Optimizing Landing Pages

How a follow-up test delivered an additional 39% increase in conversions - Part 2


We recently released the audio recording of our clinic on this topic. You can listen to a recording of this clinic here:

Landing Pages 2006 Pt.2

Part 1 of the Optimizing Landing Pages brief is available here

Test 1: Developing a New Control Page

In our last brief, we reviewed a test covering a two-week period in April of 2006. Our primary objective was to determine whether a small number of specific changes to an offer page would improve overall conversion for a company selling hypnosis products.

We applied the Marketing Experiments Variable Cluster Testing Methodology. In this way we were able to test multiple variables with a single factorial “A/B Split Testing” test design.

The optimized page included:

  • Improved credibility factors, including photo, credentials, and signatures. Greater visibility was given to testimonials, and it included a money-back guarantee.

Here were the two pages tested:

We set up an A/B split test to gauge the effectiveness of these two pages. Here were the results after two weeks of testing:

Landing Page A/B Split Test
Overall Improvement in Conversion Rate 40.7%


What You Need To UNDERSTAND: Conversion improved by more than 40% by making a number of key changes to the original page.

It’s important to note that the headline and body copy of both pages were identical. An improved eye path, and improved credibility factors (photo, credentials, and signature) all led to the overall conversion improvement.

In future tests, to improve conversion even further, we would test the page:

  • with and without the photograph.
  • with bold text to strengthen the eye path.
  • with change in the design and text of the “Add to Shopping Basket” button.
  • with a strong postscript (P.S.).
  • with more time-urgency built into the offer.

Test 2: Testing a Photograph

In our second test, we sought to determine whether the removal of the photograph on the offer page would improve conversion. Both pages were identical with the exception of the photo. On the page without the photo the credentials were moved below the signature.

Here are screen shots of both pages:

Again, we set up an A/B split test to gauge the effectiveness of these two pages. Here were the results after two weeks of testing:

Landing Page A/B Split Test
Overall Improvement in Conversion Rate 39.0%


What You Need To UNDERSTAND: Conversion improved by more than 39% by removing the photo from the landing page.

10 different pages were tested without the photo, but only one, on the subject of “shyness” showed any significant change. This may be due to the nature of the audience. Those suffering from shyness did not want to see a therapist looking out at them from the page.

This poses a very interesting question: “Are we too quick to assume that a change made to one page will have the same impact when applied to other very similar pages, but on different topics?”

Next, we would test the current winning page against the optimized page with different photo, revised placement of the bonus offer, and 30-day money-back guarantee.

Live Evaluations: Audience-Submitted Landing Pages

In our recent web clinic, we had the chance to evaluate several pages submitted by clinic participants. Below, we have aggregated the suggestions made by our analysts for three of the submitted pages. We have also included screenshots of the critiqued pages.

Page 1: William Jewell College

Although this page is very clean and professional, we had a number of suggestions for improvement.

  1. The page has no clear objective. Is it to drive new enrollments? Provide information for current students? Solicit donations from alumni? Focus the page around one clear objective and use the elements of the page to direct the visitor.
  2. There is no clear headline.
  3. Although the large image in the center of the page is attractive, it uses up a lot of key page space that could be used to direct the eye-path of the user.
  4. The file size of the page was more than 500 KB. For any visitor on a dial-up connection, the page will take two minutes or more to load. In fact, during the course of our clinic, a number of participants reported that they were not able to load the page at all.
  5. If soliciting contact with new students is the goal, then bring part of the contact form to the landing page itself. Start with the name and email address, which will allow you to follow up with the user if he or she does not complete the entire form.
  6. What are the three most important things that people want to know about the college? Consider addressing these questions on the landing page.
  7. The page copy does not “reach out” to prospective students. Consider addressing them directly.
  8. There is no “voice.” Consider a “letter” format from the president of William Jewell College.
  9. Test the colors. The existing colors may be sending the wrong signal to visitors.
  10. There are too many options. And there are too many graphics. These elements compete with each other instead of guiding the user toward the desired action.
  11. Don’t put a dominant picture in the upper left corner. This is key page “real estate” that should be utilized for a key piece of information. Pictures automatically draw attention and don’t need to be placed in dominant positions in the natural eye-path.

