Today we're releasing an update to Publisher Ads Audits for Lighthouse, which focuses on improving the real-world accuracy of mobile reports via simulated throttling. Continue reading to learn more about the new simulated throttling option and how it will affect mobile audit scores.
To address this, we are introducing a simulated throttling option. When enabled, mobile audit scores are calculated by first auditing your page under normal conditions, then using this data to simulate page load performance under mobile conditions. The simulation uses the Lighthouse mobile network throttling preset, which artificially limits both CPU speed and network bandwidth. This approximates the performance of an actual mobile device loading your site over a fast 3G connection.
Note that simulated throttling only applies to mobile audits. Desktop audits will not be throttled regardless of whether this new option is enabled or not.
While metric thresholds have been adjusted to account for simulation, we expect this new option may cause scores to change for many pages that were previously audited without throttling. If you audit your page regularly and have established a performance baseline based on the previous mobile audit behavior, you may need to establish a new baseline to account for these differences.
Questions or feedback about this or anything else ad speed related? Reach out on our GitHub issue tracker.
- Jonathon Imperiosi, Publisher Ads Audits Team
What is simulated throttling?
Until now, all audit scores were based on the result of loading your page on a desktop CPU over an unthrottled, wired broadband connection. While mobile audits request the mobile version of your page (by identifying as a Nexus 5X phone), the results did not reflect actual mobile CPU and bandwidth constraints.To address this, we are introducing a simulated throttling option. When enabled, mobile audit scores are calculated by first auditing your page under normal conditions, then using this data to simulate page load performance under mobile conditions. The simulation uses the Lighthouse mobile network throttling preset, which artificially limits both CPU speed and network bandwidth. This approximates the performance of an actual mobile device loading your site over a fast 3G connection.
Note that simulated throttling only applies to mobile audits. Desktop audits will not be throttled regardless of whether this new option is enabled or not.
How does this affect me?
The new simulated throttling option will be enabled by default for all newly generated reports, matching the behavior of other performance auditing tools such as Lighthouse, PageSpeed Insights, and the Chrome DevTools Audits' panel. It can be disabled on a per-report basis by toggling the option in advanced settings.While metric thresholds have been adjusted to account for simulation, we expect this new option may cause scores to change for many pages that were previously audited without throttling. If you audit your page regularly and have established a performance baseline based on the previous mobile audit behavior, you may need to establish a new baseline to account for these differences.
Questions or feedback about this or anything else ad speed related? Reach out on our GitHub issue tracker.
- Jonathon Imperiosi, Publisher Ads Audits Team