App landing page - a case study

1. Definition

Ater the food ordering app for both mobile and web was ready to launch, the team aimed to create an app landing page.

 

The objective was to create a very simple page that would highlight the product features; promote the app installation on mobile devices and provide a way to login and signup into the web app.

 

The visual language and the color palette should be consistent with the applications.

2. Ideation

Sketches and wireframes of several proposals were created.

Quick design sketches

And lead me to a final wireframe.

3. Prototype

After the wireframe, a mockup was created using sketch and I coded a prototype using HTML and CSS.

4. Launch

Having the html and css, it was easier for the development team to implement the final solution and the launch occurred.

5. Testing

After launch, I monitored user behaviour through hotjar recordings. The conversion rate was very low (12,2%) so we wanted to test the integration of a CTA to order since we had the suspicion that users weren’t becoming aware that they could actually login and order using this page.

5.1. A/B Testing

Using Google optimize we created two versions of the page. Since the change was so small, we didn’t even used redirect, we added the code of the new CTA button directly.

Bellow we can see the two versions of the landing page being tested. 

Fig 1. On the left hand side version A with no CTA to order and on the right version B with a CTA to order.
5.1.1. Hypothesis

The “Order now!” CTA will increase conversion rate and more users will signup/login.

5.1.2. Variables

Dependent variable: Number of users logging in.
Independent variable: Integration of a CTA button.

5.1.3. Duration of the experiment

Since the sample size was not large enough, due to a low number of visits, I decided to perform the test based on the duration using the current number of visitors. So I used A/B Split & Multivariate Test Duration Calculator by VWO. 

VWO calculator

Based on this information we setup the test to run for 30 days.

5.1.4. Test conclusions

When the test was stopped, we had a very solid and clear response from the system. The conversion rate increased significantly from 12.2% to 31.4%

Test results