Sign up for our newsletter
Receive our monthly round up of what's new with WordPress.
Black Friday is coming up and you need to act now to make sure your store can handle the rush. Web developers tend to focus on page load speeds using tools like Google PageSpeed Insights or GTmetrix, but rarely do they test for concurrency. Don’t let your WooCommerce store turn your black ink red during the busiest shopping day of the year. Prepare your site now.
In this blog post, we’re going to explore the ways you can speed up WooCommerce for Black Friday and survive the Shark Tank Effect:
Let’s start with how to load test your store with Supervisor.com.
We picked Supervisor.com because it’s easy to use and it’s based on bots using real browsers to simulate users. The use of real browsers creates a test environment that’s very similar to the loads created by concurrent real-world users on your site. The AI ensures that the simulation is realistic and not just a script being run to replicate the same behaviours over and over again.
Go to Supervisor.com’s website and start off with the free test on offer, so you will familiarize yourself with the process. Before doing so, there’s a little prep to go through:
Understanding the results
As an example, we’ve looked at results of a WooCommerce site in the marine tech space.
Here’s a breakdown of what you see in the graph:
Receive our monthly round up of what's new with WordPress.
Convesio is not just another hosting provider. Our infrastructure was designed from the ground up for hosting high performance, scalable WordPress sites.30 DAY FREE TRIAL
In this graph, we see that page load times quickly decreased as the number of users increased. Once 21 users are surpassed, page load times quickly deteriorate to more than four seconds. Even though page load times show a decrease as the number of users increases, the average load time remains north of 4 seconds. At 23 users, we see an intensified trend with page load times peaking at well above 8 seconds!
36 users far exceed the capacity of this VPS with a max page load time of 17.2 seconds. That’s a bloodbath for lost sales. We’ll see why in a little bit.
To see the data at a glance, there is an Advanced Details table beneath the graph. It shows the number of users, the average load times, minimum load times, and maximum load times.
While your test is running, use Google PageSpeed Insights to see what your individual page load times are. Just enter your website URL, including HTTPS, and keep the test results handy.
You may want to check out this article about Google PageSpeed Insights from Supervisor.com: How Google PageSpeed Insights might lead you to wrong conclusion.
Page load speed with concurrent users is the biggest metric that’s missing from the most popular page speed testing tools. These tools provide a one-dimensional measure of your website’s performance that won’t help you when you have double or triple your normal traffic. To get a clear, accurate picture of your website’s performance, you need to load test it. Here’s why: 37% of Visitors Bounce When Your Site Takes Five Seconds to Load.
In this report, the author shares the above statistic and suggests that users may hold your store to an even higher standard than Google does. The author goes on to share that a one-second delay in page load time results in a 7% drop in conversions.
Tip: Supervisor.com is easy to use, if you want to run the test yourself. However, you may miss key insights that a high availability hosting expert will see right away. If you can let an expert run your test and make recommendations to you based on the results.
1. Tell us a little bit about yourself and about how Supervisor.com was born
In 2002 I founded web hosting company Planeetta in Finland, which became the market leader in Finland. I loved performance tuning hosting and trying to make our customers’ websites faster.
We hosted many busy e-commerce sites, and my customers always kept calling me and asking whether their site could handle a TV-marketing campaign. I was frustrated about how I couldn’t easily give them a yes or no answer.
I went through all the existing load testing solutions and thought that there must be a better way to do this. So from that premise, we started building Supervisor.com a couple of years ago and launched our self-service load testing SaaS this summer.
2. How is Supervisor.com different from other load testers in the market
All the other load testers are concentrated on building tools for QE engineers to do testing. These tools only mimic the traffic your users might make, and they require days or weeks of manual work to get any meaningful test coverage.
And they only measure technical performance like how many requests per second a web server can handle. But that does not really translate well to user experience, which is important for your conversions and bounce rates.
We wanted to build a load test solution that any e-commerce merchant could run, and that would give understandable results like how many users your website can handle.
