Our approach to CRO is as follows:
- Examine all available data in order to understand the challenges specific to the business in question
- Develop hypotheses from the data and design experiments to test them
- Use the insights generated to re-iterate and thus perfect the process
Using Google Analytics, we discovered the following useful insights:
- Over 60% of total users on the Big Bus website are on mobile devices
- 54% of total transactions come from mobile users
- Conversion rate was lower for mobile than on desktop and tablet
- Of total transactions, 45% occur within the user’s first interaction on site
Because mobile devices were revealed as the leading devices on which to experience Big Bus Tours, we focused on improving the performance of mobile within the checkout process (compared to desktop and tablet).
The largest drop-off in the checkout process was on the page where the user needed to enter their details. This drop-off was seen across all devices, but this was particularly high on mobile.
One of the several hypotheses we developed to decrease this customer drop-off was using motivational messaging on the user details page. We decided to run an experiment, using a range of three messages which could be displayed in a banner at the top of the page.
We chose to display the message at the top of the page so users within the experiment were most likely to see the banner on page load.
|V1||Ready to hop on? We love New York – and we know you will too!|
|V2||Over 8,000 other adventurers loved this tour according to TripAdvisor.|
|V3||Forms suck. But this tour is worth it.|
We focused on increasing the value proposition of the product. Different versions of the message were tested, exploring varied motivational factors of the Big Bus product.
Competitor analysis further reinforced the evidence for testing the use of messaging within Big Bus Tours’ checkout journey.
Of all the locations that Big Bus Tours operates in, we chose New York due to the high level of traffic that would produce statistically significant results, quickly.
The highest performing version generated a conversion rate increase of 5% and a 7% increase in revenue compared with the control, while uncovering a whole new area for testing.