Formula One drivers and websites don’t have a lot in common, but one truth they do share is: faster is better. But how fast does your website need to be? And what’s the impact on the bottom line? Today, we look at these questions and try to find some answers with our deep dive into website speed.
How fast is your site?
First, let’s cover the basics. When we talk about website speed, we’re really talking about load time which is how quickly you can get the information on your server rendered correctly on a user’s device. Load times are measured in seconds and milliseconds.
Before you can improve your site’s load time, you need to establish how fast it actually is. Fortunately there are dozens of tools out there that will help with this. Pingdom.com is a great way to check your site quickly and easily. You can enter any URL you want and it’ll break it out for you by various elements, letting you know what takes a long time to load and what loads quickly.
Google Developers also offer a website tool that includes information about where your site speed is good or bad, broken out between mobile and desktop.
Using these tools you can find out what your site’s load time is, which takes us to our next step of finding out how fast it should be.
How fast does your site need to be?
Well, it depends on what sort of site you have and what page you’re loading. A good example of how the importance of load speeds can change is an ecommerce site. Generally for ecommerce the pages users land on first need to load much faster than the final pages of the buying process. That’s because visitors willingness to stay on a site increases along with their investment in that site. Simply, the longer someone is on your site (say, completing a purchase), the more willing they are to stay, and less likely to be affected by load speed.
So how fast should your pages be?
The best way to determine how fast your page should be is to look generally at what other sites are doing. After all, it is a users experience across the internet as a whole that sets their expectations for what is fast and slow. Moz pulled together some data on this and this is what they found:
- If your site loads in 5 seconds it is faster than approximately 25% of the web
- If your site loads in 2.9 seconds it is faster than approximately 50% of the web
- If your site loads in 1.7 seconds it is faster than approximately 75% of the web
- If your site loads in 0.8 seconds it is faster than approximately 94% of the web
If you’re aiming to be in the top 25% of the internet load times, your pages need to load in 1.7 seconds or less. For your highly trafficked pages, any landing pages, and any internal pages that attract an unusual amount of inbound traffic, we’d recommend aiming to be in the top 10% of pages on the internet, with a load time of about 1 second. This correlates with a finding that half of all web users expect a site to load in 2 seconds or less.
Of course, these guidelines should not be taken as law because a number of other factors play into the importance of load times. If you have really qualified traffic landing on your site, for example, an internal page is ranking well for a recurring long tail keyword search, load times are less important because users are pre-qualified – they’ve specifically sought out your page. They’ll likely wait for 0.5 seconds or more for it to load. On the other hand, site speed is absolutely critical for traffic where the user is poised to click away. This would include highly unqualified traffic, like traffic from PPC ads (especially if you’re bidding on keywords).
Which brings us to the question of: what’s it going to cost you?
The million dollar question: how much will a slow load time cost you?
As we have seen that depends on a number of factors (such as quality of visitor), but there is a pretty clear relationship between load time and abandonment. This makes perfect ‘cents’ – the more time users spend waiting, the more likely they are to leave.
This is especially true on mobile networks, where load times are longer due to lower quality network connection.
It’s hard to put a dollar value on this relationship but Amazon tried back in 2007, where, with A/B testing, they found that every 100 millisecond delay resulted in losing 1% of sales.
Another study in 2013 found that splitting load times in half from 15 to 7 seconds, and then again from 7 to 4, and so on, improved conversions, but had diminishing returns.
Obviously, the dollar value of these statistics is going to be different for everyone, but if you think about other techniques to optimize websites where a gain of even half a percent is a huge success, it starts to put the importance of speed into perspective.
Unfortunately, there’s no absolute figure on how fast you should be. Really, speed is entirely relative. That is, it doesn’t matter how fast you are so long as you’re fast enough for your user. For example, a blog about different types of duck feather duvets, or something else equally niche, might tolerate a slower load time due to more qualified users seeking them out. Likewise, a site relying heavily on keyword-driven PPC instead of long tail SEO would likely need a faster load time for its less qualified users.
What we do know is that over half of users require a site to load in two seconds or less or they’ll leave, and both Walmart and Amazon have reported that faster load times have had a direct impact on their conversion rates.
Our suggestion is to focus on getting all your load times to less than two seconds. Then, fold site speed into your normal optimization timelines. Now if you’ll excuse us, we’re off to watch some Formula One.