Contiki Holidays
Creating an environment of entrepreneurial risk-taking in a global enterprise, by mitigating said risks through experimentation.
Why experiment?
All businesses are risk-averse, unless that risk is underwritten. Within an enterprise business, this aversion to risk-taking often stifles progress as the potential impact of failure trumps the potential for growth.
However, a good experimentation track significantly mitigates risk, so even the largest corporation can start thinking innovatively without so much fear of failure.
In fact, failure is a good thing! An experimentation track with a 100% success rate is simply not pushing the boundaries enough. Companies operating an experimentation track should expect a 30% failure rate (that’s when the outcome of the experiment tells you that you dodged a bullet by not implementing the feature for all users).
Another huge advantage of testing is with stakeholder management. Everyone naturally has differing opinions, which can create clashes within a business and can paralyse product teams. Experimentation allows you to take the subjectivity out of decisions, which can increase velocity and confidence, not to mention improve relationships with stakeholders.
At Contiki, our experimentation track has been massively successful in encouraging the business to adopt a higher risk tolerance and has improved business unit collaboration. More below.
Example 1 - swapping button functionality to boost online revenue by six figures per month
Identified a much higher Conversion Rate for users who used the Compare functionality onsite
Ideated ways to get more users using this functionality
Split-tested the CTA shown here - 50% going to the ‘Request more info’ form, 50% adding to Compare
Results showed a significant increase in online conversions as a result, so it was rolled out globally
Control
Variant
Example 2 - split-testing pricing display to generate an uplift in revenue
We were asked to test whether removing prices from certain areas of the site generated more interest further down the funnel
We tested a few variations of showing/removing pricing, and linked up with the paid media team to run an omni-channel experiment
Whilst I can’t reveal exact metrics, certain variations of removing pricing yielded higher revenue
Control
Variant