I have a love-hate relationship with speed testing tools. Each comes with their own limitations, but each gets treated as the Truth by the client who looks at it.

This has spawned dozens of posts on how to get 100% score on Google Page Speed and other diagnostics. The dangerous thing is that 100% diagnostic score can cripple or greatly slow down your site! I’ve seen countless sites broken because people have moved JavaScript files to places that they won’t work!

Bad Recommendations get worse

Enter http/2

The arrival of http/2 has made this even worse: some recommendations given by tools like PageSpeed can actually slow your site down! A classic is recommending concatenation of javascript and css files. A known anti-pattern with http/2. Improving your score would make your site load more slowly!

But testing is really helpful!

Testing front end speeds can be a really helpful diagnostic test: it helps us to spot weak points, tidy up broken code and resources. For us, that mainly happens when we migrate a site to our Managed WordPress Services, and during development.

I started a search for something significantly better, and I found Dareboost.

The Dareboost Review

First up, Dareboost understands the difference between http/1 and http/2 and gives different scores and recommendations based on the technology available. But more than that: using the comparison tool, you can see how much of a difference http/2 makes!

dareboost review http1 http2 comparison
The comparison tool is a super-powerful feature.

Comparisons FTW

The comparison tool allowed us to do this detailed review of how much an ad network impacts page load time, and also allows us to not include ads in performance reviews. That’s a great help, due to the ‘noise’ generated by third-party resources that we can’t influence.

SEO? We can compare your site speed with your competitors, to see if that’s a ranking factor that you’re dragging behind on.

Wondering why mobile conversions are low? We can easily compare mobile vs desktop speeds.

Not sure if your browser caching is working? Test first vs second load.

High number of users in low-speed internet areas? We can compare different connection types: 3G, Cable and Fiber averages, or any custom latency and line speed that you’d like.

Comparisons allow us to focus effort into the most-important things first, and understand the trade-offs when we need to make a choice between competing priorities.

The Big Picture

Like most tools, Dareboost starts with a big-picture diagnostic. Here’s an example from a homepage of a magazine-style travel site, Indie Travel Podcast.

Seeing the big picture when it comes to speed

We can see that the simulated user was in Washington DC, using the Chrome browser on a pretty average connection speed. At 74% total score, there’s 7 “priorities” and 12 “nice to haves”; there are no browser warnings (broken javascript, anyone?), but only 88% of the resources on the page are http/2 ready… so that can be improved too!

Numbers are always subjective, but the way Dareboost breaks down the major numbers is pretty smart. With the requests vs weight pie charts, we can see if there are any major discrepancies; and each of three speed results are benchmarked against some kind of standard.

Time to first byte

Benchmarked against 200ms, the Google-recommended standard.

Start render

Benchmarked against 1 second; a generally-accepted minimum time that the page should start to be visible.

Fully loaded

I don’t know where this number comes from, but Dareboost states that “67% of users demand that a page must be loaded within 4 seconds.” Sounds good enough to me.

So our big picture has some vital information that’s missing in other tools, and it’s presented in a smarter way. But how good are the recommendations?

Dareboost’s Recommendations

As you might have seen, the more I looked into what we could do with Dareboost, the more excited I got. Better data, better tuning of the test environment. What I didn’t realise is that Dareboost wasn’t just going to improve our speeds, it was going to start a movement to improve all our front-end code.

You see, these recommendations are broken up by several categories:

  • Accessibility
  • Browser Rendering
  • Cache Policy
  • Compliance
  • Data Amount
  • Number of Requests
  • Quality
  • SEO
  • Security
  • Google Analytics
  • Google Font API
  • jQuery

Each topic is colour-coded from the outset, so you can see where you have bigger issues to fix.

What I love, is that this is about more than speed: it’s about code quality indicators overall. And while I might not agree with or be able to implement every recommendation for every project, it’s great to be reminded about them.

I’m interested in pulling down that total pageload time, so I jump into Browser Rendering recommendations to see what I can see: Each recommendation is “folded” with a bit of javascript, so I can quickly scan my issues. Clicking on one opens it up with a solid description of the issue, and specific code examples or files that need to be fixed. I’ll do the same with caching, data amount and number of requests to create a checklist of performance improvements.

