As a web developer I quite regularly optimize loading speeds of various sites.
However, what I learned stands in opposition to what you can read in most sources on the Internet.
What I learned is that:
- Some speed optimization methods increase load time
- Speed testing tools don’t rank sites by their load speed!
- 20% of speed optimization techniques give 80% of results
- Average load speed is not 2 seconds. It’s 8.
- People don’t care (that much) about load time. They care about experience.
That’s why I’ve decided to write this article.
Some speed optimization methods increase load time
The truth is that the benefits of using most optimization techniques come at a cost that sometimes outweigh these benefits.
Let me explain.
Very often Free Cloudflare CDN is recommended as an easy sollution to boost your page loading speed. However doing this also increases so called TTFB (time to first byte).
It means that even though the site might load faster overall (which is not always the case), it always takes longer(!) for the visitor to see response from the server.
Read the last section of this article to learn why it is a problem.
Another example is a popular technique of minifying and combining files. In reality:
- it may be less beneficial than loading multiple files via standard HTTP/2,
- it can drastically increase server response time if the site is not visited very often.
Speed testing tools don’t rank sites by their load speed
Testing tools rank sites by rule-sets, i.e. by how many optimization techniques they use.
Here’s an excerpt from GT Metrix’s FAQ page explaining why there’s almost always a difference in PageSpeed and YSlow scores.
PageSpeed and YSlow use different recommendations when analyzing URLs. Each service analyzes a page using a set of rules that they believe are most relevant to page speed and performance. Most of the rules overlap or are very similar to each other, but in general your scores should be comparable.
It means that striving for a perfct 100% score is an art for art’s sake.
In fact it can even be harmful for your site speed (and store’s conversion rate) if you overoptimize and your site starts to load slower.
20% of speed optimization techniques give 80% of results
The Pareto’s 80/20 principle is very much true in speed optimization.
There are more than 60 methods of improving load speed of WooCommerce and WordPress sites. However, the biggest gain is only from a few of them.
Doing advanced optimization is mostly just a waste of time.
Average load speed is not 2 seconds. It’s 8.
Here’s a nice article that not only shows statistics of load times (for 2018) but also gives good context to interpret them.
People don’t care (that much) about load time. They care about experience.
Websites are for people, not robots. And people don’t care whether a website loads in 3.1 or 7 seconds. They care whether they experience it as slow or fast.
If you clicked a button and had to wait 4 seconds for the next page to start loading than you might leave. However, you could easily wait even longer if you saw a progress bar or a loading spinner.