Landing page optimization: a hypothetical use case in Optimizely Web Experimentation

  • Updated

Relevant products:

  • Optimizely Web Experimentation
This topic describes how to:
  • Design and implement a landing page experiment
  • Set up geo-targeted landing page experiments 

Landing pages are standalone pages that serve a specific purpose, typically lead generation or to promote product-specific offers. Organizations often have several landing pages live at once, each targeted to a specific offer or customer segment. Because of these factors, landing pages offer opportunities to experiment for optimization.

Design and launch a landing page experiment

Scenario: Andrea is in charge of the digital marketing team at Adler Bragdon Insurance, a company that sells health insurance in nineteen states. She wants to test a discounted rate if a visitor buys two policies at the same time. However, the specific rates and requirements are different for each state, so Andrea creates landing page experiments for each state Adler Bragdon operates in.

Optimizely Web Experimentation provides the tools needed to create and deploy many experiments very quickly by using pages and audiences.

Step 1: Set up a landing page template

Create a reusable template—named pages in Optimizely Web Experimentation—that tells Optimizely Web Experimentation to deliver this experiment to all the landing pages supporting this special offer.

Make sure that at least one landing page is coded and added to the Adler Bragdon website because setting up a page in Optimizely Web Experimentation requires a valid URL; without one, you cannot use the Visual Editor to make content changes for each variation.

The following steps describe the process of creating an initial page that you will later re-use for a series of very similar landing pages. These landing pages will exist in multiple experiments aimed at consumers in different states.

create-landing-page.gif

  1. Go to Implementation > Pages.

  2. Click Add a Page to open the Create New Page modal. 

  3. In the Name field, give your new page a descriptive name, such as offer_1_landing_page.

  4. In the Editor URL field, type the URL of the pre-existing landing page that already exists on the Adler Bragdon site.

  5. Select Substring Match from the match type drop-down menu.

  6. Use the URL Matchfield to tell Optimizely Web Experimentation where to find the landing pages where the experiment should run. All the landing pages on Adler Bragdon's site are kept in different subdirectories stored within the landing directory. These subdirectories are named according to the purpose of the landing page. Because Andrea wants Lars to apply this experiment to all the landing pages supporting this special offer, the URL Pattern should look like
    http://www.adlerbragdon.com/landing/offer_1/

  7. Click the Create Page button to create the page.

Step 2: Create the first experiment

In this experiment, you want to know which variation of each experiment works best in each state, and are less concerned with comparing results in Florida to results in Michigan, for example.

After you create an experiment for one state, you can duplicate it for each of the other states. From there, you can use the Visual Editor on each variation to make the necessary adjustments to the copy and layout for consumers in each state. See how to create an experiment for more details.

You can also click on any of the links in the instructions below for more information on that step of the process. 

  1. First, navigate to the Experiments dashboard. Click Create New.

  2. Select Experiment from the dropdown and name your experiment.

  3. Under Pages, click the + next to the pages you want to include in the experiment. Because you are creating a landing page experiment, you’ll add offer_1_landing_page.

  4. Under Audience, you would ordinarily choose an audience for your experiment. For this experiment, you should limit your audience to visitors who live in the state this experiment is targeting. That will require creating a new audience before you can add it. So skip this step for now.

  5. Add a metric to your experiment. Because this experiment is testing the effectiveness of different landing page offers, you should track total conversions as your primary metric.

  6. Set your traffic allocation to specify the percentage of visitors who should see the experiment. You can also set up your traffic distribution, which controls the percentage of experiment participants who see each variation.

  7. Create a variation. A variation is a different version of the page where your experiment will run. Optimizely Web Experimentation measures the differences in user response between variations to determine winners and losers. Each experiment needs at least one variation, though experiments with at least five variations tend to get better results.

  8. Finally, test your experiment with the Preview tool, and publish it by clicking Publish. 

Create an audience

To limit participants in the experiment to only those visitors from the appropriate state, create an audience that targets visitors based on geographic location. Then add that audience to the experiment.

Find out more with this article on creating audiences in Optimizely Web Experimentation.

Step 3: Duplicate the experiment

The final step is duplicating the experiment for each state where Adler Bragdon sells policies. 

  1. First, click the Experiments icon in the left sidebar. 

  2. Find the experiment you want to duplicate.

  3. Open the context menu on the right and select Duplicate from the drop-down list. The Duplicate Experiment modal opens.

  4. Select the current project name from the Select a Project drop-down list. 

  5. Click Duplicate. Optimizely Web Experimentation will duplicate this experiment.

Re-name this duplicate version to include the name of the state it will target. 

Repeat the duplication process for each state. Make the necessary copy and content changes one step at a time.

For some suggestions on optimizing your landing pages, see this article on landing page best practices.