Web technology footprints

What the largest technology footprints show

The largest recorded query is jquery.min.js, with 34,960,796 homepage references. Google Fonts follows at 30,700,679, then WordPress at 24,899,751. Bootstrap and Font Awesome complete the top five. These queries describe several layers of a website: a JavaScript library, a font service, a content platform and two presentation libraries. A single homepage could contain all five. Their position in the table tells you which of these particular strings appears most often in the export.

The jQuery filename accounts for 27.36% of the documented snapshot. That figure answers a narrow question: how often did this source query appear against the export’s denominator? A differently named jQuery file would need another query. Likewise, the Bootstrap result covers bootstrap.min.css, rather than every Bootstrap build. Keeping the detection string beside the number makes the comparison usable: readers can decide whether the measured reference fits their own question.

Google Fonts has 30.7M recorded references to fonts.googleapis.com. Google’s Fonts API guide documents adding fonts through a stylesheet link. This query is useful for finding that external dependency. A site serving font files from its own domain would need different evidence. For a font-hosting review, inspect the surrounding stylesheet URL and the fonts requested before deciding which assets to move.

WordPress, WooCommerce and Shopify

The WordPress query looks for wp-content, while WooCommerce uses the more specific wp-content/plugins/woocommerce path. Their recorded totals are 24,899,751 and 5,031,078. The longer path gives a researcher a more specific lead to inspect: an asset associated with the WooCommerce plugin. The shorter path is useful when the initial task is simply finding WordPress content assets across the index.

To investigate the overlap directly, run "wp-content" "wp-content/plugins/woocommerce". Both strings must occur in the same stored homepage. The archive does not contain that combined query’s result, so its intersection should be measured separately. Dividing the two existing rows would produce a ratio of query totals, rather than an observed percentage of WordPress sites using WooCommerce.

Shopify’s 1,975,674 recorded references use cdn.shopify.com. Start with that asset host when looking for storefronts to review, then check the returned source and the current site. For an integration-specific search, combine it with "cdn.shopify.com" "static.klaviyo.com". That is a new research query; this report does not supply a historical count for the combination. The Shopify and Klaviyo guide walks through inspecting those matches.

Analytics and marketing integrations

The analytics rows answer different implementation questions. gtag/js?id=G- identifies a direct Google tag loader with the selected ID prefix. googletagmanager.com/gtm.js identifies a Tag Manager loader. Google’s Tag Manager documentation explains that a container can deploy Google and third-party tags. Finding its loader therefore gives you a place to investigate, rather than a list of every service running behind it.

The table also includes Meta Pixel, HubSpot, Klaviyo and Cloudflare Insights. Use their complete paths when researching an integration, rather than searching a vendor’s name alone. A name may appear in a testimonial, integration directory or blog article. An asset host or script filename narrows the search to source containing that particular implementation reference, which can then be checked in context.

The Cloudflare row is specifically the cloudflareinsights.com reference. Cloudflare documents its Web Analytics beacon separately from whether a site is proxied through Cloudflare. Label an exported list accordingly. “Homepages referencing the analytics beacon” preserves the evidence behind the list; “all Cloudflare customers” would describe a population this query never measured.

Why framework counts need different interpretation

React has 952,016 recorded matches for react-dom, while the Next.js row has 600,211 for __NEXT_DATA__. These strings are visible clues in HTML. The results depend on whether a site’s generated source exposes the selected clue. If an application ships its JavaScript inside a differently named bundle, searching its homepage for a library name can miss the implementation entirely.

A Next.js marker and a React reference can also occur on the same page. Counting those rows together would mix overlapping populations and different detection rules. When researching a framework, start with its technology page, read the exact fingerprint, and inspect several sources. If you need to broaden coverage, document each additional query separately before combining and deduplicating the resulting hostnames.

The separate jQuery filename report shows why query choice matters: it measures unversioned files, versioned files and CDN paths independently. Its numbers come from another documented snapshot and a different measurement process. Use that report to understand filename coverage, while keeping its sample estimates separate from the unclassified totals in this export.

Turn a technology query into a useful research list

Begin with a question specific enough to check manually. An agency reviewing WordPress commerce sites might start with the WooCommerce asset path. A developer examining payment integrations might choose js.stripe.com/v3. A font migration project might begin with the Google Fonts stylesheet host. In each case, the query defines a source reference you can point to when explaining why a hostname was included.

Open a small initial batch of returned sources before exporting a larger list. Record the matched URL, its surrounding HTML and whether the match is a script, stylesheet, comment or ordinary text. If several results come from documentation or examples, refine the query and check another batch. This review helps establish whether the string is useful for your project before you spend time enriching the domains.

Use combined strings to narrow a research question. For example, "wp-content/plugins/woocommerce" "js.stripe.com/v3" finds stored homepages containing both references. It identifies candidates for a WooCommerce and Stripe integration review. It does not establish that Stripe handles those sites’ checkout transactions; answering that would require examining the relevant implementation and checkout pages.

A top-level-domain filter can make a review manageable. Adding site:org limits the query to that suffix. The suffix describes a hostname group, rather than an organization’s location or legal status. Read the search syntax reference before combining filters, and save the complete query so another researcher can repeat the same selection.

Choose service queries that fit the task

The smaller service rows can be useful precisely because they identify a specific integration surface. Stripe’s 487,552 recorded references concern js.stripe.com/v3. That is a starting point for reviewing where Stripe.js is included. It does not establish checkout volume, subscription activity or the payment methods a merchant accepts. Those questions need information beyond the saved homepage.

Similarly, the reCAPTCHA API path can help locate pages to inspect during a form dependency review. Check whether the matched script belongs to the homepage’s form or another component, then inspect the actual form flow. The presence of a loader does not tell you which user actions trigger it. Keep the concrete asset reference attached to the task instead of assigning the entire hostname a broad classification.

For a campaign integration inventory, start with the HubSpot or Klaviyo asset host and inspect the returned snippets. If the goal concerns a particular account, find the account-specific value in source and run a narrower query. That produces a more focused investigation than treating every site that references a provider’s common script as related.

Compare snapshots without changing the question

For a later comparison, preserve the detection string, query mode, index date, denominator and count label together. If you replace a narrow filename with a broader path between measurements, the difference could reflect the new query rather than technology adoption. When the engine reports estimates, retain that label beside the values. A spreadsheet containing numbers alone loses information needed to interpret the change.

The CSV provides a starting record for this export: each row includes its source string, recorded total, calculated share, snapshot month and measurement label. Keep that original file alongside any filtered or enriched version. Add review notes in new columns, including the date you inspected a result. This keeps the published archive distinguishable from the current evidence collected for your own project.