Timeline of web vitals

The timeline currently offers focused coverage of the period until November 2024. It is likely to miss important developments outside this period (particularly after this period) though it may have a few events from after this period.

This is a timeline of web vitals, a set of metrics championed by Google that relate to how well web pages load. Of particular interest are three metrics Google calls core web vitals, whose values in field data are used to inform a "page experience" score that affects search ranking within Google.

Big picture

Time period	Development summary
2007 – 2015	This period sees the emergence of diagnostic tools and web standards that help diagnose page load speed. This is also the period where some initial concepts such as "speed index" and "performance budgets" start getting fleshed out.
2016 – 2018	This period sees the emergence of systematic datasets covering web performance for a wide range of websites, including the HTTP Archive (that uses Lab data) and the Chrome User Expeience Report (CrUX) that uses field data collected anonymously from Chrome users. It also sees the growth of public diagnostic tools for site performance, including Lighthouse, that is integrated into Chrome's Developer Tools.
2019	This is the year when Google starts getting more directly involved with improving site performance, with communication being done via a bunch of blog posts on web.dev. During this year, the understanding of site performance is expanded to include visual stability (epitomized by Cumulative Layout Shift (CLS)). Previously the focus had been on loading speed and interactivity/responsiveness.
2020	This is the year that Google officially announces three "core web vitals": Largest Contentful Paint (LCP) for loading speed, Cumulative Layout Shift (CLS) for visual stability, and First Input Delay (FID) for interactivity/responsiveness. Google also provides guidance that the 75th percentile field data values for these metrics will be used to influence search results on mobile and desktop in the coming years. In particular, Google clarifies its focus on "field data" (as collected by the CrUX report) as the data it uses for official evaluation, and clarifies the role of Lab data as an aid for self-diagnosis by people involved with developing websites. The potential of core web vitals to affect search results leads to a flurry of activity in the SEO community. A number of new diagnostic tools and data sources emerge, both from Google and from the rest of the ecosystem, and existing tools on site performance shift direction a little bit to start focusing on core web vitals.
2021 – 2022	During this period, Google starts using core web vitals to influence search results, with mobile search results being affected starting mid-2021 and desktop search results being affected starting early 2022. However, the magnitude of impact on search results is small, leading to a dying down of activity and attention to core web vitals in the SEO community. Nonetheless, the overall proportion of origins passing on the core web vitals continues to improve over the period, with CLS (one of the more actionable core web vitals) seeing the most dramatic improvement. The improvements in the case of CLS are due to a combination of definition changes and websites making improvements. The less dramatic improvement for LCP is due to a combination of Chrome browser speed updates, computers getting a litle faster over time, and websites making improvements.
2023 – 2024	Despite its much-reduced profile in the context of SEO, measured core web vitals in field data on websites continue to improve over this period, likely due to a mix of continued effort on the part of various teams at Google (including those working on Chrome), the inclusion of core web vitals as a measurement yardstick for CMSes and software tools, and continued improvements in hardware and network connectivity as users upgrade to newer devices and faster connections. The main change to metrics is the replacement of FID by Interaction to Next Paint (INP) as the interactivity/responsiveness metric. Whereas almost all websites were passing FID because the threshold was easy to achieve, INP is a somewhat harder-to-crack metric, and this metric change leads to a new era of optimization work around interactivity/responsiveness.

List of web vitals and corresponding aspects of performance

Note that for the three web vitals measured in both field and Lab data, the thresholds for good and poor for field data match the thresholds for good and poor on Lab data (measured using Lighthouse) respectively. This has been the case since June 2021 updates to field data thresholds and the corresponding release of Lighthouse v8. The desktop Lab data thresholds differ for some web vitals, and are not in the table below to keep it simple.

Web vital	Aspect of performance	Core web vital?	Reported in field data (PageSpeed Insights, CrUX report)?	Reported in Lab data (Lighthouse, PageSpeed Insights)?	Reported in WebPageTest web vitals?	Threshold for "good" on field data (if applicable) and mobile Lab data (if applicable) (in milliseconds except CLS that is unitless)	Threshold for "poor" on field data (if applicable) and mobile Lab data (if applicable) (in milliseconds except CLS that is unitless)
Time to First Byte (TTFB)	Loading speed	No	Yes	No	No	800	1,800
First Contentful Paint (FCP)	Loading speed	No	Yes	Yes	Not online, but supported in code	1,800	3,000
Largest Contentful Paint (LCP)	Loading speed	Yes	Yes	Yes (since v6)	Yes	2,500	4,000
Speed Index (SI)	Loading speed	No	No	Yes (some variant since v2)	No	3,400	5,8000
Time to Interactive (TTI)	Loading speed, interactivity/responsiveness	No	No	No (retired in v10; was reported until v9 and since at least v5, some variant since v2)	Not online, but supported in code	3,800	7,300
Total Blocking Time (TBT)	Interactivity/responsiveness	No	No	Yes (since v6)	Yes	200	600
First Input Delay (FID)	Interactivity/responsiveness	No (retired March 2024)	Yes	No	No	100	300
Interaction to Next Paint (INP)	Interactivity/responsiveness	Yes (starting March 2024)	Yes	No	No	200	500
Cumulative Layout Shift (CLS)	Visual stability	Yes	Yes	Yes	Yes	0.1	0.25

Numerical and visual data

The table below is based on monthly summaries provided in the CrUX announce mailing list. Note that the total number of origins is not necessarily the denominator for the percentages; individual percentages have as denominator only the number of origins for which there was sufficient data for that particular metric.

Overall, the data show a clear increase in the good percentage for LCP and CLS (the two metrics for which we have data throughout the period). FID shows improvement over the period for which data for it is available, though it reaches close to a flat line toward the end. Its replacement, INP, shows steady improvement over the period for which data for it is available.

In some cases, there are changes to the metric due to updates to the way Chrome measures it (the browser responsible for field data reporting) without any corresponding change to actual user experience. These changes can be both positive and negative; where identified, they are included in the notes column for that month. The bulk of the improvement in metrics seems to reflect genuine increases in the underlying conceptual metrics, and therefore, likely reflect genuine improvements in the user experience. This sustained improvement stands in contrast to ideas such as Wirth's law that would suggest that improvements in underlying hardware speed would be cancelled by bloat in browsers and wesites; it seems that even if websites are getting somewhat more bloated, the net effect on the metrics is still in the direction of improvement.

Another potential concern is Goodhart's law, which, applied to these metrics, would suggest that the way the metrics are improving is that websites are specifically targeting the metrics in ways that don't improve the facets of user experience that the metrics are intended as proxies for. Goodhart's law is more of an issue with Lab data (with hacks to optimize Lighthouse scores through cloaking-like practices) but can also be applied to field data. Generally, Goodhart's law loopholes for LCP and INP have been getting closed over time, so Goodhart's law is unlikely in practice to explain much of the improvement over time.

