In this presentation we review website performance and how it impacts ecommerce sites, specifically those using Magento. We define website performance and give tips on how to speed up websites.
6. Amazon lost 1% in
sales for every
additional 100ms
in page load time.
Walmart found
page speeds of 1-
2 seconds are
twice as likely to
convert as speeds
of 3-4 seconds.
7. How do you know your site is slow?
It feels slow
Use a stopwatch
Synthetic measures
• Pingdom,
Webpagetest, Gomez
Real user monitoring
• New Relic, section.io
✓
✓
✗
✗
8. Metrics to Measure Performance
Visually Complete Fully Loaded
Page Load TimeStart Render Time
9. < 1 second Start Render Time
2-3 seconds Visually
Complete
Good website performance
looks like…
10. • Over 50% of mobile users
abandon a site that takes
>3 seconds to load
• On Black Friday 2016,
mobile accounted for 55%
of page visits and 36% of
total sales.
Don’t forget about mobile!
11. Little Wins to Improve Magento Site Performance
Manage 3rd
Party
JavaScript
on site
Image
optimization
Remove
unnecessary
CSS code
13. • Caching
• Use a Content Delivery Network
• Varnish Cache recommended for Magento
• Upgrade to HTTP/2
• Faster and bonus of improved security
• Choose a good hosting provider
• Modern hardware, HTTP/2 capabilities, Magento-
specific considerations
BIGWINS
15. Caching: Storing parts of your page on a caching
server or CDN so requests don’t need to go back to
your website server.
Visitor #1
Cache
(Empty at
this point)
Website server
16. Caching: Storing parts of your page on a caching
server or CDN so requests don’t need to go back to
your website server.
Visitor #1 Cache Website server
20. Content Delivery Network: Servers all over the globe
bring content closer to visitors.
Website server
CDN caches
CDN caches
Visitor browsers
Visitor browsers
Visitor browsers
21. Case Study: Adore Beauty
• Online beauty store
• Used in-depth data and New Relic to pinpoint performance issues
Improvement
22. Results
• Viewed 9.4% more product pages
• Reached checkout page 15.5%
more often
• Reached checkout success page
16.5% more often
• Resulted in consistently higher
revenue
23. “We didn’t believe the results at first, so we asked
section.io to leave the A/B Testing on for a
number of weeks and each week the results
came back the same.
Their technical expertise is impressive, but more
importantly we are impressed with having more
money in the bank!”
- Kate Morris
CEO and Founder Adore Beauty
24. What we learned:
• What website performance means
• How performance impacts revenue
• Metrics and programs to measure your performance
• Steps to take to improve performance
• How caching and Content Delivery Networks impact performance
25. Get Started:
• Get the section.io Instant Global Cache Extension for Magento 2 at:
get.section.io/Magento2
• Contact section.io: sales@section.io, @section.io
• Download our guide to improving performance and scalability for
Magento: get.section.io/MagentoGuide
Editor's Notes
Website performance as an umbrella phrase can mean many different things – if your business is focused on upping conversion rate, a website that converts poorly could be considered to be performing badly for your business needs.
Website performance as an umbrella phrase can mean many different things – if your business is focused on upping conversion rate, a website that converts poorly could be considered to be performing badly for your business needs.
However, technical definition of “website performance” focuses on one element of websites, speed. Website speed tent.
impacts all the other elements of performance - bounce rate, conversion rate, cart size, and even scalability because the faster pages load, the more visitors you can serve at once.
It’s not just user experience that takes a hit – many studies have proved the impact website speed has on sales.
Was on slide: Studies by major ecommerce sites have demonstrated impact of speed on revenue:
Once you know what metrics to collect, here are some of the tools you can use.
There are two main categories to website performance monitoring tools - synthetic, which simulates real visitor behavior, and Real User Monitoring, which captures information on the actual visitors using your site.
Synthetic monitoring cannot account for the complexities of how actual visitors see and browse around your site, but for basic website performance tests it gives good metrics. Real user monitoring or RUM is an excellent tool if you can get it, but requires your website to have enough actual visitors to test and often requires buying a specific RUM platform.
