Measuring the speed of a site is not an exact science. Any test you take will be dependent on the distance between you (or the tool you used to test) and the site. You should also take into account the connection speed, in case you're doing an empirical test using your own device and ISP.
Having said that, there are a few indicators you can use to make sure that your site is fast.
We highly recommend the highly unflattering Dareboost.com for front end performance tests. With Dareboost, you can not only test any page on your site, you can also sign up for tracking. Dareboost will not only give you a page speed, but also a technical score that will change over time.
In our opinion, Dareboost has the most useful, actionable data from any front end testing platform we've seen.
But there are simpler, more flattering tests around...
The monitoring service Pingdom has a free tool that will give you a very nice-looking summary of all your site's issues.
As mentioned before, try to use the closest location to your server. If you're on one of our East Coast US servers, choose New York; if you're on the West Coast server go with San Jose. Stockholm is the best option for sites on our Europe server, and Melbourne is a good choice for AU server clients.
After you run the test, a summary will appear. This alone can give you a lot of useful data:
The most important element to notice here is load time -- under a second is good, more than two seconds is cause for concern. Also important are page size and requests, as those don't vary from user to user. The lower you keep those, the faster your site will be.
It's worth running the test again using a different location, further from the server. This will allow you to see the load time visitors from other locations are experiencing.
However, be aware that Pingdom is not a perfect tool! Its main failing is that it does not take into account HTTP/2, which is the latest protocol and which your site will be on if you are a Performance Foundry hosting client. Previously the elements of a page were loaded sequentially and therefore it was advantageous to perform optimizations such as combining (concatenating) and minifying code to help the browser read it faster. HTTP/2 processes requests completely differently, and the opposite approach takes best advantage of HTTP/2 technology, that is having code in bite-size chunks and serving elements all at once for the browser to load as soon as each one is ready. So, many of the optimizations that Pingdom recommends are actually detrimental to your performance under HTTP/2.
Another important indicator is the response time of your server. For that, you can use Google's PageSpeed Insights:
Just like with Pingdom's tool, you will get way too much information -- don't feel like you have to analyse it all! But among all those stats, one piece of information is really important: the server response time.
For Google, anything more than 0.2 seconds should be optimized, so be on the look out for that number.
Be aware that Google PageSpeed isn't a good tool for measuring the general speed of the site (despite its very misleading name).
You should keep these tools handy whenever you are creating a post or page, so you can test if the final product is becoming too slow.