Key considerations for a landing page optimization

Does the landing page has a single focused objective?
Pages like home page tend to have all things for all people or addresses various personas. Effective landing page for testing should have a very clear objective, allowing testing to be measured against a very clear single outcome. i.e. sign up, downloads, sales conversion

What is the connection to the upstream ad or traffic sources?
If you’re testing an existing landing page, it is very likely the page is already receiving traffic from some where. Study really hard around what sources are driving the most traffic and try to link that to their personas. For example, you may have a page where traffic is heavily driven by PR efforts such as links from CNet or Gizmodo reviews. If That is the case, expect to think through what your testing means to these audience. If your landing page testing’s goal is to drive sales, and current traffic or personas of the people are clear they aren’t there for shopping, then test different landing page and drive traffic through paid search targeting folks already searching for you product.

Think about post launch effect on content
You may be testing to replace a page or adding a page, but consider the consequences of adding additional page that may cannibalize your organic search equity built from let’s say Google. Google will generally charge you less for a landing page that very closely mirrors the content of an AdWords ad, and the CPC is also dependent on the quality of the page rank and relevancy of your content to search. If your tested page is to add on to currently existing page, watch out from SEO stand point on post launch of your tested winner page. In addition, if a page is way too general or not tight with future ad campaign, that may impact the overall performance.

For example, say you’re testing a page X to add to current repository of pages. Say page X’s goal is to convert generic brand searchers to sales, and it does really well in testing after targeting to branded terms searchers from Google Search. Say you push that page live, and later marketer tries to use that to convert product ABC because they’ve heard the new page A does well in conversion. You page is optimized for “Brand”, but you’re buying keywords on “Product ABC”. Your conversion may not be as good as expected, and CPC may be higher than your plan.

Keep the test simple and measure against ONE goal
Make sure to understand the different testing plan tactics A/B testing, URL split test, or Multivariate Testing (MVT). At the beginning, marketers would want to test many ideas from little button colors, to links, buttons, etc. Don’t expect all test to be successful, good testing comes from greatly thought through planning. Understand the big picture by asking what problems marketers want to solve, what KPI they want to move, etc. Make sure the goals are simply measured against one thing first especially if you’re new to testing. Many successes around many conversions, is just going to raise questions rather than agree on the outcome.

Have a framework or some sort of methodology for critiquing landing page
Here is a very good article from visualwebsiteoptimizer.com interviewing Oli Gardner @unbounce. Setting a test to go through a page prior testing could help marketing heads to be aligned. 5 second test from Oli’s article is a good one, and it goes something like…

1. What is it about?: How obvious is the core brand message?
2. Do I care?: Am I interested? Does it speak to me?
3. Is it trustworthy?: Does the design make me feel comfortable? Do they appear professional?
4. How do I participate?: What am I supposed to do first? Is the primary call to action presented in a clear manner?
5. Is it newsworthy?: Given today’s social web economy, how likely am I to want to share my experience? Is this facilitated in any way?

This rule may not apply to all pages and testing objectives, but it definitely worth the time to have some kind of framework for marketers to go through the exercise to think about landing page optimization.