Synthetic platforms we recommend include WebPageTest and Pingdom. WebPageTest will provide you with the metrics we described in the last slide, and a “waterfall test” that shows what elements are taking the most time to load and which may be delaying other elements from loading.
New Relic is an excellent RUM tool that provides advanced monitoring, including server-side statistics and scalability monitoring if you are looking for that level of detail.
Once you know what metrics to collect, here are some of the tools you can use.
There are two main categories to website performance monitoring tools - synthetic, which simulates real visitor behavior, and Real User Monitoring, which captures information on the actual visitors using your site.
Synthetic monitoring cannot account for the complexities of how actual visitors see and browse around your site, but for basic website performance tests it gives good metrics. Real user monitoring or RUM is an excellent tool if you can get it, but requires your website to have enough actual visitors to test and often requires buying a specific RUM platform.
Synthetic platforms we recommend include WebPageTest and Pingdom. WebPageTest will provide you with the metrics we described in the last slide, and a “waterfall test” that shows what elements are taking the most time to load and which may be delaying other elements from loading.
New Relic is an excellent RUM tool that provides advanced monitoring, including server-side statistics and scalability monitoring if you are looking for that level of detail.
Some recommended goals for your website are
-Under 200 ms TTFB
-Under 1 second start render
-Under 2 seconds visually complete
Don’t forget about mobile load times – these can be impacted by network speed and mobile OS so it’s important your site itself is as fast as possible. The tips we will lay out next can improve both your desktop and mobile sites.
Where should you start when improving performance? Here are some easy steps anyone can take
-Reduce image sizes and optimize them for web. Even if an image appears smaller due to HTML resizing, the browser will have to load the full image size. Avoid this by sizing images for the max width they will appear on your site. You can also save at a lower quality for many online uses. There are several free online tools that will both resize and compress images for free.
-3rd party Plugins such as Google Analytics and other marketing tools can dramatically slow down your site. Regularly review your code and view Waterfall tests to see what javascript plugins are setup on your site and how much they are slowing it down.
-Smaller CSS files will load more quickly, so review your CSS code regularly and use a free minifying tool if needed.
For real performance improvements, you’ll have to implement some of these tips. Some will require a developer to setup however the speed improvements are worth the time and cost of implementation.
-Go through this list briefly
For real performance improvements, you’ll have to implement some of these tips. Some will require a developer to setup however the speed improvements are worth the time and cost of implementation.
-Go through this list briefly
Two major types of caching –
Definitions of browser/server caching and how they speed up sites so much
Cache hit vs cache miss – aim for highest percentage of cache hit
Aim to cache as much content as possible including the HTML document – first piece of information server gives browser to begin speeding up site. This improves the TTFB.
Two major types of caching –
Definitions of browser/server caching and how they speed up sites so much
Cache hit vs cache miss – aim for highest percentage of cache hit
Aim to cache as much content as possible including the HTML document – first piece of information server gives browser to begin speeding up site. This improves the TTFB.
Two major types of caching –
Definitions of browser/server caching and how they speed up sites so much
Cache hit vs cache miss – aim for highest percentage of cache hit
Aim to cache as much content as possible including the HTML document – first piece of information server gives browser to begin speeding up site. This improves the TTFB.
Two major types of caching –
Definitions of browser/server caching and how they speed up sites so much
Cache hit vs cache miss – aim for highest percentage of cache hit
Aim to cache as much content as possible including the HTML document – first piece of information server gives browser to begin speeding up site. This improves the TTFB.
This could be part of previous slide on caching or just one bullet point if we can fit
CDNs have different strengths – as a Magento merchant you should look for those that use Varnish Cache
We performed a test with Adore Beauty, a leading online beauty store in Australia, to prove to them the impact speed would have on revenue.
Consistently those going through a faster website were shown to view more pages overall, more product pages, bounce less, and reach both checkout and checkout success page more often. We left the tests in place for several weeks to ensure results were real and found the same thing week after week.
Website performance as an umbrella phrase can mean many different things – if your business is focused on upping conversion rate, a website that converts poorly could be considered to be performing badly for your business needs.