Don’t Listen to Everything Google Tells You

In the last post on Google PageSpeed Insights, we discussed the problems with that tool and how the advice it gives is interpretive and not always beneficial from either a user experience standpoint or a search engine ranking standpoint. In this article we look at a newer tool by Google, TestMySite. TestMySite purportedly analyzes the speed of your website from a mobile device.

Google is very focused on mobile search these days. In fact, because mobile search now exceeds desktop search, they are actually ranking your site more on its mobile characteristics than on its desktop ones. So there is no question that your site needs to be mobile optimized both in terms of user experience and load speed. Google’s tools are supposed to help you with this. Only they don’t!

PageSpeed Insights gives our mobile site a grade of “Good” and an “87/100.” TestMySite, on the other hand, says that our site’s “Loading time on 3G” is 8 seconds and gives it a “Fair” grade and an estimated visitor loss of 28%. In other words, the TestMySite tool thinks our site is slow and that we will lose almost 1/3 of the visitors to our site because of this.

A key thing to note in these results is “Loading time on 3G.” It is measuring the speed of the website via a 3G mobile connection versus a faster 4G/LTE connection. While on a worldwide basis, 3G still has a large number of users, in the U.S., mobile users connect via LTE about 93% of the time. Therefore, the TestMySite tool is only accurate for the smallest percentage of potential visitors, at least for our site and for those of most of our customers because few of our customers or their customer are visiting from outside the U.S. And if they are visiting from elsewhere in the world, it’s arguable that our target audiences are going to be in developed countries where the use of LTE is as high, and sometimes higher, than in the U.S.

TestMySite is actually using the WebPageTest engine for calculating the speed of your site. If you try this tool, you can show the “Advanced” options where you can change the testing speed to LTE instead of 3G. On an LTE connection, our site loads in around 3 seconds.

So, both of Google’s primary tools for analyzing your site speed are biased and not entirely relevant to the two most important reasons to have a fast site – search engine ranking and user experience.

Then how should you measure these factors? A much better tool for a real-world evaluation is Pingdom. Like all the other site speed tools, Pingdom “grades” your site and the grade is typically not that favorable. However, it also gives the ACTUAL load time of your site, a much better indication of the true performance of your site. Pingdom, like PageSpeed, also provides a lot of information on how you can improve the speed of your site but it’s in a more attractive and easier-to-read format. Measuring our site on Pingdom, we see that the home page loads in about 1.5 seconds. That’s pretty fast and sufficient not only for decent ranking in the search engines, but it also unlikely to make users feel like they have to wait for pages to load.

Why your site loads fast or slow is based on a complex set of variables. I’m not opening that can of worms at this point, though I may in a future post. However, before even embarking on a quest to speed up your site, make sure that the information sources you are using to analyze your site are giving you a less biased view. If you just listen to Google’s recommendations, you are likely to spend a lot of time (and possibly money) trying to fix something that may not be broken.