AI-powered Ratings and Reviews on Cleartrip
HOTELS B2C
Introducing all-new Gen-AI powered Ratings and Reviews on Cleartrip
In 2000, TripAdvisor popularised the ratings and reviews for the hotel and restaurant industry.
Fast forward to 2023; video reels, rich media, summaries, and quick insights are one prompt away. However, the conventional approach to representing the Review and Ratings has remained the same.
Imagine a bite-sized visual insight backed by an in-depth summary of a decision lever to find a perfect stay.
End of conventional ratings and reviews on Cleartrip
Current representation
Travellers compare overall and individual aspect ratings, read a few reviews and book a hotel while considering other personal preferences.
Dilemma
The top 100 hotels on the search result page mostly have ratings between 3.0 and 5.0. Have you ever wondered about the difference between 3.8, 4.2, and 4.8?
An exciting takeaway from our research was that travellers are suspicious of hotels with only positive reviews.
Travellers read the negative reviews in depth to decide on a hotel. Only a few negative reviews of the hotel were good enough to drop the hotel.
Competitor's take on AI summary
Using GenAI to summarise the reviews saves time, but the trade-off of human emotions is challenging. We need more than just dumping a text summary. We need a visually appealing, human-centred representation of the review and rating.
Reimagined the ratings by discovering the missing pieces
1. Overall ratings
First, let’s talk ratings. A considerable, bold number is the first key identifier to shortlist hotels.
Ratings can be the only critical decision lever if they represent three other essential metrics: recency, consistency, and overall sentiments.
Recency: Account for recent experiences, ensuring the reviews and ratings reflect the latest traveller insights and the hotel’s standard.
Consistency: Consider & mitigate variations in a hotel’s performance resulting from management changes and renovations to ensure a consistent & positive traveller experience.
Sentiment: Traveller’s reviews are a rollercoaster ride of emotions — average the sentiment percentage to form the emotion for their hotel stay.
2. Individual aspects rating
The second concerns individual aspects of hotel amenities ratings. Each aspect hides the rollercoaster emotions of the travellers’ experiences. How can numerical representation be made more credible?
Humans have a bias toward validating what they see. What if a quick overview backs individual aspects? We explored various approaches to representing aspect ratings and have added a text summary and sentiment pills to support amenities ratings.
Quality: Provide quality by maintaining rigorous review standards, even if ratings differ from competitors.
Relevance: Value the inclusion of diverse traveller experiences and fine-tune ratings & reviews to match individual preferences.
Trustworthy: Provide an unbiased perspective to build confidence.
3. Persona ratings
The last one is Persona ratings. How is the hotel favourable among personas? We have identified five personas for a rating scale of five: couple, family, friend, solo, and business.
Overall ratings, individual aspects ratings and persona ratings — All three critical pieces of information are part of the Cleartrip Ratings section. On desktop and mobile, it takes two or three scrolls to reach it, which provides in-depth information on the hotel details page.
How do we intrigue travellers to reach Cleartrip ratings section?
Our next challenge is to summarise all this information into bite-size pieces so that travellers can comprehend it in seconds without scrolling.
We call this section Property Highlights, which appears right below the hotel name.
Top Rated: Which is the most mentioned hotel rating aspect and why?
In the Spotlight: What are the top three pivotal highlights of the hotel?
Great for: Who rated the hotel the highest among the traveller types?
Cleartrip Ratings and Property Highlights: Two sections have the necessary decision levers for glances and in-depth reviews. It’s time for visual design.
Beautifying ratings: papers to pixel and motion
It would help to start with apparent ideas to warm up the brain. To clean up the creativity faucet, you must stick with the problem longer and be exhausted with apparent ideas.
You will see the scrapped iterations, which brought everyone back to the whiteboard. These iterations resulted in earlier sketches, which helped us push boundaries when rethinking every pixel of the final version.
Considering the limited real estate and lot to fit in, we have explored mobile first iterations.
We also dig deeper to customise the experience based on the hotel chain’s star ratings. Imagine a different experience based on a hotel’s star category.
Hence, we have tweaked the card layout and label, such as “Value for Money” for 3-star hotels and “Value for Luxury” for 5-star hotels. We also decided to use different visual treatments.
All these iterations led us to the final version, a fusion of ideas
Product launch video
Thanks for scrolling, read detail case study on Medium