Officially the best small content agency in the UK 2018

How to form a culture of testing

white underline stroke

Running an A/B test now and again to improve conversion of visitors to customers isn’t a bad start.

But to be truly effective, your ultimate aim should be to develop a culture of testing in your company, and with it a deep understanding of your customers. Conversion rate optimisation (CRO) is all about ensuring you have the right tools to do just that.

People will make the difference

Training is part of the solution, but it’s the people in your business who support the growth of a culture of testing that will ultimately make the difference.

It’s about getting everyone involved in the idea of optimisation, from the person at the bottom waving the experimentation flag through to the CEO and your company’s board.

That said, not everyone needs to know exactly what is being tested every week. But a summary email, a monthly meeting – or even a passing conversation – will make the difference. Try out different mediums of communicating out what’s being tested and stick with that and make it a habit.

How do you form this culture?

The first step is identifying and solving the real pain points for your customers. For example, are people getting stuck on a particular question or are they unsure what the next step is in the journey? Focus on the experience and relentlessly improve it. By involving your marketing, product and customer service teams, you can identify, setup and run a test that aims to improve an issue rather than, say, changing the button colour for the sake of it.

The second is adopting an agile roadmap – set goals, not deliverables. For example, your goal could be to ‘improve interaction on page by X by the end of X’, rather than ‘we are going to run these series of tests’.

The third and final step of forming an experimentation culture is to invest in capabilities, hire talented people who will embrace CRO, and educate existing employees about its benefits. Finally, invest in engineering and your developers. After all, reliable tests are essential to ensuring the stats you report back are accurate without question.

If you haven’t got dedicated resources or the budget right now for CRO then an external agency can be equally powerful and can often bring in fresh ideas and concepts that you may never have thought about.

The core team and the wider business

Experimentation culture is built on:

  1. Broad participation from the whole business
  2. Transparent learning and sharing
  3. Effective process and structure

The last point, effective process and structure, is about forming a constant loop: build a backlog of ideas, prioritise them, launch tests, analyse results and start again.

Ideally you should have dedicated resources and a team to follow this process through. At the very least you need a product manager. If the product manager doesn’t have dedicated resources – like a designer or developer – then they will hit walls and it will be hard for them to run the tests.

Your backlog and maintaining a constant cycle is key. With a backlog in place, you don’t have to be creative (or panicking) each week of what to test next. Use a system for prioritisation, such as the ICE (impact, confidence and ease) rating system, Trello have a good post about how they use the ICE system.

Using a framework for your experimentation

Before experimenting, you’ll start with a hypothesis. Typically this will be along the lines of “if [cause], then [effect], because [rationale]”.

One framework you could work with to test your hypothesis is the “problem, solution, result” framework. This is broken down into the following:

Problem

The first part is identifying the problem, which informs your hypothesis. By undertaking qualitative and quantitative research you validate your assumption of a problem.

Qualitative research could be asking your customer service teams about complaints, feedback forms and doing user interviews. Quantitative could be looking at your website analytics, or if you’ve got any tracking tools like SessionCam you can see how far people are scrolling down, what they are clicking on and generate heatmaps to see where their attention is focusing on.

Sometimes it can be tempting to rely on your ‘gut’ and just run the test. Instead use your data to confirm this feeling, and this will improve the likelihood that your test will pay off.

Solution

The next part of the framework is the solution. It can’t be emphasised enough, that every test you run needs to be recorded. The solution should make sense to anyone who reads the document without pictures. Alongside your hypothesis it should contain your rationale and why you ran with that particular solution.

Result

There are two parts to the results, the before and after. The before is your prediction that by running this test it should have this outcome. It should also contain the metrics you are measuring to determine whether the test was a success or failure. Once your test has finished, the document containing the hypothesis and solution should also contain your results and what your next steps were.

This is just the start, but take these steps and you’ll be well on your way to creating a genuine culture of testing in your business.

Contact us

Michael Gearon

Mike is responsible for the development of emerging and growing technologies that relate to the customer experience across multiple touchpoints including voice interfaces like Alexa and Google Home. As well as helping businesses with online conversion (CRO) and providing UX and UI guidance and support.

Contact us