Year and month (YYYY-MM)	Number of origins covered in dataset	Percentage good Largest Contentful Paint (LCP)	Percentage good First Input Delay (FID) -- replaced by INP as a core web vital in March 2024	Percentage good Cumulative Layout Shift (CLS)	Percentage good on all core web vitals	Percentage good on Interaction to Next Paint (INP) -- becomes a core web vital in March 2024	Percentage good on CLS, LCP, and INP (the set of core web vitals starting March 2024)	Special notes
2020-05	7,103,486	46	86	64	27			Corrected data to restrict denominator
2020-06	7,501,835	47.84	86.74	57.45	25.34
2020-07	7,527,878	47.6	86.9	57.2	25.2
2020-08	7,773,359	47.1	86.9	57.1	24.9			Bump in origin coverage due to a data pipeline change that now accepts all user experiences with a FCP
2020-09	7,937,088	46.8	87.6	55.8	24.6
2020-10	8,050,755	47.0	88.2	52.3	23.7
2020-11	7,942,408	47.0	88.8	51.1	23.5
2020-12	7,629,156	47.1	88.7	51.0	23.4
2021-01	8,185,540	47.81	89.28	50.25	23.71
2021-02	8,264,371	47.99	89.46	45.99	21.98
2021-03	8,326,310	49.01	89.66	51.84	24.81
2021-04	8,423,302	49.4	89.8	53.8	25.8
2021-05	8,411,670	50.7	89.9	60.4	29.2			Metric definition updates released in early June are applied to this data, including the 5-second session window for CLS, and at least partly explains the sharp increase in the percentage of origins with good CLS.
2021-06	8,416,608	50.6	93.0	61.0	30.6			The effect of the double-tap-to-zoom FID metric change in Chrome 91, that improves measured FID, is largely seen in this month, and at least partly explains the sharp increase in the percentage of origins with good FID.
2021-07	8,174,923	50.0	93.2	61.2	30.4
2021-08	8,431,699	50.9	93.2	62.1	31.3
2021-09	8,660,068	50.9	93.3	65.4	32.8
2021-10	8,784,894	50.0	93.4	65.9	34.1			This is the first month when the percentages reported are based on the "optional FID" methodology adopted in PageSpeed Insights on June 10, 2021.
2021-11	8,733,078	50.8	93.5	66.1	34.8			The uncapped CLS metric (that was the old way of calculating CLS before a definition change around May/June) is retired from the BigQuery dataset starting with this month.
2021-12	8,398,796	51.1	93.3	66.1	35.0
2022-01	8,934,350	52.4	93.9	68.9	37.6
2022-02	8,764,246	52.7	94.9	70.6	39.0			This is the first month for which the dataset includes data on the experimental responsiveness metric that had been described by Google in blog posts in 2021.
2022-03	8,555,307	53.4	94.9	70.6	39.6
2022-04	8,602,902	55.2	94.9	71.6	41.2			This is the first month for which the dataset includes data on a new experimental metric called Interaction to Next Paint (INP), a new variant of the experimental responsiveness metric.
2022-05	11,024,795	56.5	94.7	71.7	42.0			This is the first month where effective connection type and form factor fields transition from required fields to optional fields in the CruX dataset, resulting in more eligible traffic data for websites, and therefore a larger number of origins with enough traffic to qualify for inclusion in the CrUX report. A bug resulted in this change not having its full effect (on increase in origin coverage) in this month; the full effect is expected to be seen in June. Also, the LCP improvement is attributed to an increase in Chrome's speed on Android devices, "thanks to prioritizing critical navigation moments on the browser user interface thread."
2022-06	16,230,572	54.9	94.3	72.3	41.3			There is a significant increase in the number of origins compared to last month and even more so compared to two months ago, the increase over the past two months is due to making form factor and effective connection type optional. Due to the increase in coverage of low-traffic origins, the numbers aren't directly comparable with previous months; in particular, the seeming decline in LCP is mostly a result of the expanded coverage. Restricting to the top 1,000 or top 10,000 origins shows a clearer trend of continued improvement.
2022-07	16,190,453	54.3	94.2	71.7	40.6			This release includes a change which records the CLS metric at the first OnHidden in addition to tab close. Also, there is a bug in the origin filtering causing a small number of origins to be incorrectly filtered from the release.
2022-08	16,754,655	53.7	94.2	72.5	40.7			This release continues to partially have the bug in the previous release due to which some origins are missed; the number of origins missed is smaller than last time but still not zero.
2022-09	17,715,277	54.5	94.8	72.4	41.4
2022-10	17,637,195	54.8	95.0	72.3	41.6			This release includes a breakdown of "rank" popularity for top 5,000 sites, top 50,000 sites, etc. in addition to the already-existing rank popularity for top 1,000 sites, top 10,000 sites etc.
2022-11	17,618,944	54.9	95.1	72.4	41.8
2022-12	16,824,271	54.8	95.0	72.7	41.8
2023-01	18,200,000	56.2	95.2	73.1	43.0
2023-02	18,184,396	56.3	95.3	73.1	43.1
2023-03	18,495,210	57.7	95.5	73.4	44.2
2023-04	18,406,973	57.1	95.3	72.9	43.5			This month sees a slight decrease in FCP and LCP pass rates. The FCP pass rate decrease is attributed to a paint timing fix in Chrome 109. The LCP pass rate decrease is attributed to a change to exclude low-entropy images introduced on April 6 and also, to a lesser extent, the paint timing change.
2023-05	18,377,791	56.8	95.5	73.3	43.6			This month sees a slight decrease in FCP and LCP pass rates. The FCP pass rate decrease is attributed to a paint timing fix in Chrome 109. The LCP pass rate decrease is attributed to a change to exclude low-entropy images introduced on April 6 and also, to a lesser extent, the paint timing change.
2023-06	18,065,718	57.2	95.6	73.7	44.1	77.7		The data for this month includes partial resolution of the "root page redirect issue" where some origins that redirect their root page are excluded from CrUX. This is also the first month where the INP metric is included without the experimental prefix (it's also included with the experimental prefix for backward compatibility). This is in preparation for INP becoming a core web vital in March 2024.
2023-07	17,976,663	57.8	95.7	73.8	44.6	77.8		The data for this month is after full resolution of the "root page redirect issue" where some origins that redirect their root page are excluded from CrUX.
2023-08	18,263,523	58.3	95.8	74.4	45.3	77.6	41.6	The data for this month includes the partial impact of several Chrome improvements affecting CLS and LCP, with the effect generally positive.
2023-09	18,405,462	58.5	95.8	74.7	45.7	78.0	42.1	The improvements introduced in the previous month, that continued to roll out this month, are likely responsible for continued improvements to LCP and CLS, despite slight downturns in FCP and TTFB.
2023-10	18,383,750	59.4	95.8	75.9	46.9	78.7	43.4	The final rollout of the improvements introduced in August, that continued to roll out this month, are likely responsible for continued improvements to LCP and CLS.
2023-11	18,265,721	59.8	96.0	76.2	47.4	79.2	43.9
2023-12	17,323,447	60.4	95.9	76.4	48.0	79.0	44.4
2024-01	18,583,729	61.6	96.2	76.4	49.0	80.7	45.8
2024-02	18,729,879	61.3	96.2	76.6	48.8	80.6	45.6
2024-03	18,669,191	62.2		77.1		82.1	46.8	With INP replacing FID as a core web vital, FID-based data stops being included in this month. There are two concrete sets of improvements that may be contributing to LCP and INP improvements: improved cookie caching on Android Chrome, and consent management platforms improving their code to yield more readily.
2024-04	18,703,230	62.4		77.5		82.8	47.3
2024-05	18,673,241	62.1		77.4		83.2	47.1
2024-06	18,442,199	63.4		77.8		84.1	51.0	The improvement in core web vitals is attributed as plausibly driven by seasonal factors, and the INP increase is attributed partly to changes introduced in Chrome 127, including the changed treatment of presentation modals in the INP calculation.
2024-07	18,176,653	64.1		78.0		84.4	49.2	The continued improvement in INP is attributed to the continued rollout of improvements introduced in Chrome the previous month.
2024-08	18,394,950	64.8		78.2		85.0	50.1	The continued improvement in INP is attributed to a mix of a measurement change from Chrome (no longer using pointerup events for INP calculation) and improvements in the Google Publisher Tag library to yield on out-of-viewport ads.
2024-09	18,936,684	65.4		78.2		85.7	51.0	The continued improvement in INP is attributed to the finishing of the rollout of the measurement change in Chrome from last month (no longer using pointerup events for INP calculation).
2024-10	18,814,228	65.5		78.2		85.9	51.2

