Landing page experiment walkthrough

Design and implement a landing page experiment and set up geo-targeted landing page experiments.

Landing pages are standalone pages that serve a specific purpose, such as generating leads or promoting product-specific offers. Organizations run several landing pages at once and target each one to a specific offer or customer segment. Each page is an opportunity to test and optimize.

Design and launch a landing page experiment

Consider this scenario. Andrea leads the digital marketing team at Adler Bragdon Insurance, which sells health insurance in 19 states.

Andrea wants to test whether a discounted rate persuades a visitor to buy two policies at once. Rates and requirements differ by state, so Andrea creates a landing page experiment for each state.

You can use pages and audiences in Optimizely Web Experimentation to create and deploy many experiments at once.

Configure a landing page template

Create a reusable page definition with Pages so Optimizely Web Experimentation delivers the experiment to every landing page that supports the offer. A page definition is a saved URL rule that tells Optimizely Web Experimentation where an experiment runs.

The following steps create a page definition you can reuse for similar landing pages. Reuse it across multiple experiments that target consumers in different states.

1. Go to Implementation > Pages.

2. Click Create New Page. 

   

3. In the Name field, enter a descriptive name, such as offer_1_landing_page.
4. In the Editor URL field, enter the URL of the landing page on the Adler Bragdon site.
5. Select Substring Match from the match type drop-down list.

6. In the URL Match field, enter http://www.adlerbragdon.com/landing/offer_1/. Adler Bragdon stores each landing page in a subdirectory named for the purpose of the page. Substring match therefore applies the experiment to every landing page under that path, which is the set of pages that supports the offer.

   

7. Click Create Page.

Create the first experiment

In this experiment the goal is to find out which variation of landing page works best in each state.

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

1. Go to Optimizations and click Create.
2. Select A/B Test from the drop-down list and name your experiment.
3. (Optional) Add a description. As a best practice, add your hypothesis in the description.
4. Under Targeting, select Saved Pages from the drop-down list.

5. Select offer_1_landing_page, the page definition you created in which you created in Configure a landing page template.

   

6. Click Create Experiment.

Configure the first experiment

Configure the following settings in the experiment you created:

1. In the Audiences tab, create an audience that targets visitors based on geographic location to limit participants in the experiment to only those visitors from the appropriate state. Then add that audience to the experiment. See Target group visitor behaviors for information on audience selection.
2. In the Metrics tab, add a metric to your experiment. Because this experiment tests the effectiveness of different landing page offers, you can track total conversions as your primary metric.
3. In the Traffic Allocation tab, set your traffic allocation to specify the percentage of visitors included in the experiment. You can set up your variation traffic distribution, which controls the proportion of traffic sent to a particular variation.
4. In the Variations tab, create a variation. A variation is a different version of the page where your experiment runs. Optimizely Web Experimentation measures the differences in user response between variations to determine winners and losers. Each experiment needs at least one variation.
5. Test your experiment with the Preview tool and click Start Experiment.

Duplicate the experiment

Complete the following to duplicate the experiment for each state where Adler Bragdon sells policies:

1. Go to Optimizations.
2. Find the experiment you created in Create the first experiment.

3. Click More Options (...) and select Duplicate from the drop-down list.

   

4. Click Duplicate.

   

Rename the duplicated A/B test to include the state it targets. When you duplicate a paused experiment that has no variation changes, add variation code before you start the duplicate.

Repeat the duplication process for each state. Using the Visual Editor, make the necessary copy and content changes one experiment at a time.