Page 2: RoofSmith

  1. The mechanism “Get 3 In-Home Estimates with One Free Request” is very good. However, you should capture the email address at that point, not just the zip code. Use language such as “Tell us where to send your estimate” instead of simply “Give us your email address.”
  2. The vertical layout from top to bottom is also good. It is easy to scan because of the headlines. The right-hand bar does not overly compete with the main content, so a user’s eyes are directed straight down the page.
  3. In the section, “We make finding a quality roofing contractor easy,” give more emphasis to the “Get Started” button. Make it a different color or give it a drop shadow.
  4. The credibility indicators are excellent, but they could be positioned better near key decision areas.
  5. There are perhaps too many conflicting sections, or they are not presented in the right order. Should a user enter their zip code for estimates, read the three bulleted items, click the “Get Started” or the “Learn More”? Give more emphasis to the desired primary action.
  6. The photograph is on the wrong side. Put the “3 Home Estimates” language here and put the photo where you want to draw the eyes next.
  7. The horizontal lines break the natural eye-path and separate the page into disconnected chunks. Focus more on fluidity leading to one key action.
  8. If you are trying to qualify leads, do not do so on the same page on which you’re trying to capture the most leads possible. Qualify the leads on a subsequent page.

Page 3: The Source (Canada) by Circuit City

Retail stores such as this are often much more difficult to optimize because they are not focused on selling one service, subscription, or product, but rather a large selection of products. However, there are several improvements that could be made to this page:

  1. There needs to be some kind of conversion funnel on this site. If a user comes looking for a laptop, he may have difficulty finding it if he is bombarded with free printer offers, clearance items, and so on. Consider using a configurator that allows a user to select his desired products from a series of pull-down menus.
  2. Allow users to search for items and sort items by price and features.
  3. There are so many banners that they compete with each other and destroy the natural eye-path.
  4. Don’t make the special offers on your own site look like ads. Users have been trained to ignore banners (“banner blindness”). A more effective approach for the free printer offer might be to use a simply picture of the printer on a white background with accompanying text.
  5. The “Click and Learn” section should be given more emphasis. These are the types of things that will help a visitor find the products he or she is looking for.
  6. In contrast to the special offers that look like ads, the rest of the site is a kind of washed-out gray that does not attract the user’s attention. The color strategy of this site should be completely revamped.
  7. There is no person-to-person connection. Talk to the customer, and explain your unique value proposition as quickly as possible. Be specific about what makes this store better than the competition.For starters, give extra emphasis to the free shipping offer. What else sets you apart? Do you have the best service rating or lowest price on BizRate? Tell us so. Do you have the greatest selection of products? Tell us precisely how many.

The first question to ask about any home page or landing page is this: What is this page’s primary objective?

If you have a clear answer to that question you can then look at every element on your existing version of the page and ask: Does this element directly contribute to achieving the page’s objective?

Very often pages have too many competing objectives and elements, which confuse the reader and disturb the natural eye-path.

If you have decided on your primary objective, then build a sequential, vertical communications path from the top of your center column downwards. The main, center column is the most effective space in which to communicate your unique value proposition and lead the reader forward to taking an action.

Finally, be clear about what you are asking your reader to do. Don’t confuse people with too many choices or variations on one choice. Keep your call-to-action simple, and make it compelling.

Related MarketingExperiments Reports:

As part of our research, we have prepared a review of the best Internet resources on this topic.

Rating System

These sites were rated for usefulness and clarity, but alas, the rating is purely subjective.

* = Decent | ** = Good | *** = Excellent | **** = Indispensable


Editor — Flint McGlaughlin

Writers — Brian Alt
Nick Usborne

Contributors — Eric Stockton
Guy Tasaka
Jimmy Ellis

HTML Designer — Cliff Rainer

You might also like

Leave A Reply

Your email address will not be published.