Full timeline

Inclusion criteria

The following are some kinds of events that qualify for inclusion in the timeline:

• Launch of new private or public diagnostics tools related to web vitals, or major updates to such tools
• Major version updates to Lighthouse, the main Lab data tool
• Introduction of a new web vital metric, changes to the definition of a web vital metric, or addition or removal of a web vital from the list of core web vitals
• Changes to browser behavior that could affect the data for one or more web vitals
• Blog posts or articles with insight around how to improve on web vitals, or critical discussion of web vitals and how much importance to give to them
• Important conferences about site performance (such as the performance.now() conference)

Here are some examples of things that will not qualify for this timeline:

• Minor version updates to Lighthouse
• Performance trajectories of specific websites, except to the extent that the analysis of such performance qualifies as valuable insight into web vitals
• Individual talks at a conference (such as the performance.now() conference) unless the specific talk breaks a lot of new ground and gets a bunch of commentary; in many cases, talks at these conferences summarize past discussion so the original post or articles for such discussions should be included instead

Timeline

Year	Month and date (if available)	Event type	Type of tool, entity, or change	Aspect of performance (specific metrics in parentheses)	Type of measurement (field data (from real users; aka Real User Monitoring (RUM)) or Lab data (from a bot or synthetic local testing))?	Event
2007	000000002024-07-01-0000July	New product	Public diagnostic tool	Loading speed	Lab data	Yahoo! launches YSlow, a Firefox add-on built on top of Firebug, to help debug speed issues on websites. YSlow operationalilzes Yahoo! Chief Performance Officer Steve Souders' 13 rules for high-performance web sites (also covered in his book "High Performance Web Sites"). Development on it would cease after 2014 as people migrated to other tools like Google's Page Speed, Webpagetest, and later Lighthouse.
2009	000000002024-06-05-0000June 5	New product	Public diagnostic tool	Loading speed	Lab data	Google releases Page Speed, a Firefox add-on built on top of Firebug, to help debug speed issues on websites. Commentators consider this to be Google's equivaent to Yahoo's YSlow tool.
2012	000000002024-04-26-0000April 26	Documentation commit	Web standard	Detailed timing information	Field data, Lab data	The first commit is made to the repository with the specification for the Navigation Timing API. The API allows for access to information on the fetching of the root document and its loading in the browser.
2013	000000002024-04-09-0000April 9	Documentation commit	Web standard	Detailed timing information	Field data, Lab data	The first commit is made to the repository with the specification for the Resource Timing API. The API allows for access to information on the timing of download of various resources.
2014	000000002024-09-29-0000September 29	Documentation commit	Web standard	Loading speed (Speed Index)	Field data	The first commit is made to the repository with the specification for the RUM Speed Index (RUM stands for real user monitoring), an attempt to calculate a Speed Index based on field data. As of 2021, Speed Index is still calculated in the Lab.
2014	000000002024-11-18-0000November 18	Insight	Metric definition, concept development	Loading speed	Field data, Lab data	Tim Kadlec publishes his landmark post titled "Performance Budget Metrics" that classifies metrics used for performance budgets in four categories: (1) milestone timings (measuring when specific milestones in loading and rendering are achieved), (2) SpeedIndex (measuring how a page loads from start to finish), (3) quantity-based metrics (such as total number of requests, overall page weight, total image weight), and (4) rule-based metrics (such as PageSpeed and YSlow score). Performance budgets would eventually be supported by build tools and diagnostic tools, and would be the subject of further analysis; one analysis on web.dev would use a similar categorization of performance budget metrics (without a separate category for SpeedIndex).
2016	000000002024-01-15-0000January 15	Code commit	Public diagnostic tool	Loading speed, interactivity/responsiveness	Lab data	The first code commit to the Google Lighthouse GitHub repository is on this date.
2016	000000002024-06-30-0000June 30	Release	Public diagnostic tool	Loading speed, interactivity/responsiveness	Lab data	The Lighthouse v1.0.3 tag is created, suggesting the official release of Lighthouse 1.0.
2016	000000002024-09-14-0000September 14	Documentation commit	Web standard	Loading speed (FCP)	Field data, Lab data	The first commit is made to the repository documenting the Paint Timing API. This API is used to calculate metrics such as the First Contentful Paint (FCP). Initially supported only on Chromium browsers, the Paint Timing API would eventually be supported by Firefox (Gecko) and Safari (WebKit) as well.
2017	000000002024-02-01-0000February 1	Data collection	Public dataset	Loading speed	Lab data	Data for the HTTP Archive's Loading Speed dataset is available starting this day.
2017	000000002024-08-24-0000August 24	New version	Public diagnostic tool	Loading speed, interactivity/responsiveness	Lab data	Lighthouse v2.0.0 is released. The last commit for it is from May 19. At this point, the Performance section of Lighthouse has the following metrics: First meaningful paint, First interactive (beta), Consistenly interactive (beta), Perceptual Speed Index, and Estimated Input Latency.
2017	000000002024-10-22-0000October 22	Insight		Loading speed, interactivity/responsiveness		Alex Russell publishes the first post of his multi-year series "The Performance Inequality Gap"; in the first post, Russell, who is at Google at the time, lays out suggested targets for page load speed (5 seconds to interactive for first load, 2 seconds for subsequent loads) as well as the budgets this places on the amount of resources of various kinds (HTML, CSS, JavaScript) on the page. Russell also provides extensive reasoning for his suggested budgets. In later posts, published over the years, Russell updates his suggested budgets, taking into account the evolution of devices and network connectivity. Overall, Russell is critical of the heavy use of client-side JavaScript in modern web apps and of the focus on "developer experience" in web development at the cost of user experience.
2017	000000002024-10-01-0000October 1	Data collection	Public dataset	All	Field data	Data for the Chrome User Experience (CrUX) report, collected anonymously from Chrome users who have consented to the anonymous data collection, is available in a BigQuery dataset starting this day.
2018	000000002024-07-11-0000July 11	New version	Public diagnostic tool	Loading speed, interactivity/responsiveness	Lab data	Lighthouse v3.0.0 is released. The last commit for it is from June 28. This release includes changes to performance weights and scoring thresholds.
