Creating a culture of testing and learning in a digital product team is hard.
Too often, the day-to-day gets in the way. Listening to web analytics, user feedback or A/B test results comes last. This hinders the cycle of continuous product improvement.
Thankfully, two experts with experience overcoming this came to Code to explain how.
Simon Elsworth has more than eight years’ experience in digital optimisation. He heads up optimisation for Sky.com. He’s passionate about helping businesses foster experimentation.
He spoke alongside Code’s Performance Director Matt Lacey. Lacey also has tonnes of experience in this field. He has run more than 50 conversion programmes across a range of clients and industries.
Here’s a summary of their advice:
1. Start small – focus on shared routines and tools
Elsworth explained how Sky rebuilt its Conversation Rate Optimisation (CRO) centre of excellence. Sky’s teams then went from simple A/B testing, to complex server-side experimentation.
First off, focus on simple routines, rather than complex agile methodologies. “Your optimisation team is already the most agile team in the business,” he stressed. Integrate them into routines, like daily stand-ups and sprint planning, reviews and retrospectives.
Optimisation specialists can also integrate by involve developers in tooling selection. Helping them use the tools in the right way won’t hurt either. This ensures developers buy into and value the analysis.
“It’s crucial developers understand the value of their work,” Elsworth says.
2. Agree shared KPIs and metrics
Elsworth also said that everyone needs to agree to the same KPIs – right up to director level. This allows collaboration and puts optimisation into the DNA of product teams.
3. Be determined and sell the power of testing
Amusingly, Elsworth also said sometimes have to be a pain in the proverbial. It’s not easy to build experimentation into teams, so determination is key.
Getting your product owners on board requires clear, persuasive communications. Show the promise of your tests. Show what it means to the customer. Do this, and you’ll deliver unrivalled product improvements.
4. Get in the habit of testing
Elsworth believes that there is no such thing as “we can’t test that” – for Elsworth, if you can build it, you can test it.
When approaching testing, he reminded us that it’s important to not assume too much. After all, what has worked in small-scale test environment might not at a larger scale.
And he stressed the importance of on-going data analysis and insight generation. Even small changes to competitor behaviour can affect user behaviour.
5. Develop a product mindset by winning trust
Our Performance Director Matt Lacey started with the story of Code Computerlove. He showed how we embedded experts in CRO and web analytics into our product teams. To Lacey, this experience underlined the importance of championing a product thinking mindset.
Experimentation is a core element of a product thinking mindset. But too often, it’s left to a separate conversion team. This is a mistake. Experimentation should come from an integrated product team. And managers should be onboard, said Lacey.
Granted, this isn’t easy to achieve. For starters, Lacey advised taking extra care when communicating test results. Do that right, and you’ll be able to manage expectations and earn managers’ trust. From there, you’ll be able foster a culture of experimentation at the top of the organisation too.
Product thinking is growing in popularity. Successful dot-coms love it, and others are adopting it. That’s according to research by Econsultancy. In these businesses, continuous delivery and continuous improvement interlock. Lean engineering and CRO and analytics complement one another.
These businesses also have other similarities. They value products over projects. They value over volume. They value experiences over design, and insights over opinions.
6. Pick your tests wisely and know when you’re wrong
“People want to test a million things at once,” said Lacey. Engineers always want to test another approach. Plus, management executives want test data help them forecast. Prioritising your tests by impact and effort is key, he said.
What’s more, testing can prove you wrong as well as right. So, it’s important to understand and accept the findings of a test when does prove you wrong. Don’t get too attached.
Both Lacey and Elsworth stressed that once this culture takes hold, results follow. You’ll be able to add much better features and make radical improvements to your product.