Last week I did a post on how the latest version of Google PageSpeed Insights works.
That got me to thinking about what the speeds of the arguably most important New England tourism websites - the 6 official State tourist-board sites - are like, and how difficult or easy it might be to improve these.
So today I ran the mobile version - since most tourism websites are now getting 50%+ of their traffic from mobile devices, and because it's a tougher test - of the Google PageSpeed Insights tool on these 6 websites, and recorded the most basic performance data for each, along with Google's #1 recommended fix - the change to website coding with the most potential for shortening loading time.
Here are the results:
As you can see, unlike Google.com which scores 96 on this test, none of these State tourism websites are exactly balls of fire. (In this test Google considers scores of 0 to 49 to be "Slow", 50 to 89 "Moderate".)
VisitMaine.com got the only Moderate score, while all the others got labeled "Slow".
The "Time to First Meaningful Paint" numbers - the times it took the sites to produce recognizable content - are fairly tightly clustered: between 3.1 (ME) and 5.1 (RI) seconds.
But the "Time to Interactive" numbers - the times it took for the pages to become fully functional - working hyperlinks, etc. - are much more diverse, ranging from 8.0 (VT) to 21.8 (RI) seconds.
What's causing these delays in interactivity? In 4 out of 6 cases (ME, MA, RI, CT), the #1 actionable fix the tool prescribes involves images - offscreen images that slow down page rendering, images in traditional formats that are slower than those in "next-gen" formats, or images that have been inefficiently encoded.
The PageSpeed Insights tool thinks the developers of MassVacation.com can reduce the loading time of their home page by 3.9 seconds by encoding images more efficiently!
Some caveats on these results:
Note that I've indicated the time and place of these tests. All were done in a span of about 20 minutes this morning via my Verizon Internet connection in Boston. Repeating the tests later today or from a different location will almost certainly change the reported loading times, possibly significantly.
So the important takeaway from these tests is: the recommended fixes. Even if the measured times change a lot, the recommended fixes likely will not.
And all these recommendations are actionable.