2018	000000002024-08-11-0000August 11	Release	Public ranking tool	Loading speed, interactivity/responsiveness	Lab data	The earliest Wayback Machine snapshot of webperf.xyz, a site that maintains a leaderboard of sites based on their performance, is from this date. At launch time, it uses webpagetest.org for its data generation. The tool would be referenced in an Atlantic article on August 23 about their efforts to speed up ad loading on their site.
2019	000000002024-01-16-0000January 16	New version	Public diagnostic tool	Loading speed, interactivity/responsiveness	Lab data	Lighthouse v4.0.0 is released. The last commmit for it is from January 15.
2019	000000002024-05-07-0000May 7	New version	Public diagnostic tool	Loading speed, interactivity/responsiveness	Lab data	Lighthouse v5.0.0 is released. The metrics used in the performance score, with weights, are: First Contentful Paint (FCP) (23%), Speed Index (SI) (27%), First Meaningful Paint (FMP) (7%), Time To Interactive (TTI) (33%), First CPU Idle (FCI) (13%), and Max Potential FID (0%).
2019	000000002024-05-13-0000May 13	Documentation commit	Web standard	Visual stability (CLS)	Field data, Lab data	The first commit to the GitHub repository for the Layout Instability API is made on this day. This is a proposal by Google for adoption as a web standard, though as of June 2021 it has not been adopted and is only used on Blink-based browsers, which is effectively just Chromium browsers (mainly Chrome, Edge, and Opera).
2019	000000002024-05-15-0000May 15	Documentation commit	Web standard	Loading speed (LCP)	Field data, Lab data	The first commit to the GitHub repository for the Largest Contentful Paint API is made on this day. This is a proposal by Google for adoption as a web standard, though as of June 2021 it has not been adopted and is only used on Blink-based browsers, which is effectively just Chrome browsers (mainly Chrome, Edge, and Opera).
2019	000000002024-06-11-0000June 11	Announcement	Metric introduction	Visual stability (CLS)	Field data, Lab data	The blog post "Cumulative Layout Shift (CLS)" is published on web.dev. In the post, Google describes a new metric, Cumulative Layout Shift (CLS), that measures the totality of unexpected shifts during the lifecycle of a page. This builds upon the Layout Instability API that Google started workingg on recently.
2019	000000002024-08-08-0000August 8	Announcement	Metric introduction	Loading speed (LCP)	Field data, Lab data	The blog post "Largest Contentful Paint (LCP)" is published on web.dev. In the post, Google describes a new metric, Largest Contentful Paint (LCP), that measures how long it takes for the page's main content to have loaded.
2019	000000002024-09-19-0000September 19	Insight	Public diagnostic tool	All	Lab data	The page "Lighthouse performance scoring" gives insight into how Google Lighthouse scores are calculated. In particular, it explains that the log-normal distribution used for translating each of the web vitals into a score component is based on real website performance data on HTTP Archive.
2020	000000002024-04-30-0000April 30	Announcement	Metric selection, thresholds	All	Field data, Lab data	Google announces the selection of three of its metrics as core web vitals: Largest Contentful Paint (LCP), First Input Delay (FID), and Cumulative Layout Shift (CLS). The post also states that the 75th percentile of fielld data should be used for any assessment using web vitals, and specifies thresholds for good and poor values for each of the core web vitals.
2020	000000002024-05-05-0000May 5	Insight		All	Field data, Lab data	Google publishes three blog posts, one for each of the three core web vitals (LCP, FID, and CLS), on strategies to optimize that core web vital.
2020	000000002024-05-12-0000May 12	New version	Public diagnostic tool	All	Lab data	In version 21.07, WebPageTest adds support for web vitals. WebPageTest, available at webpagetest.org, is a tool that can be used to report performance data on any web page.
2020	000000002024-05-19-0000May 19	New version	Public diagnostic tool	All	Lab data	Lighthouse v6.0.0 is released (it is expected to ship in the DevTools of Chrome 84, that releases July 14). The performance scoring is updated, with three new metrics added: Largest Contentful Paint (LCP), Total Blocking Time (TBT), and Cumulative Layout Shift (CLS); two of which are among Google's three core web vitals. Three metrics from Lighthouse v5 are removed: First Meaningful Paint, First CPU Idle, and Max Potential FID. The weights are as follows: First Contentful Paint (FCP) (15%), Speed Index (SI) (15%), Largest Contentful Paint (LCP) (25%), Time To Interactive (TTI) (15%), Total Blocking Time (TBT) (25%), and Cumulative Layout Shift (CLS) (5%). This is the first time that Lab data includes a metric for visual stability (the specific metric being CLS).
2020	000000002024-05-21-0000May 21	Insight	Thresholds	All	Field data	In a lengthy blog post, Google describes how it decided on the 75th percentile of field data as the point to use for assessing each of its metrics, and also explains how thresholds were selected for each of the threee core web vitals: LCP, FID, and CLS. The decision for each metric is made by a combination of acceeptable user experience and achievability by webpages with current technology.
2020	000000002024-05-27-0000May 27	Release	Private diagnostic tool	All	Field data	Google updates Google Search Console adding a section on core web vitals. This section reports on field data performance on both mobile and desktop on each of the three core web vitals (CLS, LCP, and FID). It is based on the CrUX report. Rather than making all the data available for each url, Search Console groups urls together and reports on aggregate performance of the url groups.
2020	000000002024-05-28-0000May 28	Announcement	Search algorithm update	All	Field data	Google announces that at some point in the future, it will start using "page experience" as a ranking factor in search. The page experience signal combines core web vitals, mobile usability, security issues, whether the site is HTTPS, and no intrusive intersitials.
2020	000000002024-06-24-0000June 24	Implementation	Web standard	Loading speed (FCP)	N/A	A blog post on the Wikimedia Foundation's tech blog describes work done by Wikimedia Foundation engineers to implement the Paint Timing API in WebKit, the engine used by Safari, thus making it available in future versions of Safari.
2020	000000002024-11-10-0000November 10	Announcement	Search algorithm update	All	Field data	Google announces that it will start using "page experience" as a ranking factor for mobile searches starting May 2021. The page experience signal combines core web vitals, mobile usability, security issues, whether the site is HTTPS, and no intrusive interstitials.
2020	000000002024-11-16-0000November 16	New version	Public diagnostic tool, private diagnostic tool	All	Lab data	GTmetrix, a performance optimization tool, switches from its legacy Timings and PageSpeed/YSlow measurements to using Lighthouse as its underlying engine; the new data includes Performance and Structure tabs, an overall GTMetrix grade, and a web vitals section that includes LCP, CLS, and Total Blocking Time (TBT).
2020	000000002024-12-17-0000December 17	New version	Public diagnostic tool	All	Lab data	Lighthouse v7.0.0 is released. It is expected to be shipped with Chrome 89 (released March 2, 2021) and becoms part of PageSpeed Insights on February 19, 2021.
