Do you think page load speed is your target performance KPI? Think again. I want to introduce you to 2 metrics you may not have heard of, but influence both SEO and site conversion.
TTFB: Time to First Byte
This metric measures the time it takes for a browser to receive the first byte from a web server when requesting a given URL, an indicator of back end performance.
TTI: Time to Interact
A new metric coined by Radware, time to interact refers to the point which a page displays its primary interactive (think clickable) content, rather than full page load.
Why do these metrics matter?
TTFB for SEO
Research by Moz and Zoompf found a correlation between speedier time to first byte and higher search rankings. Yes, correlation does not equal causation -- but consider the same study found that the time to load the full HTML did not correlate with better search rankings. The researchers surmise the reason TTFB may factor in search ranking is Google crawlers can easily measure TTFB, whereas measuring the full page load is better handled by a full browser. Also noted is load times can vary across browsers due to their own rendering capabilities aside from how well a website is optimized for performance, thus TTFB may be a more reliable indicator of positive user experience to Google.
These findings suggest it's really about how quickly servers and back end infrastructure can deliver web content, not front-end website performance that really matters when it comes to SEO advantage.
What influences TTFB?
- Network latency between a visitor and the server.
- How heavily loaded the web server is.
- How quickly the website's back end can generate the content.
Tuning the back end is just as important as optimizing your HTML. Moz community member Billy Hoffman of Zoompf has written an excellent tutorial on how to determine your TTFB with tips on how to tune your back end.
TTI for CRO
Conversion rates increase when bounce rates decrease, if for nothing else than more customers stay on your site. Time to Interact is a key indication of your site's ability to provide a good user experience, and reduces perceived page load time, serving as a good measure of real-user performance.
The median TTI for top 500 retailers is 4.9 seconds -- well above the perceived page load threshold of 2 seconds. A study by Akamai found in 2006 that 25% of visitors would abandon a site that took longer than a perceived 4 seconds to load, a threshold cut in half by 2009, with 50% of visitors saying they bail at the 2 second mark. Only 8% of top 100 sites in the Radware study achieved a sub-2-second TTI.
In 2010, Strangeloop Networks (now part of Radware) found "a site that loads in 3 seconds can expect 22% fewer page views, a 50% higher bounce rate, and 22% fewer conversions than a site that loads in 1 second, while a site that loads in 5 seconds experiences 35% fewer page views, a 105% higher bounce rate, and 38% fewer conversions."
Tuning the front-end for faster TTI
As hinted by the Netflix, iHerb and LivingSocial examples, best practice is to optimize your pages to load important, interactive content first, and defer the less important.
- Load what's above the fold first
- Load interactive elements quickly
- Defer loading scripts that support user actions that aren't useful until after the page has fully loaded (e.g. drag-and-drop or comparison functionality)
- Defer loading third party scripts like analytics, social sharing
- Use AJAX to load additional page content when the user requests it, rather than loading it up front
- Leverage a CDN (content delivery network)
- Compress text and images, minify code and resize or reformat media
- Use keepalives
Your site performance is a mix of front-end and back-end optimization, but simply measuring and tuning for page load speed misses the opportunity to improve Time to First Byte and Time to Interaction, both of which have real impact on search bots and site visitors.
Stay "tuned" for this week's infographic, which will share more highlights from Radware's State of the Union for Ecommerce Page Speed & Web Performance [Summer 2013]