12th Dec 2019
How might we...
...help people understand, and have control over, the ads content delivered to them
Making feedback a fair, and fun, exchange
Ads-driven services take account of a great many complex signals for delivering ads to people. How can services create meaningful experiences that help people understand and control the algorithms that deliver ads to them and how their data is used to make those choices?
Product Context
Friendlee is a social app that displays targeted ads. Through its feedback function people can gain insight into the AI behind the ad targeting and can provide detailed feedback on their ad experience which is valuable for the Friendlee platform. Through a playful interaction people are invited to plant a tree in return for their feedback or have an ad-free day. This is a way to display Friendlee’s commitment to do right, in a real-life context, when they get it wrong!
Problem & Opportunity
Friendlee is a social app that allows people to organize their friends into groups to share photos, videos, messages, and attend events. Every Friendlee account is personalized, ranking content based on the profile and behavioral input from the person using the app. Friendlee has a news feed that shows content and targeted ads based on what they like.
Friendlee takes account of a great many complex signals for delivering ads to people. The challenge for Friendlee is to create an experience that helps people understand and control the algorithms that deliver ads to them and how their data is used to make those choices.
Design Features
Gaining insights on why ads display
When an ad appears on a person’s feed, they can touch a ‘banner’ that shows why this ad was displayed. An overview of the most relevant triggers for the ad is shown as a starting point for the person to explore these in more depth.
Additionally, Friendlee is transparent when it doesn’t have a clean or clear reason that an ad is shown. In reality, some signals and sources are so complex across the internet that Friendlee may not know exactly why this particular ad was shown.
Design Features
Get something back for feedback
Friendlee understands the effort needed to provide feedback about ads, therefore they try to give something back in exchange for the feedback which is of course valuable to train their AI systems. When a person submits feedback about an ad, which in turn helps Friendlee build better advertising profiles, they receive an option to go ‘ad-free for a day’, or to have Friendlee plant a tree as a friendly gesture.
Next steps
One thing to consider for Friendlee is how to avoid gaming the system. It would be possible for a person to submit random feedback to an ad each day to enjoy an ad free experience, or plant a tree. For the latter the worst outcome wouldn’t be so bad as a few extra trees would be planted!
Additionally, Friendlee can explore how more detailed options can be offered to people to gain insights and tweak ad settings in their personal profile. They could be assisted with options to update their profiles and inferred interests as inputs for ads, or just blocking the outputs of certain topics.
How might we build on this concept to...
- ensure that ad feedback remains a positive experience?
- ensure a fair value exchange is maintained in displaying ads and providing feedback on the ad experience?