2021	000000002024-02-17-0000February 17	Release	Threshold update	All	Field data	Field data available in Google Search Console and PageSpeed Insights are updated to consider a metric value at a threshold as being on the good side of the threshold. For instance, for CLS, the the threshold for good CLS is 0.1; a 75th percentile CLS value of exactly 0.1 would now be considered good.
2021	000000002024-04-07-0000April 7	Announcement	Metric definition update	Visual stability (CLS)	Field data, Lab data	Google announces a planned update to the calculation of Cumulative Layout Shift (CLS) to be fairer to long-lived pages. The update changes the definition of CLS to look at the largest CLS over 5-second windows with a 1-second gap. The rollout of this definition update would happen in June.
2021	000000002024-04-19-0000April 19	Release	Private diagnostic tool	All	Field data	A new section called "Page experience on mobile" is added to Google Search Console (where people can access this data for their own sites only). This is restriced to mobile searches; it reports on the percentage of urls and number of search impressions that have "good page experience" i.e., that perform well on core web vitals, mobile usability, security issues, whether the site is HTTPS, and ads experience.
2021	April	Insight	Data	All	Lab data	Searchmetrics publishes a study of core web vitals, looking at Lab data for over 2 million URLs on two of the three core web vitals (LCP and CLS) and a proxy for the third core web vital (using Total Blocking Time (TBT) instead of FID). YouTube is highlighted as an outlier that skews measurement.
2021	000000002024-04-19-0000April 19	Announcement	Search algorithm update	All	Field data	Google announces an updated timeline of mid-June to August for its rollout of page experience as a ranking factor for mobile searches. The previous announced rollout time was May 2021; the time extension is allegedly to give websites more time to prepare.
2021	000000002024-04-25-0000April 25	New version	Public ranking tool	All	Lab data	The backend for webperf.xyz is switched from webpagetest to Lighthouse on this date (or least, this is the earliest date with data recorded using Lighthouse).
2021	000000002024-05-04-0000May 4	Insight	Cross-browser comparison of metric definition	Loading speed (FCP)	N/A	A blog post on the WebPageTest blog describes challenges with comparing First Contentful Paint (FCP) across browsers. FCP is one of the first web vitals to be available across all major browsers with the release of Safari 14.1.
2021	000000002024-05-18-0000May 18	Insight	Business impact	All	Field data (mostly) with some discussion of Lab data	In a blog post on web.dev that also includes a video, two Google staff members discuss the business impact of (core) web vitals. The post includes case studies of: Vodafone, iCook, Tokopedia, Redbus, and several others who have improved one or more of the core web vitals and seen improvement in traffic, user engagement, and revenue. These impacts are prior to (and therefore not caused by Google starting to use page experience (including core web vitals) as a signal for search ranking.
2021	000000002024-05-25-0000May 25	New version	Browser behavior	Interactivity/responsiveness (FID)	Field data	Chrome 91 disables double-tap-to-zoom when the viewport meta tag specifies width=device-width or initial-scale>=1.0, even when implicitly doing so, like for example in minimum-scale=1.5, maximum-scale=2. Because DTZ negatively impacts FID and the amount of pages where DTZ is disabled is increased, Google expects some sites to see FID improvements.
2021	000000002024-06-01-0000June 1  – 2, some changes a little later in June	Release	Metric definition update	Visual stability (CLS)	Field data	Google updates its field data tools including the Chrome User Experience (CrUX) report, PageSpeed Insights, and Google Search Console, to use an updated definition of Cumulative Layout Shift (CLS) that looks at the largest CLS over 5-second windows with a 1-second gap. The planned change itself had been announced back in April, but it had not yet been reflected in the tools Google used to report on field data, as well as in Google's Lighthouse tool available on PageSpeed Insights and in Chrome Developer Tools (canary channel only, so not part of officially released Chrome yet).
2021	early 000000002024-06-01-0000June	Release	Metric definition update	Loading speed (LCP)	Field data	Some changes are made to LCP calculations to account for offscreen images and multiple images of the same size.
2021	000000002024-06-02-0000June 2	New version	Public diagnostic tool	All	Lab data	Lighthouse v8.0.0 is released. The release is made available on PageSpeed Insights immediately, and is expected to ship as part of Chrome 93. This includes Lab-side changes corresponding to field data updates for CLS released at around the same time; it also includes threshold changes for TBT and FCP and a reweighting: FCP: 15% to 10%, SI: 15% to 10%, TTI: 15% to 10%, TBT: 25% to 30%, and CLS: 5% to 15%. A new diagnostic aide called the Lighthouse Treemap is also released.
2021	000000002024-06-10-0000June 10	New version	Public diagnostic tool	All, FID	Field data	PageSpeed Insights field data now reports field data even for pages where some (but not all) of the field data metrics have insufficient data; in such cases, only the metrics with sufficient data are reported. Rick Viscomi further clarifies on Twitter that this will most affect pages missing FID, since FID data tends to be much more likely to be missing than LCP or CLS data. Further, an origin with insufficient FID data will be considered to pass overall in core web vitals if it passes LCP and CLS; however, an origin must have LCP and CLS data in order to be considered passing on core web vitals.
2021	000000002024-06-15-0000June 15  – September 2	Release	Search algorithm update	All	Field data	The rollout of Google's "page experience on mobile" update happens during this period. The update is limited to mobile searches, and gives weight to a "page experience" factor that includes performance on core web vitals, mobile usability, security issues, whether the site is HTTPS, and ads experience.
2021	000000002024-06-21-0000June 21, 000000002024-11-03-0000November 3	Proposal	Metric definition update	Interactivity/responsiveness (responsiveness metric)	Field data	In two blog posts on web.dev. the Chrome Speed team proposes a new responsiveness metric, that is expected to correlate with the Lab data metrics of time to interactive (TTI) and total blocking time (TBT) and whose goal is to measure responsiveness over the page lifecycle. It is expected to overcome some of the shortcoming of first input delay (FID) and be a more holistic measure of responsiveness.
2021	000000002024-06-24-0000June 24	Release	Public diagnostic tool	All	Field data	The HTTP Archive's Core Web Vitals Technology Report is released. This builds on the Chrome User Experience Report (CrUX report) as well as the HTTP Archive's data on CMS and technology stacks used by various websites, to provide a big-picture view of performance on core web vitals by CMS and technology stack.
