September 01, 2010
How To

Page Tests Lift Site Registrations and Conversions: 3 Examples that Stopped Site-Design Bickering

SUMMARY: Sometimes your team can agree that a website needs changes, but how it should be changed is a different story. If team members have different thoughts on the right direction, it's time run tests and dig into the data -- because no one can argue with metrics.

See the top three tests an international roommate-finding service ran to circumvent office conflict and redesign key pages. You'll find before-and-after images of all three tests, which increased memberships 8% and paid conversions 3%.
by Adam T. Sutton, Reporter

EasyRoommate operates a network of roommate-finding websites in 29 countries, including the U.S., U.K. and France. The team has a manager in each country to operate a local website and its business.

Baptiste Intsaby, Director, Business Development, EasyRoommate, is in charge of marketing and site testing across the entire EasyRoommate network. At times, there has been friction between him and managers -- particularly last year when many managers wanted to tweak their websites.

"It was a bit of a mess," Intsaby explains. "Lots of fighting was coming from different sources. We never managed to really decide what was good for our users and what was not good, because all the ideas seemed good."

Intsaby and his team wanted to improve site performance without relying on hunches and opinion alone. Last year, they invested in an A/B webpage testing tool to determine the best page designs for the U.K. site. Then, they planned to apply their lessons on other sites in the network and measure results.

The strategy worked. Site performance improved across the network, and the team now has data-based design changes to share with site managers.

Below, we describe the top three tests the team ran last year and their international results (see before and after creative samples in the Useful Links section below):

Test #1. Improve "Need a Room" registration form

EasyRoommate's visitors typically have one of two goals: to search for a room or to post an available room. Visitors searching for rooms must register with the site to freely browse and contact others. Basic registration is free.

Intsaby and his team saw an opportunity in the "need a room" registration form. The page requested 27 fields of information organized into six categories, which looked cumbersome.

The team individually tested four changes to the page:
o Reorganize categories from six into four: search criteria, preferred roommate attributes, personal attributes, and ad description
o Removing six required fields
o Organize fields into a single column, as opposed to two columns
o Combining winners of the three previous tests

Each test took about two and a half weeks to reach statistical significance. The entire process, including planning, design and coding, took about three months.

Tests that improved results:
o Reorganizing the fields into four categories improved results
o Organizing the fields into a single column.

Test that did not significantly improve results:
o Removing some of the fields

The team combined these lessons for its final test and realized a 7.5% lift in the form's completion rate.

- Greater international impact

The team applied the changes on the corresponding pages of the 28 other EasyRoommate sites. On average, the new page increased form completion 10% on the "Need a Room" page, globally.

Test #2. Improve "Have a Room" registration form

EasyRoommate visitors wanting to offer a room had a similar registration page to fill out. Early on, the team spent three months testing changes to this page, including:
o Reordering the field categories on the page
o Changing the fields' positions
o Splitting the page into four separate steps/pages

"We did a lot of different tests for this page, and they were not successful at all," Intsaby says.

Following the success of the tests on the "Need a Room" page, the team decided to test the same concepts on the U.K. "Have a Room" form. They organized the fields into a single column down the page and reduced the number of categories from ten to five.

"The design was successful here as well," Intsaby says. "We had an 8% uplift in terms of conversions.

- Similar international impact

As with the previous test, the team applied these lessons to the other EasyRoommate websites. On average, the page had roughly the same impact Intsaby's team saw in the U.K., but there was some variance.

"The uplift is really different based on the country. For example, we saw an 8% uplift in the U.K., a 15% uplift in the U.S., and only a 3% uplift in France, and so on."

Test #3. Improve payment submission form

EasyRoommate users can purchase premium memberships to contact a broader range of users and get access to dedicated 24-hour customer service. The team used a single page to collect the user's desired membership length and preferred payment method. Users then clicked to proceed to the next step of the payment process.

That first page was so vital to selling memberships that the team avoided changing it last year when EasyRoommate completely redesigned its website.

"We never touched it," Intsaby says.

Once the team had begun testing other pages, though, they revisited the membership submission form to research potential changes. The team tested ten different variations. Here are three key elements:

- Copy changes

The team removed a four-sentence paragraph of copy displayed just above the membership options. This paragraph included text intended to persuade visitors -- e.g., "You're just one click away from contacting new flatmates!" -- and details about premium service options.

In place of that paragraph, the team entered the following personalized sentence:
o "[First Name], find the perfect place in 30 days -- guaranteed!"

Next to this sentence was a small link:
o "Find out more," which described the team's satisfaction guarantee.

- Emphasis on percentage savings for longer-term memberships

The team tested significant changes to how it displayed its three premium membership options: 5-day, 10-day and 30-day memberships.

The existing page displayed only the full price for each option. The test page first showed the number of days in the membership, followed by the price-per-day, and below that, in a smaller and different color font, was the full price.

On the new page, the 10-day and 30-day memberships also featured bubbles displaying the percentage savings customers received by selecting these options over a 5-day membership. The font used for the percentages was larger than any other used in the options area.

- Graph listing membership features

The older page featured a box on the left-hand-side that listed the three main features available with premium (paid) memberships, followed by the two features available with basic (free) memberships.

For the test page, the team removed this box and replaced it with a graph that let visitors compare basic and premium membership features side-by-side.

The graph was headlined "Membership Features," and listed seven features in vertical rows. Two columns for basic and paid memberships showed check marks illustrating which features each included.

The column for basic memberships only featured two check marks, whereas the column for premium memberships had checks in each of the seven categories.

-> Results: Conversions lift, but not internationally

After incorporating all the winning variables from each test onto the page, the team achieved a 3% increase in conversions.

"So that's direct revenue that comes to our company," Intsaby says.

As with the previous tests, the team made these changes on the other countries' sites. However, the changes did not appreciably improve results outside the U.K. in this case.

Intsaby hypothesizes that the form became less persuasive after translation, and may have confused some visitors since some sites offer different features.

Useful links related to this article

Creative Samples from EasyRoommate's page tests

Members Library -- How to Plan Landing Page Tests: 6 Steps to Guide Your Process

Members Library -- Master the Art of Multivariate Testing: 7 Lessons from Avis Budget Group

Maxymiser: Provided the team's testing tool and aided testing process

EasyRoommate



Improve Your Marketing

Join our thousands of weekly case study readers.

Enter your email below to receive MarketingSherpa news, updates, and promotions:

Note: Already a subscriber? Want to add a subscription?
Click Here to Manage Subscriptions