14th Nov 2018
How might we...
...Provide feedback on data inferences drawn from online behaviors
Letting people have the final check in their digital mirror
Offering people easy ways to review their online profile and how it impacts recommendations.
Product Context
Suppa is a food delivery service that serves meals at home by collecting and delivering food from restaurants nearby. Suppa includes a few social features around food that allows people to recommend places they've liked.
In order to provide the service, Suppa is powered by different types of data, including:
- Location data, which is used to connect people to their nearest food retailers. Aggregated data is shared with third parties, providing vital revenue to the company.
- Suppa periodically reminds people of their history and content of deliveries, and how this information is used for ad targeting & discounts. Control of this information can also be accessed at any time in settings.
Problem & Opportunity
Can we provide people with greater control over their data history and how it is used to provide personalised marketing offers ?
Suppa's Team in the middle of the ideate phase
Design Features
Putting technical processes into a more human, contextual format
By adding structure to the generic listing of past actions you'd expect in a profile record, this interface surfaces 'who they think you are' and 'why they think so', demonstrating what ads and discounts they believe are relevant for the user.
The ways in which data records are processed in the background are presented in a relatable and contextually relevant manner through food profiles. This human format helps users to decipher app mechanisms by explaining how profile building works through the connection between the data collected - the type of food, geolocation, favorite moments of the day - and the conclusions drawn from it, so that a Pizza lover regularly has pizza while an East London Expert has made multiple orders from a specific area in town.
Design Features
Did we get this right? Editing inferences through instant controls
In addition to being transparent about how data records are used, Suppa lets people have a say in whether they got their inferences right or wrong. Hitting the right balance between transparency and control without overwhelming a user or hindering experience through information overload and time-consuming interaction is no easy task. Here, all that is required from a user is to rank the type of food they like best through a simple drag interaction. This time-efficient solution requires no additional control panel, seamlessly blending with the interface but nevertheless allowing the user to know what is done with their data. It takes people less than a few seconds to make edits, allowing them to change their mind at any time as their preferences change.
Next steps
People want services that are efficient and relevant to their needs. While making assumptions about users can make their lives easier if they're right, they generate frustrations when they differ from reality. However, it shouldn't be necessary for people to have deep technical understanding to benefit from inferences. Building simple but efficient ways to review what data is gathered to build user profiles and how they impact experience is key to the long-term success of a service.
How might we build on Suppa's ideas to...
- Strike the right balance between inferred profiles and user opinion
- Build similar features around more sensitive data sets.