2021	000000002024-08-17-0000August 17	Insight	Search algorithm update	All	Field data, Lab data	A blog post on the Semrush blog reviews the rollout of Google's page experience update. It goes over trends in field data and Lab data performance of mobile urls before and during the rollout, as well as the correlation with mobile search rankings. It detects a small correlation in the expected direction.
2021	000000002024-08-17-0000August 17	Insight	Public diagnostic tool, private diagnostic tool	All	Field data, Lab data	A blog post on web.dev describes how field and Lab data differ conceptually, how the values of field data and Lab data might differ in general, and concrete field/Lab differences associated with the three core web vitals.
2021	000000002024-08-31-0000August 31	New version	Browser behavior	Visual stability (CLS)	Field data	Chrome 93 makes changes to its calculation of cumulative layout shift (CLS) to ignore changes to offscreen elements countered with scroll anchoring, as well as layout shifts while dragging or resizing elements with a mouse. This is expected to improve the performance of websites on CLS.
2021	000000002024-09-16-0000September 16	Insight	Search algorithm update	All	Field data	A blog post on the Sistrix blog titled "Core Web Vitals is a Measurable Ranking Factor" documents the observed effect size on ranking of sites passing all core web vitals versus the average and versus sites failing all core web vitals.
2021	000000002024-09-21-0000September 21	New version	Browser behavior	Loading speed (FCP)	Field data, Lab data	Chromium is updated so that paint timing is not triggered by content that is not visible due to being under the effect of an opacity: 0 style. This change is part of the Chrome 94 release on September 21, 2021, and is also expected to affect other Chromium-based browsers such as Edge and Opera. The change is expected to increase (worsen) measured FCP (in field and Lab data) for some sites, without directly changing the end user experience.
2021	000000002024-10-01-0000October 1	Data collection	Public dataset	All	Field data	This is the first day for which data is available in any dataset of the RUM Archive. RUM stands for Real User Monitoring. The RUM Archive is an archive managed by Akamai for which data is collected via boomerang.js and mapped using Akamai mPulse. Data would initially be available only for the first day of the month, before transitioning to being available for every day, starting September 1, 2022.
2021	000000002024-10-27-0000October 27	Insight	Search algorithm update	All	Field data	A post on Search Engine Land describes the relatively minor impact that page experience has had on search rankings so far, as well as SEOs questioning the value of spending time on improving page experience. It also discusses other potential benefits of improving page experience beyond the search ranking impact.
2021	000000002024-11-03-0000November 3	New version	Public diagnostic tool	All	Field data, Lab data	A new version of the PageSpeed Insights UI is released. Around this time (by November 15), PageSpeed Insights is moved to pagespeed.web.dev.
2021	000000002024-11-13-0000November 13, November 15	New version	Public diagnostic tool	All	Lab data	Starting November 13, PageSpeed Insights starts using Lighthouse 9.0.0. On November 15, Lighthouse 9.0.0 is officially released. It is expected to be part of stable Chrome starting Chrome 98. One of the new functionalities it offers is support for user flows, making it easier to run Lighthouse for users with a warm cache, and to carry out user interactions.
2021	000000002024-11-16-0000November 16	New version	Browser behavior	Loading speed (LCP)	Field data, Lab data	Chrome 96 fixes a bug in LCP calculation where a previous intended change to ignore images occupying the full viewport had been applied incorrectly to iframes to use the iframe's full viewport rather than the page's full viewport.
2021	000000002024-12-08-0000December 8	Insight	Public diagnostic tool	All	Lab data (with some discussion of field data)	In a blog post on Calibre's blog, Karolina Szczur argues against focusing on the Lighthouse performance score. Arguments include: it’s impossible to describe user experience with one metric, Performance Score is relative to device type and speed trends, Performance Score is prone to variability, Performance Score doesn’t make it easier to talk about speed, Performance Score can be gamed, Performance Score doesn’t matter for SEO, and Performance Score has been overtaken by Core Web Vitals.
2021	000000002024-12-22-0000December 22	Insight	Public diagnostic tool, private diagnostic tool	Loading speed (LCP)	Field data	An article in Smashing Magazine describes in detail their efforts to get their core web vital of LCP to a passing state, including their use of tools such as Google Search Console, web-vitals JS library + CrUX dashboard, PageSpeed Insights, and the CrUX BigQuery dataset. The main methods used include the removal of author images from mobile and the skipping of loading webfonts for users with slow connections and/or data-saving preferences.
2021		Launch	Private diagnostic tool	All	Field data	RUMVision, a product and eponymous small company, launches. RUMVision helps its customers collect real-time web vitals data using their JavaScript, which is built on top of Google Chrome's web-vitals library. RUMVision would later provide anonymized data for other analyses including Google's performance team (for an INP analysis) and the 2024 Web Almanac (providing some specialized sub-part metrics for the analysis of LCP and INP in its performance chapter).
2022	000000002024-01-17-0000January 17	Release	Private diagnostic tool	All	Field data	The page experience section of Google Search Console, that previously only had "page experience on mobile" data, now has separate data for mobile and desktop. This update is made ahead of the rollout (scheduled for February and March 2022) of page experience on desktop as a search signal on desktop.
2022	000000002024-01-18-0000January 18	Insight	Public diagnostic tool, private diagnostic tool	All	Field data, Lab data	An article on web.dev provides a summary of the various diagnostic tools available for monitoring and debugging web vitals.
2022	000000002024-02-01-0000February 1	Release	Metric definition update	Visual stability (CLS)	Field data	Starting with Chrome 98: "The cumulative layout shift score will be recorded the first time a page moves from the foreground to the background, which can fill in for the layout shift values that are not able to be recorded at tab close, the current CLS reporting point." The change would show up in the CruX BigQuery dataset starting with the July 2022 dataset.
2022	000000002024-02-10-0000February 10	Insight	Business impact	All, but concrete examples focused on visual stability (CLS)	Field data	A sponsored post by Clickio on Search Engine Land makes the case for the importance of optimizing Core Web Vitals, mentioning a few cases involving their customers. The post ends with a plug for Clickio's web vitals monitoring product.
2022	000000002024-02-22-0000February 22  – 000000002024-03-03-0000March 3	Release	Search algorithm update	All	Field data	As previously announced on November 4, 2021, Google executes on the rollout of using page experience as a ranking factor on desktop during this period. Page experience on desktop includes all the core web vitals as well as all the other page experience signals on mobile except mobile-friendliness.
