Once you’ve mastered the basics and you’re doing data science well, the final piece of the puzzle is testing. It’s also the most important element. Even if you’re doing everything else right, it’s pointless if you’re not conducting any tests.
Marketing is always a work in progress. Things shift and move constantly, so taking the time to test things out and see if they work for your business is an essential part of understanding your customers. This circles back to our original blog where we discussed the basics of data – the key is to never make an assumption unless you’re willing to test it and prove it right or wrong.
"People click through a homepage slider 32 times out of 500 visits, a click-through rate (CTR) of 0.65%."
For example, data has shown that people will only click through a homepage slider 32 times out of 500 visits. That’s a click-through rate of 0.65%.
What this shows, without question, is that homepage sliders are bad for almost everyone. However, almost all website builders will have dozens in their template libraries, so we assume that it’s best practice to have one. This should make you question what else you consider to be best practice just because everyone else is doing it. Best practice is only best practice if it works for you. How do you know if it works for you? Test it.
"45% of companies aren’t conducting any UX tests."
So, why don’t more businesses make testing a part of their process?
In fact, almost half (45%) of companies don’t conduct any UX tests at all. Given the importance of testing to really know what does and doesn’t work for your customers, this stat seems shocking. But there are several reasons why companies don’t encourage it as part of their culture. The main one being that people don’t have the right technology and, as it’s seen as an extra ‘nice to have’, there’s no desire to invest.
There are also often issues with bandwidth and capacity. In larger businesses especially, there can be several lines of sign off for a test so unless you already have a culture that encourages it, you can very easily hit barriers. Also, it can be very hard to justify the positive impacts of testing vs the possible implications of getting it wrong.
Have a culture for testing
Building testing into the DNA of your business is really important, but it’s sometimes the hardest thing to achieve. Ensure you have a proper testing environment with sample sets so that the implications are reduced. Continue to run your control data as usual and conduct tests on a subset. This should limit the impact of any failed tests. And don’t be scared to fail fast, the more you test, the more likely that it’ll net out positive. If you ensure you have the right process in place, then you can control the business impact.
Have the right technology
The Google stack has A/B testing capacity through Optimize and Optimize 360, although these can be a bit complicated to use and do require some degree of development skill. However, there are other platforms available. Both Optimizely and VWO, for example, are far simpler to use and are good for getting things off the ground quickly. They’re also inexpensive.
Give tests long enough to actually show results. Sometimes, if you make changes to your site, UX for repeat visitors becomes worse as they can’t locate things as easily and your KPIs will drop. But, as soon as familiarity kicks back in, they’ll bounce back and exceed what they were previously. You need to keep going for long enough to see the results come through. As long as your testing procedure is robust and you’re using good data to inform a hypothesis, you should be able to hold your nerve and trust yourself
This doesn’t mean being reckless! Have the courage to want to improve, make a difference and drive things forward, but don’t dive in without taking all elements into consideration. You have to do the leg work, manage the process, and mitigate the risks.
For more on testing, or any of the other sections we’ve focussed on in this blog series, check out our full data and customer insight webinar.
If you need some help with the basics, data science, or testing, you can get in touch with us at email@example.com