Rentalcars.com – search results: add five-star ratings component
The objective of the experiment was to add five-star ratings to the search results.
When customers give rating on certain items, it is generally recommended to simplify the process so that other users can intuitively know what the objective is. The users want to reward good content and punish bad content. They want to express their opinion and feel that their opinion is being taken into consideration.
The user is emotionally attached to the five-star rating system which is a simple way to capture the entire sentiment: very bad, bad, neutral, good, very good. Its simplicity will make the user observe better and set expectations about a specific supplier. Also it will make the act of rating fun/game-like, as such for a period of time the act of rating is inherently worthwhile.
What I Did
- Problem definition. I identified the problem: the website doesn’t get enough engagement with the rating system. This is because users have a certain mental budget they will invest in rating suppliers. The more work they do for each decision, the fewer decisions they will make. Another consideration is that customers only give feedback if they are motivated to do so. With the previous rating system, the organisation only ranked suppliers without having any visual clues.
- Observation. My process continues with observing analytics and identifying the users who will be using the solution. I found out how users interact with the ratings. From user testing insights I understood how users perceive five-star ratings.
A five or close to five star rating, leads to a higher number of bookings. However, this must come correlated with multiple user reviews. Otherwise the high ratings do not have a significant booking impact on conversion rates. When the user has to make a decision between two cars with equal rates, users are more likely to book a car with slightly better reviews. Also, if the care has a higher rate, the users will still book the car with the better reviews than the one with the better rating.
- Ideation. Brainstorming ideas based on what I found during the previous two steps: observation and problem definition. I made several possible visualisations of star ratings with rating accompanied by color change of the stars themselves – gold, silver and bronze are straight forward.
- Prototyping. I developed several prototype versions of the accepted ideas and also integrated HTML
- Mockups, wireframes;
- Hi-fi prototypes.
Results & Business Impact
The experiment is currently fully implemented. Next step is to add the five-star rating to all the pages in the funnel.
- After being A/B tested with users become a successful experiment with a high clickthrough rate;
- Star rating outperformed previous non-star version on conversion rate with more than 15%.