2022	000000002024-03-29-0000March 29	Release	Public diagnostic tool	All	Lab data	A blog post on the Chrome developer blog announces the Performance insights panel in Chrome Developer Tools. This panel has some of the functionality of the performance panel with a greater orientation toward producing actionable insights. It features diagnostics for each of the core web vitals.
2022	000000002024-05-06-0000May 6	Proposal	Metric definition update	Interactivity/responsiveness (Interaction to Next Paint (INP))	Field data (but also available in Lab tools)	A blog post on web.dev announces Interaction to Next Paint (INP), a new experimental variant of the responsiveness metric, that may eventually replace First Input Delay (FID).
2022	000000002024-05-12-0000May 12	Insight		All	Field data	A 2-year anniversary conversation about core web vitals is published to Twitter Spaces. The conversation features Annie Sullivan from Chrome, Barry Pollard, Crystal Carter, and Henri Helvetica.
2022	000000002024-05-16-0000May 16	Insight	Metric definition update	Interactivity/responsiveness (Interaction to Next Paint (INP))	Field data	A blog post on web.dev looks at field data across a variety of frameworks for the experimental responsiveness metric titled Interaction to Next Paint (INP), and also describes what the frameworks (React, Next.js, and Angular) are doing to improve INP performance.
2022	000000002024-07-19-0000July 19	Insight	Public diagnostic tool	All	Field data	Documentation on the Chrome User Experience (CrUX) report moves to a new home.
2022	000000002024-08-15-0000August 15	Insight		All	Field data	A blog post on web.dev describes various sources of difference between the official CrUX report and real user monitoring (RUM) data collected by websites using JavaScript.
2022	000000002024-09-01-0000September 1	Data collection	Public dataset	All	Field data	The RUM Archive's rumarchive_page_loads table data starts being available for every day starting on this date. The RUM Archive is an archive managed by Akamai for which data is collected via boomerang.js and mapped using Akamai mPulse. The RUM Insights charts also date back to approximately this time (the earliest start date for the charts appears to be September 6).
2022	000000002024-10-27-0000October 27 – 28	Insight		All	Field data, Lab data	The performance.now() 2022 conference is held in Amsterdam over these two days. Talks from the conference would be released on YouTube in the coming days. Many of the talks would be about web vitals both in the field and in the Lab.
2022	000000002024-11-04-0000November 4		Public diagnostic tool	All	Lab data	pagespeed.web.dev is updated to include the entire Lighthouse report, not just the performance section of the report that it previously made available.
2023	000000002024-01-10-0000January 10	Insight		All	Field data	A blog post on web.dev includes a list of top recommendations to improve core web vitals.
2023	000000002024-02-07-0000February 7	Release	Public diagnostic tool	All	Field data	The CrUX History API is introduced; this API provides access to CrUX data at the level of url and origin for the past 6 months.
2023	000000002024-02-09-0000February 9 (announcement), ~February 16 and March 29 (rollout)	New version	Public diagnostic tool	All	Lab data	Lighthouse 10 is announced; this introduces the scoring change of removing Time to Interactive (TTI) from the score calculation. There are a number of other changes to the available audits and the code now ships with full TypeScript declarations. Lighthouse 10 is shipped to PageSpeed Insights in mid-February and releases as part of Chrome 112 with its release on March 29.
2023	000000002024-03-01-0000March 1	New version	Browser behavior	Loading speed (FCP)	Field data	Chrome 111 includes two bugfixes: (1) a bug that causes FCP to be reported as smaller (better) than it actually is gets fixed, so that the actual FCP now represents the more pessimistic true value, (2) a bug that causes the actual FCP event to sometimes get unnecessarily delayed is fixed. The net effect of these changes is that reported FCP increases (gets worse) but actual FCP decreases (gets better). This would be cited as a reason for the decrease in FCP and LCP pass rates in the CrUX mailing list announcements for April 2023 and May 2023.
2023	000000002024-04-06-0000April 6	New version	Browser behavior	Loading speed (LCP)	Field data	A change is introduced to Chrome to exclude low-entropy images from the pool of candidates for the LCP image. This could affect measured LCP if the low-entropy image would have been the largest of the LCP candidates, and the effect could be an increase or a decrease depending on whether it loads before or after the LCP candidate after the application of the exclusion. The change is applied to all Chrome versions 109 onward, but is largely coupled with Chrome 112 that was released a week ago. This would be cited as a reason for the decrease in LCP pass rates in April 2023 and May 2023.
2023	000000002024-05-10-0000May 10 (announcement) for planned change March 2024	Release	Metric status update	Interactivity/responsiveness (First Input Delay (FID) and Interaction to Next Paint (INP))	Field data	Google announces that Interaction to Next Paint (INP) will replace First Input Delay (FID) in the list of core web vitals, with the replacement planned for March 2024. As of the date of announcement, INP is immediately changed from an "experimental" metric to a "pending" core web vital. The thresholds for FID had made it a very easy metric, with almost all sites passing FID; INP seems to have stricter evaluation criteria, with fewer websites expected to pass it and more optimization work needed to get to passing. On the same day, it is announced that the INP metric is available in the CrUX BigQuery, API, and History API both with and without the experimental prefix.
2023	000000002024-06-05-0000June 5	Insight		All	Field data	Dan Shappir's talk comparing JavaScript frameworks in terms of their performance on real-world data is published, along with a transcript.
2023	000000002024-06-21-0000June 21	New version	Private diagnostic tool	Interaction to Next Paint (INP)	Field data	Google Search Console is updated to include reports on INP. Google had already announced in May that INP would replace FID as a core web vital in March 2024; the addition to Google Search Console gives websites information to help improve their INP about nine months ahead of the switch.
2023	000000002024-08-11-0000August 11	New version	Browser behavior	Loading speed (LCP), visual stability (CLS)	Field data	An experimental change is rolled out to 50% of Chrome users starting on this date, starting with around Chome 116, whereby the priority of the first five sufficiently large images is increased from low to medium from their default of low; images with an explicit 'fetchpriority' are not affected. The experiment also allows for two medium-priority resources at a time to be loading while Chrome's loading scheduler is in “tight mode” (up until the body has been inserted into the document). This is expected to improve both LCP and CLS, with a bigger effect on CLS (as the images are more likely to be downloaded prior to rendering).
