Insight

4 of the best website optimisation tools, and how to choose the right one

September 13, 2017
Jamie Evawin

So you've got a new product and you can't flippin' wait to start experimenting and optimising your website. I can't blame you!

But hold your horses, rein it in, and ask yourself: What software am I going to use to start experimenting? Because if your only tool is a hammer, every problem looks like a nail.

Start by asking yourself exactly what it is that you want to test on your website:

  • Do you have a whole new website and want to test it against the old one?
  • Are you looking for a personalisation solution to optimise towards specific audiences?
  • Or are you planning on replacing all your blog post cover images with doughnuts to increase click-through-rates?  

It's a good idea to determine your essential needs beforehand to make the process a lot quicker. Another element you need to factor into your decision is how much traffic your website receives. This is important, as a lot of tools have different pricing plans based on the level of traffic. In fact, most premium tools will not supply a quote unless you have this figure in mind, so jump into your web analytics and get that data beforehand.

The more traffic your website receives, the more investment you're likely to have to make.  Once you have your key ingredients, you can start to filter down your options to find your dream testing tool.  

In the list below, we've highlighted what we consider to be the best tools for an optimisation programme for specific needs. Please note, however, that this is a fairly generalised set of recommendations, and you should always conduct your own research into testing software before making any decisions.  

Best value for money: Visual Website Optimizer

Visual Website Optimizer (VWO) is commonly considered the best value solution. Starting from a mere $49 per month, VWO provides the most bang for your buck in terms of both features and traffic allocation.  

However, one drawback of VWO is its lack of more advanced features. Experiment segmentation is very basic, so any targeting options are best written within the code of the experiment.  VWO also has a bit of a fiddly integration with most external measurement tools. It's super simple - just a tick of a box, specify your custom variable/dimension slot and you're away. Only problem is, it very rarely works first time. In most instances, you'll need a bit of analytics know-how to get third party integrations working. For this reason, we'd recommend implementing any third-party tracking through manual tagging.  

Best free solution: Google Optimize

Google Optimize has been the talk of the town in the world of optimisation for the past few years.  Initially, when Google announced their free experimentation software, every other budget-friendly testing tool was quaking in their boots.

However, reality soon set in and people realised that they had a long way to go to catch up to the others. The beta was buggy and had a very unusual UI compared to the competition. However, there was one huge advantage of Optimize: its Google Analytics integration.  

All experiments run out of Google Optimize are automatically integrated with whatever Google Analytics profile that you're utilising for analysis. This is a lifesaver for less technically-minded users that are looking to run some simple A/B experiments.  

Best for personalisation: Qubit

A firm favourite at Code Computerlove, thanks to its developer-focused user interface, Qubit is our optimal platform for clients looking to implement sophisticated personalisation.  

There are a few reasons for this, the primary one being its advanced segmentation options. When using Qubit, you have a few different ways to categorise your users into specific segments:

  • Integrated segmentation: You can create useful segments from rich customer data collected from a wide variety of sources, then target them immediately with personalised experiences.
  • Create the right segments for your brand: Use data points to create segments with broad groupings to reach lots of visitors. Alternatively, target clusters of select individuals with the most precisely tailored content.
  • Self-segmentation: Don't just assume customers are in a segment based on their behaviour; ask them. Use Visitor Pulse surveys to let customers express preferences, and use those preferences to deliver personalised and relevant journeys.  

Once you've segmented your users, you can then use the experience editor to begin building personalisation campaigns ('experiences' is essentially Qubit jargon for campaigns). And this is where I get a little fanboyish over Qubit‚ Both the editor and triggering options for experiences are very developer-focused. Triggering options can be written utilising JavaScript, allowing you to create more advanced conditions for when your personalisation campaigns are activated. The editor is also heavily developer focused, and this is the only tool that encourages code-written campaigns as opposed to the typical clunky WYSIWYG editors.  

Best all-rounder: Optimizely

Optimizely takes our best all-rounder spot for a multitude of reasons. Simply put, it's our favourite tool in almost all instances. Firstly, Optimizely pretty much has every gizmo you'd want for experimentation. It offers a great depth of features‚ not only can you run A/B tests, multivariant tests, split tests and personalisation campaigns, but also server-side experiments which allows you to test directly within application code using Full Stack.

They are also releasing a new program management area of Optimizely X, meaning your entire optimisation program could soon be consolidated into one online platform. We're a big fan of this as we're yet to find an experimentation tool that efficiently merges management, build, and running and analysing experiments.  

Best for split testing: No tool required

Site testing is a term we typically use for simply splitting traffic between two web pages. This is one of the more common methods we use at Code, as it allows us to experiment with large scale changes such as a website redesigns and determine what impact they have before releasing.  

Generally, we would recommend not using any experimentation tools at all for split testing. Due to a bit of a performance obsession, my general opinion is that no matter how fast your experimentation tool is, how quickly your server responds or how optimised it is for delivery, you will still suffer some kind of lag when split testing through an experimentation tool. In some instances it could take a matter of nano-seconds to redirect traffic to a different web page, but you're still relying on more variables than necessary to do so.  

Splitting traffic is easy, and any developer worth his weight in salt should be able to set this up with relative ease. There are a multitude of free split testing scripts online as there's a multitude of ways to do it; our best use-cases come from using a proxy to shift the traffic.  

So which is the best one?

Only you can make the decision on what experimentation tool is best for you. Annoying answer, right?  No tool is currently the undisputed best one out there. In Smart Insights' Essential Digital Marketing Tools guide, they specify that Optimizely and VWO have around 55% of the market share, which sounds realistic due to their maturity.  

 

However, as mentioned previously, the right tool for you is totally dependent on the requirements you have. The above chart only displays the technologies based on the amount of tags installed on different sites, not based at all on capabilities or suitability.  

Which is your favourite optimisation tool and why? Let us know by Tweeting us @Computerlovers.

Image source: picjumbo.com

Continue reading