Two issues to fix here!

These recommendations tick the boxes. They are:

  • Sensible
  • Specific
  • Actionable

They’re also colour-coded for importance.

Like any tool, this is a diagnostic. 

It’s not always possible to defer and async JavaScript, so we might be stuck with a bad score here until we re-write a whole chunk of code. We might decide that the feature that that creates this code is worth the hit. We might decide it’s a good candidate to get rid of.

In a WordPress environment, this tool helps us to identify plugins that are bad actors when it comes to loading their code inefficiently, and on pages where it’s not actually used. But sometimes, you just have to suck it up though! The fastest webpage is a blank one 🙂

A time machine

At our service level, we’re storing data for a month, which allows us to look back in time and see what impact our work is having. This is becoming especially interesting as we integrate Dareboost into our new Managed WordPress Hosting client reporting.

I’d love to have a longer dataset, but a month is a good period for many iterative projects.

One thing I’d love to see is the ability to compare a historical test with one today: at the moment, you can hit ‘retest’ to replicate the exact conditions, which is great. I think it’d be even more powerful to see an A/B comparison of the past and current conditions, like we can with the other comparison tests.

After publishing this article, Dareboost got in touch to tell me that it’s already possible! To compare any two existing reports, use the Audit screen (Dashboard > More) and select any two existing reports. At the bottom of the report list, hit “Compare”. It’s magical!

dareboost review - compare reports
The missing feature… That wasn’t missing!

Dareboost has more than one-off reporting tools, but that’s our primary use case right now. After we master stage one, we might start to dig into user journey analysis, regular page monitoring, and other automated tools. Overall, we’re sold.

How is Performance Foundry integrating Dareboost?

The depth and accuracy of the Dareboost speed test is starting a slow but powerful sea change at Performance Foundry.

New projects

We have a higher bar to pass with new development projects: not only for speed, but also for basic SEO and accessibility rules. This is being worked into our QA processes and code reviews to ensure we’re not releasing code that could be better.

New hosting clients

All our new hosting clients get a free speed report as part of their migration. We’re moving away from the old combination of Pingdom, GTmetrix and Google PageSpeed to using Dareboost recommendations for front end recommendations*. This saves us time, reduces confusion, and gives better results. (Although Pingdom load time results are normally more flattering!)

Standalone speed tests

We have a standalone website performance package that differs from our on boarding speed testing, in that we are working with a third-party server environment; but Dareboost is becoming the tool of choice here too.

*In both speed reports, we use NewRelic and server scripts to log and deeply understand server-side data.

Staff training

Dareboost’s detailed recommendations provide a great starting point for training new staff members on performance issues, and kicking off conversations around front-end quality.

Security and Performance Improvements Foundry-wide

Dareboost has opened our eyes to a raft of improvements that can be made to our client sites across the whole hosting network. We are currently working on a new tool, code-named Anvil for all our hosted WordPress sites that will allow us to quickly ‘fix’ many of Dareboost’s recommendations, making sites faster and more secure without any additional development costs for our clients.

And that’s what we’re after: creating the best WordPress services company in the world. We can’t wait to show Anvil to you.

Want to try Dareboost? Coupon code here!

Anyone can do a few free tests with Dareboost, but to make the most of it, sign up!

Dareboost is a premium tool, with the ability to subscribe and get regular reports on your speed, security, accessibility and performance. Site owners can keep their finger on the pulse, agencies can white label reports. There’s a lot of bonus tools we haven’t touched here.

Use the coupon code FOUNDRY20 to get a 20% discount on all Dareboost plans, available for both yearly and monthly subscriptions. This isn’t an affiliate code; this isn’t a paid ad placement.

We’re excited to be using Dareboost and want to see them grow, flourish and keep providing us with great tools to help make the web a better place. They’re comp’ing some reports for us and we’re hoping to feedback to the product team as we find our way around, and give back that way.

Here’s to #webperf!