2023	000000002024-08-22-0000August 22	New version	Browser behavior	Loading speed (LCP)	Field data	Over a 14-day rollout period starting on this day, Chrome changes the way animated images and videos are considered for LCP. For animated images, the change "counts the time that the first frame of the animation is displayed, if the image becomes an LCP candidate." (In contrast, the prior behavior was to wait for the animation to fully load). This change is expected to improve LCP. For videos, they were previously ignored as LCP candidates unless they had a poster image; now they are included as LCP candidates, with the first video frame being used to determine the time of loading. This change is expected to worsen LCP for pages where the video element would become the LCP element.
2023	000000002024-08-03-0000August 3 (release tag creation), 000000002024-08-09-0000August 9 (update in PageSpeed Insights), 000000002024-08-24-0000August 24 (announcement) 000000002024-10-17-0000October 17 (release in Chrome)	New version	Public diagnostic tool	All	Lab data	Lighthouse 11 is announced and is now the Lighthouse version in use in PageSpeed Insights; this is to become the Lighthouse version shipped with Chrome starting with Chrome 118. There are not many changes to the performance portion; the main changes are the update to the name for INP to remove the "experimental" label on it, and changes involving service workers, resource summary, and legacy navigation. The PageSpeed Insights release notes give August 9 as the date that PageSpeed Insights is updated to Lighthouse 11.
2023	000000002024-10-15-0000October 15	Data collection	Public dataset	All	Field data	The RUM Archive's rumarchive_resources table data starts on this date, with daily data starting on this date. The RUM Archive is an archive managed by Akamai for which data is collected via boomerang.js and mapped using Akamai mPulse.
2023	000000002024-12-23-0000December 23	Insight		Interactivity/responsiveness (INP)	Field data	A blog post by Rick Viscomi uses data from RUMVision, specifically, long animation frames (LoAF), to understand their impact on INP. The analysis uses anonymized data that RUMVision has shared from across its customer base. The analysis identifies the third-party scripts with the worst and best performance in terms of INP oerformance as determined via LoAF script attribution.
2024	000000002024-03-12-0000March 12	Release	Metric status update	Interactivity/responsiveness (First Input Delay (FID) and Interaction to Next Paint (INP))	Field data	Google replaces FID by INP as a core web vital. This follows the plan announced in May 2023.
2024	000000002024-04-09-0000April 9 (announcement, the work likely happened prior)	New version	Consent management platform	Interactivity/responsiveness (INP)	Field data	In the message in the CrUX mailing list announcing the data for March 2024, Barry Pollard mentions that Google has been collaborating with consent management platforms (CMPs) (including OneTrust, Complianz, and Axeptio) to improve their code to yield better, and that this was a likely driver for the INP improvements seen in March. A case study of Google's collaboration with PubConsent, PubTech's CMP product, is also published as a case study on web.dev.
2024	000000002024-04-22-0000April 22 (release tag creation), 000000002024-05-10-0000May 10 (PageSpeed Insights), 000000002024-06-05-0000June 5 (release in Chrome)	New version	Public diagnostic tool	All	Lab data	On May 10, the Lighthouse version used in PageSpeed Insights is updated to Lighthouse 12. Lighthouse 12 becomes the Lighthouse version shipped with Chrome starting with Chrome 126. While this version does not make many changes to the performance portion, there are a bunch of updates to lantern, the underlying code used in Lighthouse to simulate webpage loads.
2024	000000002024-07-17-0000July 17	New version	Browser behavior	Interactivity/responsiveness (INP)	Field data	Chrome 127 introduces two changes to the calculation of INP: (a) Enable EventTimingKeypressAndCompositionInteractionId by default (this is not expected to have much effect in practice as the newly enabled event types tend to be fast), and (b) Enable EventTimingFallbackToModalDialogStart by default (this change is expected to help improve reported INP by covering additional ways a webpage may respond to a user action other than painting something on the page; for instance, it covers the case that the webpage shows a modal (by calling a browser function such as "alert"). This is expected to reduce (i.e., improve) measured INP for websites using modals to respond to user input, without having any effect on the underlying reality. This change is cited as a contributor to INP improvements seen from May 2024 to June 2024 as well as from June 2024 to July 2024.
2024	000000002024-07-30-0000July 30	New version	Front-end ads code	Interactivity/responsiveness (INP)	Field data	Google Publisher Tag (GPT), the front-end ads code used for Google's ad stack on publisher websites, that is used by a large share of publisher websites, makes improvements to its ad yielding behavior. The improvement makes out-of-viewport ad insertion logic yield the thread, while also giving publishers the option to make the ad insertion logic yield the thread even for within-viewport ad insertion (the option is initially called adYield and later becomes threadYield). Initial data suggests a significant improvement in mobile INP pass rates in August 2024, the first full month with the change.
2024	000000002024-08-21-0000August 21	New version	Browser behavior	Interactivity/responsiveness (INP)	Field data	Chrome 128 introduces a change to the calculation of INP; it excludes pointerup events that are the start of a scroll interaction. These events had been included in the INP calculation in contravention of the intended use of INP. The change is expected to improve measured INP in principle by reducing the set of interactions, though the practical impact is expected to be limited because the browser tends to be quite responsive to these kinds of events, so they are unlikely to be the slowest event that would determine INP.
2024	000000002024-11-11-0000November 11	Insight		All	Field data, Lab data	11 chapters of the 2024 Web Almanac (a project of the HTTP Archive) are published, including three that are related to web vitals. The Performance chapter augments Lab data from the HTTP Archive with field data from the Chrome User Experience (CrUX) report with some additional metrics from RUMVision, since the HTTP Archive is limited to Lab data. The chapter goes through metrics in all three categories (loading speed, interactivity/responsiveness, and visual stability) and identifies trends in the field and Lab metrics in each category over the past five years, correlating the field data with the Lab data where both are available. In addition, the CMS and ECommerce chapters also touch on core web vitals and related considerations within their respective contexts.

Meta information on the timeline

What the timeline is still missing

• PageSpeed Insights throttling factor change

See also

• Timeline of Google Search
• Timeline of search engine optimization