It’s fairly common knowledge that the faster your website loads—the better. Faster websites rank higher, convert quicker, and provide a better user experience than their slower counterparts.
Studies have indicated that people prefer a website that loads in less than 3 seconds, and Google has been noticed to penalize websites (by ranking them lower in search engine results) that take longer than 10 seconds to load.
However, page speed can be a tricky thing to fully understand. How do you measure it? How slow is too slow? What causes slow loading times?
Your Experience Doesn’t Equal the User’s Experience
The internet connection in my office is about 250MBs. A few programmers I know in Canada & Columbia have 1GBs in download speed, meaning pretty much every website renders quickly for us. Your office may have faster or slower internet (use a tool like http://www.speedtest.net/ to check, don’t just take your ISP’s work for it).
However, your users could all be living in Montana, which has some of the slowest internet speeds in the world. The point is, website speeds vary drastically between cities, counties, states, and countries.
Just because the website loads quickly on your computer, or slowly on your computer, doesn’t mean that all your website visitors have the same experience.
Using Google Analytics Site Speed
Google Analytics provides a Site Speed report, showing how quickly users are able to see and interact with your content. It helps identify problem pages and tracking your overall page speed over time. It helps solve the issue of “your experience doesn’t equal the user’s experience.”
Here’s an example:
Above is a screenshot of a page load report for one of our clients. You can see the average page load time is 4.93 seconds, but that Safari (in-app) and Samsung Internet both have slower loading times – 7.93 seconds and 13.79 seconds!
While this may seem like a problem, it’s really easy to generate false positives. Like any data, you need to review and think about it.
Respond to the data, don’t react to the data.
How Site Speed Measures Data & How to Read the Data
You can see in the screenshot from the prior section that 282 page views were used as a sample. The total traffic over that period of time (90 days) was 12,292 visitors. This Google Analytics tool usually only tracks about 1 – 2 % of the visitors’ page speed, which means that 98% of the traffic isn’t being measured.
From there—think—what could cause traffic on Safari or Samsung browsers to be slower? Think about the times you’re out and about and trying to find something quickly on your phone, noticing that (1) you have 1 bar and (2) the service sucks. If Google is measuring you trying to get the phone number to the local steakhouse in that circumstance — it’ll give a disproportionately negative return.
Different countries also have drastically different internet speeds. Canada, Europe, and South America have blazing fast internet. America’s internet is alright, while different parts of Asia/South Asia can be worse.
So, we’ll look at the Country breakdown for the same report:
This is important for several reasons. One, India and Russia are not countries that are target markets for this company, and are, frankly, irrelevant to us. How quickly the page loads in those countries won’t affect the bottom line.
I cannot overemphasize how important it is to double check this. Most find that, for their websites, the regions of the world that you’re concerned about are actually performing much better than regions of the world you’re less concerned about. Again, if no one from India or Mexico can (or would ever buy from you) focusing on the load time in those countries is a fool’s errand.
Document Interactive Time
This is a hidden, hidden nugget within Analytics.
Up until this point, I’ve shown you tracking the entire time it takes to load a page. What we’re more concerned about, however, is how long it takes before a user can interact with the page.
For example, let’s say your webpage has a page load time of 8 seconds. That might seem bad, but most of that load time is on a video that is near the bottom of your website. Users won’t even see this video until they scroll down past your priority content.
This is where a metric called “document content load time” comes in.
This is where the page is still loading but the users can begin to interact with the page. In the above example, the Average Page Load Time is 4.93, but the average document content load time is about 1.93 seconds.
This is extremely important because load time is all about user perception.
If you have a website that is interactive AND loads at precisely the same time, say 5 seconds, it will be perceived as much slower than a site that becomes interactive at 2 seconds, and fully loads at 7 seconds. Even though the latter site technically takes 2 seconds longer to load—the user isn’t measuring how long it takes to load each individual picture.
Unfortunately, Google doesn’t have a way to dump that out into an easy-to-read number like “Average Page Load Time is 4.93 seconds & Document Interactive Time is 2 seconds. It breaks it down by day:
I know this is confusing. The purpose up until this point is to show you the tools through Google Analytics that can give you some insight into your actual page speed.
The important part to remember so far is this:
- Page speed is based on perception, not necessarily data
- Users like to see a page load as close to instantly as possible
- Google will likely penalize your site if the page load is longer than 10 seconds
- The page speed for your target market is the only metric that matters
Using the page speed as the be-all-end-all of your page speed metrics won’t help you. The beauty of this report is to make sure:
- Average load time is under 10 seconds
- You’re only targeting countries that matter to you
- You drill down by page
Drill down by Page
This is the most relevant part of the report:
I blurred out the page names for privacy reasons, but there is one page (4th one down) that is almost 100% slower than the rest of the pages on the website. When we check out that page, we can see that the blogging author uploaded images without properly sizing them.
Generally, there are two big issues that slow down websites:
- Images are too large
- Website code isn’t optimized properly
- Server sucks
As mentioned above, the large images slowed down an important part of the website, hurting the user experience.
I Want My Website to Load Instantly, Damn It!
First, you’re going to be limited by your infrastructure. If you only want to pay $15 a month for a shared HostGator server… you’re not going to be able to have more than a few lower quality images on your page before instant goes out the window.
Secondly, if you want to hack together a few WordPress themes without spending a lot of time on code optimization… your instant loading speed goes out the window.
Finally, if you want to have big, beautiful 4k images on your website (and don’t want to pay for a CDN or other service). You guessed it—instant goes out the window.
Web users have been spoiled with expensive servers, Content Delivery Networks, and other “under the hood” type improvements in industrial level servers. Even a smaller insurance company with only a few million dollars a year in premiums can rack up $5,000+ in monthly hosting costs to get the “instant” feel for their website.
Okay, So I Guess I Don’t Want Instant. What Do I Do?
Contact your local web developer. Chances are they’ll figure out a good plan to optimize your website across devices.
Seriously, it really is case-by-case. The priorities of your website will depend on your target market, your location, and who you are as a person. Just understand that your website speed will greatly be based on your infrastructure, coding decisions, and content decisions on a website.