The Impacts of Page
The load time of a web page has become a very hot topic here at Arcane. A slow webpage can negatively impact our clients’ customers and consequently, their bottom line.
Page Load depends on several factors and there are many tools for testing the time it takes for a web page to load. Google Lighthouse audits webpages on their performance, accessibility, and search engine optimization. Google Test My Site tests a webpage and provides insights around the potential missed revenue due to the page’s performance. The main problem is that different tools use different base metrics for their scoring. At Arcane, we needed to unify the metrics and speak a common language surrounding page performance, both internally and with our clients.
When we refer to Page Load, we needed one metric to rule them all. Time to Interactive is our base metric — the time it takes for a web page to be ready for the user to interact with. Some testing tools use the Time to First Meaningful Paint — the time it takes for a web page’s primary content to appear on the screen.
Page Load Benchmarks
Let’s talk benchmarks! We ran Google Lighthouse tests on the top 2,000 websites (based on traffic). From the 2,000 domains, we selected three random pages and performed 25 Lighthouse tests on each page, resulting in 150,000 performance scores. The data we gathered showed us that the average Time to Interactive was 10.2 seconds — Google recommends 3.0 seconds. We set our goal at 3.5 seconds — more on that later.
How does your website stack up against the internet average? How do your competitors’ websites compare? Does your website out-perform your competition?
Using the data we collected, we spent time testing and validating Google Test My Site logic to provide comparable insights using Time to Interactive instead of Time to First Meaningful Paint metrics, keeping our measurements consistent. We collected our own data on these metrics for two reasons:
- Most data of this nature is out of date (2-3 years old)
- We needed it to validate the translation of Time to First Meaningful Paint to Time to Interactive and normalize the tools used for testing.
Which factors affect page load? When searching for insights into the impact a slow Page Load, Google Test My Site can provide some pretty shocking projections. There are several contributing factors that can affect the time and in turn, the amount of traffic lost.
Time to Interactive
A key piece of data is the amount of page traffic that is lost for every second a page takes in its Time to Interactive. The internet average loses 18% of the traffic going to the page. Arcane’s goal of 3.5 seconds would lose 6% of the page’s visitors. The percentage lost is more significant when traffic is directed to the page through paid methods — ads get the user to the page, but a substantial amount of users are lost because of the time it takes for them to view and interact with that page.
Hosting, Servers, & Programming
Where the page is hosted, the physical server, its specifications and configuration can impact how long it takes for a browser to communicate with it. The programming that the page is created with and its efficiency — as well as the page’s functional components — can also add substantial time.
Assets & Third-Party Integrations
Assets including images, videos, and animations usually have the largest impact on the time it takes for a page to load. Optimizing or eliminating these assets is paramount to a high performing page. Third-party integrations can be problematic as they are out of the control of those that maintain the page.
The last factor that influences Page Load is the speed of the connection of the user. The majority of internet traffic uses mobile devices and thus mobile connection speeds — which can be considerably slower than desktop speeds. Additionally, data plan limitations can be barriers for some users.
The Financial Cost of a Slow Page Load
By now, you should be wondering what the cost of an average page load might be! As a simple example, let’s take Company X and their web page that loads in 10.5 seconds (the internet average). The estimated monthly revenue generated by this page is $30,000. The page has an average of 10,000 visitors per month and has a conversion rate of 3%. That equates to 300 conversions per month and their average order value is $100.
However, a 10.5-second page load means that the page loses 18% of the traffic that has intent. 18% of 10,000 visitors is 1,800 visitors lost waiting for the page to load. If we plug those 1,800 visitors into the same formula for estimating revenue, we determine that Company X misses 54 conversions per month — which equates to $5,400 in lost revenue!
Finding the Root Cause of Page Load Issues
How can you diagnose where problems lie with a web page? We performed 220,000 Google Lighthouse reports and plotted the Time to First Meaningful Paint and Time to Interactive. Comparing these metrics provides an indication of the root cause of a performance issue.
The overall cause of poor performance is the web page’s payload — the total size of the page measured in kilobytes. Web pages have been (on average) getting larger and larger due to internet speeds getting faster and faster.
Setting Arcane’s Page Load Goal
Remember what Google’s recommendation is for the Time to Interactive? 3 seconds.
Let’s look at a landing page for Company X with a Page Load Time of 4 seconds. The functionality of the page is limited to a lead-generation form accompanied by a few sections of marketing content. Here’s a breakdown of the page load:
- 0.4 seconds for the server response
- 0.4 seconds for a Google Font
- 0.7 seconds for Google Recaptcha (used as an anti-spam mechanism on the form)
- 1.5 seconds for Google Tag Manager (used to track page interactions).
That’s 3 seconds! We haven’t even loaded the code and visual assets yet — which load in 1 second, for a total of 4 seconds for the Time to Interactive.
For Arcane, 3 seconds isn’t an achievable goal when key Google technologies occupy that time and there’s nothing left for code and design. We’ve set our own goals for clients and test them with Google Lighthouse. We achieve this by having a Performance score over 90/100 and Accessibility and SEO scores of 100/100. We can only achieve an overall score over 90/100 if the page loads in approximately 3.5 seconds.
One last thing to remember is that the “rules” these testing mechanisms use change, as do the weighting of certain metrics. This can result in better or worse performance scores, depending on the change. This is why continual testing and optimization are integral to keeping your webpages at their peak performance.
For a PDF version of the infographic, click here.