We do the load testing by AI-driven software robots on a massive scale (from 1 to 100,000s). The robots use real web browsers to open the web site, and they will click through the links, add products to the shopping cart, scroll the pages, and do whatever a regular user would do.
All the robots also report on the critical KPIs affecting user experience like page load time. So we can show you how fast your website is on Black Friday.
All this is fully automatic, and the load test could be a matter of a few minutes to complete.
3. What’s on the roadmap for Supervisor.com?
Lately, we’ve been working on reporting. We will collect everything that is happening, including the HAR file, on each page load that our robots make. We can then drill down to single requests and do some data mining to find patterns and root causes for problems.
We have just incorporated the Web Vitals metrics like LCP and CLS to our reporting and visualizing those on our test runs. It seems like everyone is trying to improve those metrics for better UX, but there’s a lot of confusion on what they mean.
In the long run, we will keep improving our AI capabilities. We would like our robots to understand simple commands like “Buy 500 white Nike sneakers from yourshop.com.” So we are working on machine learning and computer vision things.
4. With Black Friday coming up, do you have any advice for WooCommerce store owners?
Successfully hosting a WooComerce site requires a lot of specialized knowledge from the web hosting provider. WooCommerce isn’t scaling that well out of the box, and your site could crash with fewer users you could even imagine.
You can’t fix something if you don’t know it is broken. So start by testing your current performance. You can do it for free at Supervisor.com for up to 50 users.
Even if the news is not good, it’s better to get them now and not on Black Friday with real users. You will have time to fix the issues, and then you can happily wait for Black Friday that will be bigger than ever online this year.
More about Lauri and Supervisor.com:
Lauri Kasti on LinkedIn
If for any reason you don’t want to use Supervisor.com here are a couple of solid alternatives:
Load Ninja works in a similar way to Supervisor.com, simulating the actual end-user experience across tens of thousands of browsers using their TrueLoad technology. Their interface is clear, logical and easy to use – a big bonus.
Here some of its main features:
Load Ninja is priced quite high – USD $289 / month for 100 virtual users (per test) and a maximum of 8 hours, so not accessible for smaller WooCommerce stores but it given the value it providers it pays itself quickly.
Loader.io may not simulate usage with the same level of sophistication and realism as Supervisor.com and Load Ninja but is much more affordable (including a free tier). Pro plans star from USD $99.95 / month covering 100,000 clients / tests.
It’s easy to use even for the less technically savvy but dev-friendly too, as it can be integrated into deployment routines.
If you’re new to load testing Loader.io is a good place to start, and a platform you can trust as it has been developed by the good folk at SendGrid.
There are three things you need to do to ensure that your website is ready for the busiest online shopping day of the year:
Tip: read our 7 Ways to Make WooCommerce Faster Before Black Friday blog post to get started with WooCommerce optimization.
Our server architecture is built with scalability in mind so that your shop handles even the largest traffic spikes just as smoothly as it does during low-traffic periods. We use Docker containers for this — you can read more about this here: Docker Hosting for WordPress: What It Is, How It Works, And Why It Makes Sense
It works like this: instances of WordPress run in their own separate containers and when the platform anticipates a surge in traffic it instantly spins up additional containers share the load. Once the spike is over the extra containers are removed. Also, WooCommerce hosting should be complemented by a set of caching rules so that much of the traffic can be handled before it even hits the server. We use Cloudflare for this (learn more about it here) and bundle a number of tools and services in our WordPress hosting plans.
This approach is much more effective than traditional set ups. If you’re running WooCommerce on shared hosting 99 times out of 100 your website will grind to a halt or crash if there is a significant spike in traffic. A VPS will be able to handle a bigger load but doesn’t scale unless you have a complex (and expensive) set up.
In this article, we talked about how to load test your store using Supervisor.com and why that’s important. We also covered what you need to know and what needs to be done so that your website can handle the Black Friday rush.
Finally, we shared how we can help your website handle the shopping frenzy that’s coming in a couple weeks.
If you need help with any of this just reach out. We’re offering a free review that includes the testing covered in this article as well as checking page load speed and the overall health of your WordPress install.