This post is part of a series: First – Next
An Extension of WordPress
A look at Google Trends indicate that WooCommerce has been gaining popularity very quickly over the past three years. This could be due to WordPress already being prolific on the content management side of things. Users who already have a WordPress site can easily implement a web store of their very own. This is in contrast to Magento, which is built for true enterprise-level transactions (even if it’s just Magento Community Edition).
Background and Setup
If you missed our previous blog post, we are scoring six different e-commerce platforms based on web performance. We continue our e-commerce benchmarking series by giving you a glimpse of WooCommerce by scoring it against the same standards as our Magento store in the previous blog post.
WooCommerce with Sample Data. It’s so clean!
Novices won’t find WooCommerce difficult to set up since it is just an extension of WordPress. I didn’t break a sweat installing and configuring WooCommerce, even with sample data! I imagine this won’t be the case for all platforms I’ll be testing. Anyway, after setting up my sample WooCommerce store, it was time to begin testing! Again, we are going to run three separate load tests and choose the test with the median average response time.
Testing Phase
The Three Load Tests for WooCommerce (click to expand)
Performance Errors
The median response time happened to come from the second test in the series. Now let’s check out LoadStorm’s error breakdown page of test #2. I split up the error breakdown into two graphs: 500 type errors and request connection timeout errors. Read request timeout will be omitted for this example.
(Figure 5) Breakdown of 500 Errors throughout Test #2
(Figure 6) Breakdown of Request Connection Timeout Errors for Test #2
Not looking too good. We see 500 Internal Server errors near the beginning of the test (Figure 5). At five minutes and 512 VUsers, 500 errors already reach eight percent! We also see that request connection timeouts are a problem (Figure 6). Request connection timeouts mean we can’t even establish an HTTP connection from our engines to the target server. This doesn’t build a good case for WooCommerce. Error percentage is an important marker of performance.
Score
After selecting test #2, we graded against the scoresheet. The WebPageTests were run at their usual time of 0, 30, and 50 minutes. Notice how WebPageTest scored generously with decent load times. The Achilles heel though is the estimated scalability of WooCommerce at about 412 concurrent VUsers! This is also apparent in the other two load tests. For such a clean and minimal store (with sample data), you would assume that it could handle much more than that. Next time we’re going to inspect osCommerce and see if it has better scalability.