|The Power of Small Changes Tested|
|Tuesday, 21 March 2006|
Topic: The Power of Small Changes — How minor changes to your website can have a major impact on your conversion rate.
We recently released the audio recording of our clinic on this topic. You can listen to a recording of this clinic here:
Here is the question we asked ourselves: In our focus to maximize increases in conversion rates by completely changing key site pages, are we ignoring the potential of small changes?
Often, with the best intentions, we apply all we have learned over the years, and make significant changes to a site page. Frequently the results are very encouraging.
However, once we have optimized a page, are we still leaving money on the table?
Once we have achieved the "big increase," is there a danger that we might sit back and feel the page is now fully optimized?
Our testing suggests that even with a highly-optimized page, it is still worth following up by testing additional, small changes.
Test #1: Offer Copy
In our first test, for a large online publisher, we made only one small change in a single sentence near the order button. It had the following impact on conversion rate:
What You Need To UNDERSTAND: Page B improved overall conversion from 1.58% to 1.77%, an increase of 12%.
What was the simple change that resulted in this 12% gain?
Page A had this copy:
Page B included these simple changes:
The words "Free" and "Risk Free" in each page were bolded to highlight the key benefit.
By changing just a few words of your offer copy, you can have a positive impact on your conversion rate.
KEY POINT: When testing small changes, pay special attention to key decision points and significant points along the primary eye-paths.
Test #2: Page Layout
In our second experiment, for the same company, we tested two pages in an eight-day micro-test.
Here are the results of this micro-test:
What You Need To UNDERSTAND: Conversion of Page B (one column) was 88% better than that of Page A (two columns).
Simplifying the page by combining two columns into one resulted in a surprising gain.
KEY POINT: Not only text, but simple changes in page design elements can yield a significant increase on conversion. We will suggest a number of areas for improvement in the final section, below.
Test #3: Small Changes in a Multivariate Testing Environment
Our first two tests were simple A/B split tests. Multivariate testing allows you to test a number of small changes at once. For an introduction to these topics, see our recent reports:
In a third experiment (for a different company), we focused on testing a number of small changes simultaneously using multivariate testing. We ran a micro-test focusing on the following page elements:
This is an ongoing test, but results were gathered after the first eight days of testing. Below are the results for each variable in isolation.
We tested three subscription forms:
What You Need To UNDERSTAND: The "Standard" form out-pulled the more colorful variations. It performed 28% better than the next best form.
We tested six different headlines:
What You Need To UNDERSTAND: The fifth headline ("Identify Registered Sex Offenders Living Near You") converted 34% better than the next best headline.
We tested three different logos:
What You Need To UNDERSTAND: Logo C (silver eagle) out-pulled the next best logo by 19%.
Taken together, these three small changes amounted to a combined 104% increase in conversion over the next best set of variables. This is the compounded gain of the three small (cumulative) changes taken together.
KEY POINT: A good multivariate testing platform will allow you to test multiple small changes at once and optimize your site more quickly than testing each change in isolation.
Small Changes to Test:
Above are examples of some of the small changes you might test on your own site. But don't stop there. Following is list of other small changes that can be systematically tested on your own site; through individual A/B splits, or together in multivariable test groups.
Finally, you can measure the cumulative effects of multiple changes made over time with our spreadsheet tool:
RELATED MEC REPORTS:
Editor — Flint McGlaughlin
Writers — Brian Alt
Contributors — Aaron Rosenthal
HTML Designer — Cliff Rainer