadian or US dataset, because its visits from that one country fall short. The same page notes that country dimensions are not available in the CrUX APIs, only in BigQuery.
Definition
- Missing, not zero
- The rule that an origin absent from a dataset is recorded as having no measurement, never as having a measurement of zero. Zero is a value; missing is the absence of one, and a plan must keep the two apart.
Which kinds of publishers are exposed to missing CrUX data?
Google does not publish a breakdown of which sites fall below the threshold, so no one can give a sourced share by publisher type. A planner can still read the documented rules and see who is exposed to each one. A publisher exposed to several rules at once is more exposed to coming back with no record, whatever the quality of its journalism or the loyalty of its readers.
| Publisher situation | Documented rule it meets | Check before concluding anything |
|---|---|---|
| Local, niche or new publisher with a modest readership | Undisclosed popularity threshold, applied again per country | Look for the origin in the global dataset as well as the country dataset |
| Readers who arrive mostly from links inside Android apps | Android apps using WebView do not contribute; Custom Tabs do | Ask the publisher how readers reach it, in aggregate |
| Readership weighted toward iPhones and iPads | Chrome on iOS does not contribute | Treat the missing record as uninformative, not as low reach |
| Articles on a different host from the homepage | An origin is one scheme and host, such as https://www.example.com | Query the host that serves the articles, not only the brand domain |
| Site that recently moved hosts or redirects its old address | Status codes other than 200 after redirects fail discoverability | Query the final address the reader lands on |
The pattern matters for fairness. Popularity thresholds and browser coverage gaps can fall on the independent, local and specialist publishers that the fair publisher planning hub argues deserve the same review as large media groups. A plan that reads absence as weakness turns a dataset limit into a bias against small publishers.
How should a plan show a site without popularity data?
Give missing data its own visible state. A blank cell invites a reader to assume the worst, and a zero is simply false. One workable convention is a short set of evidence states used on each row of the plan, written so that a client who never reads the methodology still understands what each one means.
| State | When to use it | Wording the client sees |
|---|---|---|
| Ranked in market | The origin has a rank in the brief's country dataset | Popularity band in Canada (or the US), with the month of data |
| Ranked elsewhere only | The origin appears worldwide but not in the brief's country dataset | Ranked worldwide; no country record this month |
| Not in the dataset | A correct lookup of the right host found no record | No public popularity data; not a measure of audience |
| Lookup not confirmed | The host, protocol or redirect target is still uncertain | Popularity not yet checked |
| Publisher evidence supplied | The publisher shared aggregated figures with a period and geography | Publisher reported, with its period and source named |
Four rules for presenting missing popularity data
- Sort by fit before popularity. Order the plan by relevance and quality, so a missing band never pushes a strong site down the list on its own.
- Never fill the gap. Do not estimate a rank for a missing origin or guess where the hidden threshold sits; write the state instead.
- Date every reading. CrUX data on BigQuery is released on the second Tuesday of the following month, so name the month each band comes from.
- Keep sources apart. Show CrUX bands and publisher supplied figures in separate columns, because they measure different things over different periods.
What other evidence can support a small site?
Popularity is one question among several. For a small site the other questions can carry more weight, and several of them can be answered from public material. The evidence below supports a decision to include, test or leave out a site without leaning on a dataset the site may never enter.
Evidence a small site can be judged on without a CrUX record
- Several recent articles were read and match the brief's topics, not just the homepage.
- The publishing history shows regular new articles with named authors or a named editorial team.
- The site's ads.txt file (opens in a new tab) lists at least one seller, and the matching exchange sellers.json entry agrees.
- The page layout leaves the article readable, without heavy or rapidly refreshing ad slots.
- The publisher's language and regional focus fit the brief's market.
- Any audience figures the publisher shares are aggregated, dated and tied to a named geography.
- The reason for including the site is written beside it, so the client can weigh it.
Publisher supplied figures deserve care. Ask for totals over a stated period and country, never for data about individual readers. The niche publisher discovery guide covers how to find specialist sites in the first place, and the CrUX popularity rank glossary explains what each rank band means when a record does exist.
How does PlacementGraph handle sites without popularity data?
What does a missing data review change on a real plan?
Fictional example
A fictional independent bookstore group in Vermont and New Hampshire
A fictional group of independent bookstores in Vermont and New Hampshire briefs a display plan beside book reviews, author interviews and regional arts coverage, in English. The first draft keeps only sites with a US popularity band and leaves out every row with a blank cell.
The planner reviews the blanks before accepting the cut. Two literary magazines had been looked up under their brand domain while their articles live on a separate host; queried on the right host, one has a US band. A small regional arts site appears worldwide but has no US record that month. A volunteer run book review site has no record anywhere.
The final plan sorts all of them by fit, labels each popularity cell with one of the five evidence states, and adds a column for the publisher supplied figures one magazine shared. The client sees which sites are measured, which are not, and why each one is there.
Frequently asked questions
Can a site with no CrUX record still have a strong audience in one city?
Yes. Country datasets apply the threshold to visits from that country, and there is no city level dataset at all. A site read closely in one city or region can fall short of the national threshold while serving that place well. Judge local reach from the content, the publisher's own aggregated figures and the brief's geography.
Will a small site appear in CrUX once it grows?
It may, if its eligible Chrome visits pass the undisclosed threshold and its pages stay publicly discoverable. Because Google does not publish the number, no one can say when that will happen. Check again when a new monthly release is out, and record the month of each check rather than assuming the site's status.
Should a planner ask Google for a site's CrUX numbers?
There is nothing more to ask for. CrUX is a public dataset, and origins below the threshold are not included, so no fuller version exists for an individual buyer. The useful questions go to the publisher instead: aggregated totals for a stated period and country, and how readers reach the site.
Does the CrUX API show country data for a site?
No. Google's dimensions page states that country dimensions are not available in the CrUX APIs. Country records come from the BigQuery datasets, released monthly. An API lookup can confirm whether an origin is in the global data, but it cannot answer whether the site has a record for Canada or the United States.
Sources
- CrUX methodology (opens in a new tab)Google Chrome for Developers
- CrUX dimensions (opens in a new tab)Google Chrome for Developers
- Using the Chrome UX Report API (opens in a new tab)Google Chrome for Developers
- CrUX on BigQuery (opens in a new tab)Google Chrome for Developers
- ads.txt (Authorized Digital Sellers) (opens in a new tab)IAB Tech Lab
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.