Connect Google PageSpeed Insights
In the client’s workspace, open Sources, add Google PageSpeed Insights and choose the account to sync. If you don’t have access to the account, send the client a connect link where the platform supports one.
Analytics integration
Turn lighthouse performance, accessibility, best-practices, and SEO scores into 34 governed metrics for dashboards, agent answers, and cross-channel reporting.
What you can do with Google PageSpeed Insights in Maven
Maven models Google PageSpeed Insights into 34 governed metrics from the semantic layer. Track Downtime Seconds, Unmonitored Seconds, Uptime Percentage, Uptime Seconds and Web Accessibility Score, break results down by Signal Type, Account, Target and Strategy, and report Google PageSpeed Insights alongside Google Ads, Facebook Ads and Google Search Console without rebuilding exports or metric logic for every report.
What you can measure
Maven keeps Google PageSpeed Insights numbers consistent across dashboards, reports, and answers so the team is not debating which spreadsheet is right.
Total monitored seconds in a down state across selected checks and dates
Total seconds without monitoring coverage across selected checks and dates
Monitored uptime seconds divided by uptime plus downtime seconds as a percentage
Total monitored seconds in an up state across selected checks and dates
Average provider-native accessibility score across represented page checks
Average provider-native best-practices score across represented page checks
Cumulative Layout Shift score.
First Contentful Paint in milliseconds.
Largest Contentful Paint in milliseconds.
Numeric value reported by the provider. Interpret with metric_name and stream.
Average provider-native web performance score across represented page checks
Average provider-native SEO score across represented page checks
Speed Index in milliseconds.
Total Blocking Time in milliseconds.
Average active (interacting) time per recorded session in seconds (Clarity EngagementTime activeTime), weighted by recorded sessions. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Average total time per recorded session in seconds (Clarity EngagementTime totalTime), weighted by recorded sessions. Differs from avg_session_duration_seconds (web analytics sessions). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Average uptime-check response time in milliseconds. Unweighted mean of the provider's daily average per check (Pingdom and compatible uptime monitors); not request-weighted.
Average percent of the page scrolled per recorded session (Clarity ScrollDepth averageScrollDepth), weighted by recorded sessions. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Share of recorded sessions with at least one of the clicks on page elements that produced no response (0-1): sessions with the signal (Clarity DeadClickCount sessionsCount x sessionsWithMetricPercentage) divided by sessions evaluated, re-weighted across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Occurrences of clicks on page elements that produced no response in recorded sessions (Clarity DeadClickCount subTotal). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Share of recorded sessions with at least one of the clicks followed by a JavaScript error (0-1): sessions with the signal (Clarity ErrorClickCount sessionsCount x sessionsWithMetricPercentage) divided by sessions evaluated, re-weighted across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Occurrences of clicks followed by a JavaScript error in recorded sessions (Clarity ErrorClickCount subTotal). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Share of recorded sessions with at least one of the scrolling through a page more than expected (0-1): sessions with the signal (Clarity ExcessiveScroll sessionsCount x sessionsWithMetricPercentage) divided by sessions evaluated, re-weighted across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Occurrences of scrolling through a page more than expected in recorded sessions (Clarity ExcessiveScroll subTotal). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Share of recorded sessions with at least one of the navigating to a page and quickly returning to the previous one (0-1): sessions with the signal (Clarity QuickbackClick sessionsCount x sessionsWithMetricPercentage) divided by sessions evaluated, re-weighted across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Occurrences of navigating to a page and quickly returning to the previous one in recorded sessions (Clarity QuickbackClick subTotal). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Share of recorded sessions with at least one of the rapid repeated clicks in one spot (0-1): sessions with the signal (Clarity RageClickCount sessionsCount x sessionsWithMetricPercentage) divided by sessions evaluated, re-weighted across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Occurrences of rapid repeated clicks in one spot in recorded sessions (Clarity RageClickCount subTotal). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Recorded sessions classified as bots (Clarity Traffic totalBotSessionCount). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Distinct users across recorded sessions (Clarity Traffic distinctUserCount). Non-additive: a sum over dates or dimension values over-counts returning users. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Sessions recorded by the session-recording tool (Clarity Traffic totalSessionCount, bot sessions included). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Share of recorded sessions with at least one of the JavaScript errors (0-1): sessions with the signal (Clarity ScriptErrorCount sessionsCount x sessionsWithMetricPercentage) divided by sessions evaluated, re-weighted across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Occurrences of JavaScript errors in recorded sessions (Clarity ScriptErrorCount subTotal). Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Average pages per recorded session: page views behind Clarity's pagesPerSessionPercentage divided by recorded sessions, so it re-weights across dates and dimension values. Session-recording measure from the behavior rows of obt_web_ops (Microsoft Clarity live insights), not GA4/web-analytics sessions. Daily snapshot of the trailing 24 hours before the sync; group by device_category, browser or country (device/browser/country call) or page_location (URL call) for the breakdown rows.
Connection and sync
You connect Google PageSpeed Insights with API credentials from the Google PageSpeed Insights account. Maven stores them encrypted. Each Google PageSpeed Insights sync takes a fresh copy of the current records, so the workspace always reflects what is in Google PageSpeed Insights now.
Daily PageSpeed/Lighthouse web performance snapshot.
Fields: account name, target URL, strategy, performance score, accessibility score, best practices score, SEO score, first contentful paint ms, largest contentful paint ms, cumulative layout shift, and 2 more
How it works
No SQL or warehouse setup. Connect the source, then build the client’s report by asking Claude or ChatGPT.
In the client’s workspace, open Sources, add Google PageSpeed Insights and choose the account to sync. If you don’t have access to the account, send the client a connect link where the platform supports one.
Maven pulls the client’s Google PageSpeed Insights data (1 objects) and models it into documented tables. No SQL, warehouse or pipeline setup.
Build the client’s report or deck by asking Claude or ChatGPT through the MCP, set alerts on what matters, and publish to the client’s portal once you approve it.
Works well with
FAQ
Maven exposes 34 governed Google PageSpeed Insights metrics, including Downtime Seconds, Unmonitored Seconds, Uptime Percentage, Uptime Seconds, Web Accessibility Score and Web Best Practices Score — each defined once so it means the same thing in every report.
You can break Google PageSpeed Insights metrics down by Signal Type, Account, Target, Strategy, Performance Score, Accessibility Score, Best Practices Score and Seo Score, and more.
Yes. Google PageSpeed Insights sits in the same client workspace as sources like Google Ads, Facebook Ads, Google Search Console, Shopify and HubSpot, so one report or question can use them together.
Maven syncs pageSpeed performance snapshot from Google PageSpeed Insights.
Maven refreshes Google PageSpeed Insights on a schedule after the source is connected, and shows each source’s last sync time in the client’s workspace, so you can check the data is current before a report goes out.
Yes. Maven encrypts integration credentials and tokens, scopes access to the workspace, and uses the connection only to sync the reporting data needed for governed analytics.
Set up your first client, build their report by asking Claude or ChatGPT, and publish it to their portal.
Nothing is charged unless you subscribe.