CHALLENGE

Christine Mykota, Manager, IT and Americas Marketing, Business Objects, had a software trial download landing page that was underperforming relative to her other pages. She knew that a landing page test could help boost those conversions, but she had to clear two hurdles.

First, she needed to convince her bosses of the potential return on investment when asking for a budget increase to perform the test.

Second, she knew that corporate website design standards placed tough restrictions on layout and design, which limited her team’s testing flexibility.

“I might not necessarily agree with branding standards from a lead generation perspective, when design is important,” says Mykota. “I didn’t realize it was going to be as restrictive as it was.”

CAMPAIGN

Mykota and her team simply worked within the company’s brand design guidelines to develop variations on the Crystal Reports 2008 trial download page. That desktop report-generating software is a relatively low-cost, high volume product, so Mykota’s team targeted a 20%-25% increase in conversions to justify the investment in the test.

Here are the steps they took to develop two new page variations and collect data on conversion rates:

Step #1. Examine brand design guidelines to determine testing options

The team studied the company’s design guidelines to discover which page elements they could change, and which aspects were off limits. Among the rules that limited their flexibility:
- Pages had to include a standard, corporate banner at the top, which took up valuable real-estate.
- Pages could only employ a two- or three-column design, with standard widths for each column.
- Only certain fonts and font sizes could be used.
- Design elements had to stay within an approved brand color palette.
- Images could only come from an approved list.
- Buttons must use approved colors and sizes.

Step #2. Develop hypotheses for changes that could influence conversions

Those restrictions forced the team to think creatively about new designs that could improve conversion rates. They evaluated page elements according to several criteria that influenced a prospect’s actions, including:
o Value proposition
o Relevance to the traffic source
o Clarity of the presentation
o Anxiety/distraction
o Urgency

That process identified several factors the team targeted for improvements:
o Value proposition
o Clarity of the presentation
o Distractions on the page

Step #3. Create two page variations

*Variation A: Stronger call-to-action*

The team’s first landing page variation employed changes intended to reduce distractions and emphasize the call-to-action. In certain instances, they had to receive special approval from the corporate branding team for ideas that pushed the envelope of company’s design standards.

Significant changes included:

- Improving the prospect’s eye flow on the page.
Sticking with the corporate design standards, the team employed a two-column design, but changed the position of the narrower column from right to left.

This move allowed them to place the product hero shot on the left, where it would attract the eye and lead prospects directly to the headline in the center of the page. Although they wanted to try an alternative image, the team decided that change would violate brand design standards.

Instead, they received approval to make the existing image slightly larger. They also added a caption link just below the image that took visitors to the trial download registration page.

- More relevant headline.
The team wrote a new headline that reinforced the message used on the search ad that brought visitors to the landing page – “Free Crystal Reports 2008 Trial Download.”

- Consolidating information links.
The team reorganized additional links that were present on the control page to reduce distraction from the trial download call-to-action. Links for customer support, developer resources, training, and additional information about the product were moved to the bottom of the page, below the primary call-to-action.

The team also moved a box containing links to facilitate immediate product purchases, either online, through the telephone or through a reseller. That box was moved from the upper right of the page to the left side, just below the product hero shot.

- Add large, orange download button.
The existing page’s call to action was a text link inviting visitors to download a trial. The team believed that design was too small to attract visitors’ attention, and wanted to use a large, orange “Download Now” button.

The proposed button was larger than allowed in the design standards, however, with a color that pushed the limits of the approved palette. So, Mykota had to push for the change with the brand marketing team.

She presented examples of landing pages the company had used in the past that featured orange buttons. She also reached out to new colleagues at SAP, which had recently acquired Business Objects, for past testing data that showed the effectiveness of large, colored buttons.

“They do more rigorous testing, and I was able to get evidence to pull out and forward to our brand people.”

The brand marketing team approved the orange button.

*Variation B: Registration form on landing page.*

Both the control landing page and Variation A required prospects to click through to a registration form to receive a software trial. The team wanted to test whether removing that extra step would improve conversions. They built a page that featured the registration form immediately on the landing page.

The headline, graphics and layout of Variation B remained the same as in Variation A. But instead of a large orange button on the top of the page, the team added the registration form just below the headline and a brief product description.

The company’s standard registration form had more than a dozen fields for prospects to fill out – along with several qualifying questions they had to answer. Mykota’s team created a design with a shorter registration form – to no avail.

The brand marketing team denied their request to swap the existing form for a new one. “So many folks use that form, there was no way we could change that.”

Step #4. Test new pages against control with live search traffic

With the new designs set, the team conducted tests with traffic from Google search advertisements.

- Traffic was randomly divided between the three landing pages.
- Team tracked conversion rates (software trial downloads) on each page.
- Tests were run for 17 days to generate reliable data.


RESULTS

The test produced a clear winner, and helped achieve an even bigger lift than Mykota had targeted:
- Variation A, with the large, orange button, boosted conversions 32.5% over the control page.
- Variation B, despite using the long form, also delivered a 17% conversion lift over the control page.

“I was very, very surprised,” says Mykota. “I knew that we would have improvement, but I also knew that our branding restrictions and long form presented challenges.”

Even better, the test achieved a 154% return on investment, convincing Mykota’s previously skeptical boss. “Now she’s a champion within the organization to look at landing page testing and to do more of it.”

Based on the test results, Mykota’s team implemented similar design changes to other landing pages. In six weeks, they were able to achieve a 26% lift in conversion rates across the board.

Now she’s confident she can work with the brand marketing team on further tests that allow for flexibility within design standards – as long as the data can demonstrate an improvement.

“I’m very confident that we can effect change from the grass roots -- we’re seeing it already,” says Mykota. “I’m sure the branding team is very appreciative of any information they can garner to help tweak and improve the brand.”

Useful links related to this article:

Creative Samples from Business Objects' Landing Page Test
http://www.marketingsherpa.com/cs/businessobjects/study
html

WiderFunnel Marketing Inc. helped develop the landing page variations and conduct the tests:
http://www.widerfunnel.com

Business Objects
http://www.businessobjects.com/