The goal of the project was to remove stress during holidays by being one step ahead of users and by providing them relevant information just in time.

But first, let’s start with the context: holiday trips.

Holiday trips are always exciting and something to look forward to. However, the process of preparing, traveling and arranging stuff is often perceived as stressful.

Together with Coen van Hasselt, we ideated on a predictive travel experience that would reduce the amount of stress travelers perceive along their journey. Take airports for example. They’re like mazes where a lot of decision-making is involved.

Our goal: be one step ahead of our travelers by providing needed information just in time.

This project to me was a first experiment to ideate on predictive user experiences & anticipatory design patterns. Besides explaining the concept, I’ve also included my observations and reflections.

Anticipatory Design within Travel

The premise behind Anticipatory Design is to reduce stress by making decisions on behalf of the user. The decisions we’re making are basically only related to way finding. The TravelBird Highlight Cards give directions, exactly when our travelers need them.

This way, travelers can look forward to their upcoming trip without having to stress about which luggage band, gate or direction they have to go. The Highlight Cards got their back.

Feedback = Trust

Holiday trips should be experiences of joy. Many travelers fear disappointment. Disappointments regarding hotels, facilities, service or transport. We’ve tapped into this need by introducing a continuous feedback loop that asks feedback, just in time, at every step in the journey.

Think of travelers who are able to feedback about their hotel after they have checked-in at their hotel. This gives them the feeling that they are in control and able to voice their opinion.

Having the opportunity to feedback builds trust and allows us to follow-up pro-actively when travelers are not satisfied with provided services or facilities.

Design Process

We followed the following process to map needs and pain points, to create a first predictive user experience.

1. First, we mapped the journey of our travelers and investigated what pains, gains and tasks they had during a trip.
2. Pains, gains and tasks we’ve found formed a backlog for possible states we needed to anticipate on.
3. We’ve designed an experience map to make the journey visible within our office, including emotional states.
4. Based on the evidence-based experience map, scenarios were crafted/defined covering all pain points along the journey at which travellers had much decisions to take or at which there was much uncertainty.
5. These scenarios resulted in timelines with specific states.
6. Because this design sprint was about testing the concept of anticipatory design, we came up with a rule-based algorithm.
7. Prototyping: we made mid-fidelity prototypes to test on user groups.
8. User testing: we tested several states in short user test cycles.
9. Design & development of concept.

The ‘Highlight Cards’ were a first opportunity to ideate on predictive user experiences within travel. This service is already live and I love to hear what your thoughts are about the concept!

Learnings & Reflections

This project was very insightful and made me reflect about Anticipatory Design & predictive UX in general. About its possible impact and scalability. A few points keep returning in discussions:

1. The importance of algorithms to be more empathic and able to identify meaning behind actions (relevant because of understandability).
2. People want to know why you recommend a certain action or event (this is currently a UI-challenge for us).
3. Too much reliability on anticipated and predictive services: we are distilling knowledge into algorithms so tech can guide us with tasks. With this approach we remove ourselves further and further away from understanding what and WHY we do things.
4. Personality is key for smart AI’s. For us to get meaningful data, people need to use services like Siri so we learn from there queries. The engagement with these smart operating systems is crucial. Therefore, I think we should add more personality to these AI’s so that people can relate more to them.
5. There’s a need and urge to re-imagine ML methods because of current levels of serendipity.
6. Design challenge: how to design for a multi-user automated experience? And how do we decide who’s higher in rank?