Load Speed

According to the Site Performance feature in Google Webmaster Tools, your pages load reeeeealy slow but, other external tools or monitoring services tell a different story.

What should you believe?

First, it's important to understand the differences between these tools, the data they capture and how it's measured.

Page Speed evaluates the performance of a specific web page and individual elements in the browser. As a result, this type of testing may not accurately reflect latency experienced by users. Page Speed is for testing and improving speed for individual pages.

Tools like webpagetest.org and monitoring services often test latency for a specific URL at various times of day and locations around the world. As a result, these kinds of tests may not reflect latency as perceived by users in the region the site targets.

Google Webmaster Tools Site Performance data is collected from actual Google Toolbar users in the same geographic region as the target audience of the site. This data can be measured in several ways. One being, time between when the user clicks on a link "until just before that document’s body.onload() handler is called." If for example, if a user clicks on a link, is then redirected and then redirected again, that delay should be recorded and reflected in Google Webmaster Tools Site Performance data. These are the kinds of delays that impact users and Googlebot and that are totally missing from other tools including analytics.

Speed doesn't currently have a major impact on rankings but, slow pages deter users and hamper crawl efficiency. Crawl efficiency can be a major factor for pages with lower PageRank because " the number of pages Google crawls is roughly proportional to PageRank".

Matt Cutts first mentioned speed publicly, as a potential ranking signal in November 2009 but, speed has always been important at Google. Google's homepage for example, is intentionally sparse so that it loads quickly. Larry Page recently said he wants to see pages "flip" online. Clearly the concept of speed and it's importance at Google is nothing new. Robert Miller, actually conducted the first research in this area over 40 years ago. According to Miller, when a computer takes more than one tenth of a second to load, the user feels less and less in control. When Google and Bing conducted their own tests in 2008 the results were similar to what Miller had predicted. Bing experienced a 1.8% reduction in queries when slowed by 2.0 seconds and Google experienced a 0.59% reduction in queries when slowed by 400 milliseconds. Bottom line, fast pages help marketers because users are far less likely abandon fast loading sites. Users expect pages to load quickly!

Page Speed was introduced as a ranking signal for English queries executed via Google.com nearly a month ago. It's defined as "the total time from the moment the user clicks on a link to your page until the time the entire page is loaded and displayed in a browser." In their first major update to Webmaster Guidelines in over a year, Google recommends webmasters monitor site performance and optimize page load times on a regular basis.

Because Page Speed data is missing from Analytics and other tools, it's best to use the Site Performance in Google Webmaster Tools for regular monitoring and optimization. To view Site Performance data in Google Webmaster Tools, you'll need to add and verify your site first. In Google Webmaster Tools, Site Performance data is processed in aggregate and without personally identifiable information. "Example" URLs are derived from actual Google Toolbar user queries and as a result query parameters are removed. "Suggestions" are based on URLs crawled by Google and not truncated. "Average" site load times are weighted on traffic. Site Performance data accuracy is based on the total number of data points, which can range from 1-100 data points (low) up to 1,000+ data points (high). As mentioned earlier, this data comes from Google Toolbar users with the PageRank feature enabled. While Google hasn't provided Toolbar PageRank user demographic information, this data seems fairly reliable. If anything, it would seem that Toolbar PageRank bias would point to more savvy users and that as a result Google Webmaster Tools Site Performance data might be faster than actual.

During our "Speed" session at SMX, Vanessa Fox and Maile Ohye (Google) seemed to agree that less than 4.0 seconds was a good rule of thumb but still slow. According to Google AdWords, "the threshold for a 'slow-loading' landing page is the regional average plus three seconds." No matter how you slice it, this concept is fairly complex. Page Speed involves code, imagery, assets and a lot more not to mention user perceptions about time. For example, the user's internet connection (Google Fiber), their browser (Chrome), device(Nexus One), its processor (Snapdragon), the site's host and other issues all impact user perceptions about speed. Don't worry though, according to most expert estimates, 80% to 90% of page response time is spent on the front-end. At the end of the day, relevance is critical but speed follows quickly behind.

Speed Resources:

Matt Cutts provided some interesting details about where the industry is headed, last week at PubCon.

During the "Interactive Site Review" session, Matt suggested investigating the history of each domain name you own or plan to purchase. He suggested avoiding domains with a shady history and dumping domains that appear to have been burned in the past. To investigate the history of a domain, Matt suggests Archive.org. Matt said, blocking Archive.org via robots.txt is a great indication of spam when already suspected.

Matt mentioned speed several times. During the "Interactive Site Review" Matt said that webmasters need to pay more attention to speed. He pointed out that landing page load time factors into AdWords Quality Score and said speed will be a big trend in 2010. During Matt's "State of the Index" presentation, he pointed out Google's tools for measuring page speed and even mentioned webpagetest.org a third party tool. According to Matt, Google is considering factoring page load speed into rankings. Matt said, that Larry Page wants pages to flip for users on the internet. He illustrated this point with Google Reader's reduction of pages from 2mb to 185kb. Nothing official yet but, something to keep an eye on for sure!

During Q&A for "The Search Engine Smackdown" session Matt explained Caffeine as being like a car with a new engine but not an algorithm change. Matt said, Caffeine will help Google index in seconds and that it should be active within a few weeks on one data center. That said, Caffeine won't roll out fully until after the holidays. Matt pointed out that Google is built for load balancing and for that reason isolating individual IPs for Caffeine testing access is difficult. Matt also mentioned that AJAX SERPs and Caffeine aren't related but that Google will continue testing AJAX SERPs.