One thing that I want to make sure you understand before we get into the various testing tools is how you should test. First, when testing, don’t just test your home page. You’ll want to run tests on the various page types on your site. So test your home page, but for eCommerce also look at testing category pages, subcategory pages, and item pages. Most of those will cover the bulk of your site but any page that runs differently is worth testing. Just make sure to test the page or at least page type that you want to rank and perform well.
Second, when testing with tools that give actual load times, you may need to run the test multiple times. Each test, whether from your browser, or an online service, can be subject to delays across the internet. So if you run a test and get a horrible load time, try retesting. If you keep getting bad times, then I’m afraid it’s real. But if not, then it could be a temporary delay. We’ve had sites where we run a timed test, it looks bad, then we optimize and we know it’s better and the first test shows a worse time. But you look at the info and see that it should run better (better images, files, etc) and a second test will usually show it. What’s key to know is that your fastest test will be as fast as it can go, but it can always get slowed down by things between the testing site and your site (or any 3rd party). So if things seem out of line, try another test in a bit and see if you get better numbers.
Now how should you test speed. Well there are a lot of tools out there and some are better than others, or some test different things than others. Unfortunately there is no “perfect” tool, but many of them overlap so when you target one, you often are targeting all the others too. The nice thing is that all of these are free. 🙂
Google Page Speed Insights – https://developers.google.com/speed/pagespeed/insights/
Obviously this is from Google and so this is the one most users see, and use. The current version of Page Speed Insights does a good job of testing both mobile and desktop load. Most of what it reports are best practices in regards to load time. It won’t report actual load times, except if you have a particularly horrible server response time. We have seen some cases where it reports incorrectly some scripts causing delays in load, when in reality they’re loading at the end of the page, but overall it does a great job and is a good place to start. Also, in regard to images, you’ll want to try and make “google happy” but also look at the actual image sizes. There are ways to get the image off Google’s list, but the file doesn’t shrink much (or can grow). So obviously two files off the list are good, but if one has actually reduced by 50% and the other by 5%, then for users there’s still a difference.
BONUS: – Drag this link to your shortcut bar and you can click it to test any page in Page Speed insights – Insights
Pingdom – https://tools.pingdom.com/
Pingdom has been around for a long time and they actually have some good tools to continually monitor your site. Pingdom tests, as many of these do, things that overlap with Page Speed Insights. The nice thing is they also give you a nice waterfall graph along with time tests (and of course a pretty letter grade). The benefit of the time tests and waterfall is if you are seeing slow load times, even when doing things “correct”, you can nail down the culprit services or files.
WebPageTest – https://www.webpagetest.org/
WebPageTest is also similar to Pingdom. They give a letter grade, show waterfalls, etc. They are a bit more aggressive on images and so will downgrade the score based on images a bit more. They do provide a lot more details than Pingdom and Google if you need to delve into the details.
Y Slow – http://yslow.org/
Y Slow is different than other tests mostly because it’s a browser addon. So you run it in your browser and test. It does test a lot of the same things and so honestly I prefer one of the other services because you can test different locations (for Pingdom and Web Page Test), but it is another option.
GT Metrix – https://gtmetrix.com/
Now GT Metrix is interesting, and debatable if you should use it. The nice thing is that it will test using Y Slow, Google Page Speed insights, and do a timed test. It also can keep and compare historical tests. So from that standpoint it’s pretty nice. The downside is that it uses the Google Page Speed Insights API which unfortunately hasn’t been updated in awhile. So if you test on Google’s site and compare the score vs GT Metrix, they will differ. Most of the concepts are the same, but a few have changed in what Google is looking for, so those can throw off the numbers. Also, when doing comparisons, we’ve some cases where the calculated numbers vary slightly and so sometimes even if nothing has changed in one area, it will show one test with a higher score. Not always, but it does happen. But the ability to save tests and compare before and after is very nice to have.
Google Chrome (and other browsers)
Just like Y Slow, most modern browsers have tools to look at waterfall graphs and load times. They won’t break out the details nicely for you but they can make a quick way to check if you’re seeing a performance issue. You have to get into the developer tools (often F12) and then from there either go to Network or Performance Tabs. They can be a bit geeky and overwhelming so if you don’t want to get into the nitty gritty details, you may want to skip these.
Frequently Asked Questions
Is Google Page Speed Insights the perfect test?
Simple answer, no. It’s got great guidelines but does have it’s imperfections. We’ve seen a number of cases where it may be ok with a heavier image file, but then a lighter one it won’t be (without further optimization). It also doesn’t test timing and connectivity issues with various 3rd parties which can bring up problems.
I tested my site home page, isn’t that enough?
Well it is if visitors only go to your home page and never visit another page, and you don’t want search engines sending them anywhere else. Otherwise you should try to test at a minimum your home page, section page and item pages. But if you have other pages of value with different layouts (section vs subsection, custom layouts, etc) you may want to test them as well. Also, try testing some of your largest pages to see how they fare.
Do you have a question? Feel free to ask one of our wizards via the form below…