Personalized Publisher Content

All partner names & content have been anonymized for privacy.

Team

I acted as the design lead on this project. I collaborated closely with a team of data scientists, platform engineers and a product manager.

Goals & Needs

Currently, on loyalty apps, content is very static. Different users are shown the same offer to the same brands. There’s an opportunity to personalize this content based on user’s interests and past purchases. The main challenge with this problem is that many publishers aren’t willing to give up prime real estate on their apps until they see a real value.

Users

There are two types of users to solve for. The end user who downloads the apps, and the loyalty publisher.

The loyalty publishers goals are:

1. Higher engagement

2. Increased revenue

3. Control over their app’s experience

The ultimate goal of the user is:

1. Get deals on brands they’re interested in

Discovery

Many loyalty publishers have a dedicated “promoted” or “featured” section in their apps, which is prime real-estate for displaying a personalized offer. The challenge is, brands pay publishers big bucks to be featured in those areas. Button needed to prove to these publishers that using our personalization content adds more value than their static placements.

Competitive Analysis

When doing competitive analysis, we looked at other loyalty apps that have promoted sections on their home screens. Here are some examples of that:

User Research

In collaboration with the PM, we created a list of hypothesis and assumptions of this product:

• Users would convert higher if they’re shown a personalized offer at a brand they shop at frequently

• Users would convert higher if they know an offer is “exclusive” or personalized to them

• Users would convert higher if they’re shown an offer to a product they’re interested in

In order to give users a personalized offer, we needed to work closely with the data team to create a “decisioning model” that would take into account a user’s interests and serve a brand offer based on a set of backend logic.

Wireframes

For MVP, we had three Publishers agree to allow us to test a personalized offer on their app in order for us to measure success. We needed to test out a few different designs based on what users would be interested in seeing. Here’s some wireframes of ideas we brainstormed:

Prototype

After getting publisher approval of designs, the next step involved building the personalized offer card. During this phase, I acted as a front-end developer on the project, since the team did not have a dedicated engineer for this work. In the code, we factored in the decisioning model logic to display a dynamic brand and special cash-back offer.

Testing

We first did some user testing through Validately to get feedback on our designs. We wanted to understand if:

A) the offer details, designs and messaging were appealing to the users and

B) what they expect to have happen post-tap.

We went through several rounds of this until we were satisfied with the final design.

When we were ready for release, with the help of the data team, we introduced a series of variation tests to determine which type of offer would provide the highest value. These variations had different user preferences in their logic (i.e. users who have seen an offer before, users who have tapped on an already seen offer). In the beginning, our variation tests were performing 8-10% tap-through-rate lift. After a few rounds of learning, we tweaked the logic of the decisioning model and saw a 25% TTR lift on our content.

Impact

Next Steps

We are working with several new Publishers to release additional personalization offers within their apps. We’re looking for new ways to introduce personalization that looks past just a standard offer.

Additionally, we’re in the process of releasing an A/B test to understand the effectiveness of different messaging and creatives in boosting tap-through-rate.