Conversion Rate Optimization Case Study: How a travel website doubled site conversion rates in one year

This case study was originally published in the MarketingSherpa email newsletter.

THE CUSTOMER

TravelPass Group serves its B2C customers by helping them book hotel rooms and serves B2B customers like hotels, wholesalers and travel agencies by getting their products in front of the right consumer.

This leisure, travel and tourism company offers consumers an expanded inventory that is difficult to find in any one place. “If we don’t have rooms available, no one does,” said Bryce Miller, former Senior Optimization Manager at TravelPass Group (and current Marketing Director at Assemble).

TravelPass goes to market through a family of travel-related websites, including Reservation Counter, Reservation Desk, Nitecrawler, RentalCounter and Jauntaroo. It has surpassed $2 billion in gross hotel booking revenue across these sites.

One of the requirements for working at TravelPass is a customer focus, and according to the company’s About Us page, “We believe in putting people first.” Part of the organization's value proposition to the B2B segment is its focus on optimizing the user experience.

THE OPPORTUNITY

One way to put people first is with A/B testing, to create websites that better serve customers which ultimately generates higher sales.

“The company had done a lot of testing before we started to change things, but we were able to kick testing into a new gear. We altered the focus of testing from basic benchmarks like number of tests or orders, as well as loose testing practices, to a more rigorous process of user discovery. We traded number of tests for quality of tests by testing and iterating on concepts even if they did not pan out the first time. In doing this, we were able to confidently find enormous gains that would have been left undiscovered without this process improvement,” Miller said.

Miller managed the product user testing process at TravelPass, and transformed the product team into an optimization team focused on customer data, testing and results.

First, his team helped set the groundwork for testing by implementing tracking changes. This allowed them to track tests across websites and phone calls. Then they set about building hypotheses to determine how to better serve the customer.

“By focusing our efforts on specific elements in the mind of the user, we were able to isolate what types of tests we should focus on. Combining that data with analytics from our analytics platform told us we were testing the right type of tests in the right places. So when we did not get a successful test the first or second time, we kept iterating, allowing us to eventually find the improvements we knew were there,” Miller said.

Here is a closer look at three tests his team ran.

TEST #1: Checkout flow

When visitors search on Reservation Counter, they are given options of different rooms to book from different hotels. The challenge is that each hotel chain or individual hotel can be named something slightly different, or come with different features or promotions.

“Because the industry does not have standard naming conventions, the user suffers when trying to understand what is going on, so we simplified it a bit for them,” Miller said.

The team wanted to discover the best way to display these options to increase revenue.

Control

The site was serving up rooms to customers simply based on keywords and availability.

Treatments

The team tested several treatments. The treatments had the same design as the control, but tried different ways of prioritizing the options displayed to the customer.

“We did not do complete room-type matching, but rather a quick and dirty approach that made the experience better for the user,” Miller said.

The highest-performing treatment showed room-type matching by keyword in the name only. Standard keywords like king, queen, double, suite, deluxe, etc. It only showed a maximum of eight rooms. It always showed one non-refundable room and one refundable room per type as long as there was availability. It showed the lowest price rooms selected by lowest total cost, and then sorted by lowest nightly price.

“One of the secrets to succeeding in this life, and is no different in conversion rate optimization, is to know what not to focus on. If you can decide only three things that matter on the page or experience, and what order your customer wants to address those, you will have a lot more success when you emphasize those three things in the proper order and de-emphasize the rest,” Miller advised.

Results

“In the end, the highest revenue treatment actually came from showing the lowest priced rooms available in our inventory,” Miller said.

The highest-performing treatment produced 19% more revenue on desktop and 31% more revenue on mobile.

“By focusing on what the user was going through, we were able to combine hotel room offerings from multiple providers in a way that allowed the user to decide what was best for them without feeling overwhelmed,” he said.

TEST #2: User testing-derived design

The design team conducted user research, including user testing, to create a new design for the hotel landing page template on Reservation Counter.

The testing team was concerned that the new design, while attractive looking, could hurt the user experience, so they ran the following test.

Control

Here is what the original hotel landing page template looked like.

Treatment #1

This is the new design based on user testing.

Treatment #2

Created by the testing team, this treatment focused on the thought sequence of the visitor.

Treatment #3

Treatment #3 was similar to Treatment #2 but had a few small changes in design and color.

Results

Treatment #1 generated 37.8% less online orders than the control. Treatments #2 and #3 generated 29% more online orders than the control (there was no significant difference between Treatment #2 and Treatment #3).

“Using the MECLABS Institute thought sequence process, I was able to immediately see problems with the work of designers who were not familiar with the process. This allowed me to have the confidence to work with the designers to come up with something that would be more helpful to our users, and positively impact the bottom line,” Miller said.

TEST #3: Impact of phone number placement

The testing team discovered that having the contact center phone number prominently displayed on the website increased orders. Next, they wanted to optimize phone number placement.

Control

The Reservation Counter website originally had the phone number in the header on the search page.

 On the hotel landing page template, the control had the phone number under the H1 headline.

Treatment (Desktop)

For the treatment, the team kept the phone number in the header on the search page and under the H1 headline on the hotel landing page template.

But, the team also added the phone number on the left-hand side of the hotel landing page, under the hero image on the hotel landing page, as well as at the bottom of the page.

Treatment (Mobile)

On mobile, the phone number was included at the top of the hotel listing under the hotel’s address, on the hero image, and in a sticky footer.

 

Notice how the team had different phone numbers on the desktop and mobile versions to be able to track which device type led to the phone call.

 

Results

For the desktop version, the treatment did not hurt conversion rates noticeably online but had a slight increase in phone call conversion rates.

The mobile treatment experienced a 14% drop in online conversion but saw a 14% lift in phone call conversion rates. This was a win for the team since there are a lot more conversions from the call center than from mobile devices through the website.

KEY TAKEAWAYS

Over the course of a full year of testing and optimizing the website, Miller and his team were able to learn enough about the customer to ultimately double website conversion.

“Know who your primary customer is. Some marketers optimize for their boss. If your boss has never learned the MECLABS process, then you will end up testing a lot of your boss’s ideas and be blamed for not improving the product or site when his or her ideas fail. In my opinion, you have to realize your primary customer is the end user. Focus on that customer, follow the MECLABS methodology, and you will find opportunities no one else thinks of and ultimately improve performance,” Miller said.

Key Contributors

Miller expressed his gratitude to his team for the success profiled in this case study.

“First, this would not have been possible without TravelPass Group giving us a chance to take control of the sites and truly own the outcome. The entire product team made this possible, but Scott Knudson, the senior product manager, was the oil and the glue of the entire operation — and none of it would have happened without his trust and belief that we could make a difference,” Miller said.

Related Resources

MECLABS conversion marketing training – Learn how to apply the MECLABS process to your marketing and business challenges

Unlock the Power of Your A/B Testing Program

How to Drive Conversion Using a Value Proposition-focused Testing Strategy in Email Marketing

How We Stopped a Beautiful Page From Ruining Everything (via Scientific Web Optimization)

How The Website Team Increased Call Center Orders by 14% (via Scientific Web Optimization)

The Marketer as Philosopher: 40 Brief Reflections on the Power of Your Value Proposition (book by MECLABS CEO Flint McGlaughlin)