Performance Benchmark: Q and A Sites (90 Second Analysis)
Sharing interesting finds from our benchmark feature, coupled with quick-hit analysis. Make your own by signing up for a free Yottaa account.
For this benchmark we compared a few of those always-popular “”Q and A”” sites. (These four are only a small selection of the many sites focusing on Q and A – a snapshot, not a comprehensive study).
Overall Performance – Load Time
First, looking at Time to Interact, i.e. total load time, we see a pretty tight spread of performance, and consistency over the last month. Yahoo Answsers is fastest and also has the smoothest line. They must be using Yahoo’s own Yslow rules for fast performance.
Blue – Wiki.answers.com; Red – Ask.com; Green – Answers.yahoo.com; Purple – Blurtit.com
All the sites fall between 4 seconds and 7 seconds on average. This roughly reflects internet-wide averges: the mean (the yellow line) is just under 6 seconds, which is around the median for all sites on the web. So far, so average.
Time to Start Render is Fast
This table shows the page load sequence for the four sites. (The shots were taken from the samples closest to the median).
What this tells us is that across the board, the pages start rendering visible content really darn fast. 3 of the 4 show a Time To Start Render under 1 second (shown as 1000 ms above). The median across the web for Time to Start Render is 2.5 seconds, meaning all four of these sites come in well under the median – in fact, they’re mostly in the top 25% of the web. Perhaps it’s a testament to how competitive the Q and A space is that these sites are pushing the envelope on start render time, hoping to avoid losing visitors that bounce back to Google when they don’t see any progress in page load.
Page Weight Over Time
This chart shows the page weight trending over time. The green and red fields show the top and bottom 25% of the web for weight, so none of the sites are either exceptional or awful for page weight. But check out Ask.com – it’s all over the place! Some sort of content is appearing on the site making the average for some days jump up to 1.5 MB.
We tracked down the culprit – we looked into the individual test samples that compose the benchmark and, in a sample from one of the “”bad”” days, we found a 3.7 MB flash video advertisement from a third party that brings the total weight for the page that day to 4.7 MB. That’s a LOT of extra bits to load for the browser.
So with all that weight, why isn’t Ask.com significantly worse in performance? For one, the ad only appears in some samples, so when performance was averaged over time for the benchmark, those samples were smoothed out. Anytime the video appears, the load time is longer. But more importantly, the team has done well to place the file at the very end of the load order, so the rest of the page content can load undisturbed, and the video (which takes almost a full second to load) will only parse later. This doesn’t help the overall load time, but it does ensure a good user experience.
Want Us to Cover Your Industry?
Leave a comment or send an email to firstname.lastname@example.org to let us know what sites you’d like to see in a benchmark. Or even better, create your own with the Yottaa Benchmark feature in a free Yottaa Monitoring account – you’ll can use the full feature set including backend performance, content complexity, and individual-sample waterfall analysis.