Psychology of rating systems in digital interfaces
Uber, like other consumer services, has an interesting rating system for both passengers and drivers. Arjun leads the passenger rating system with 4.91 points, not because he has developed some kind of strategy, but because he is a pleasant conversationalist!
He recently noticed that one of the drivers gave him 1 star. Arjun was taken aback. Out of curiosity, he asked the driver why the rating was so low. The driver smiled and explained, “Sir, this is not a low grade. I gave you the # 1 rating! “
Should we stop using stars?
Earlier in 2017, Netflix took a big step away from star ratings, replacing it with thumbs up and thumbs down likes and dislikes to suggest possible matches based on your past likes.
YouTube introduced a similar system almost a decade ago, in 2009. When it comes to ratings, it s almost always all or nothing.
They motivated this by the fact that the vast majority of YouTube videos had a five-star rating. And user reaction was at two extremes. Either the video was great or pathetic. For the rest, they didn t care about reactions and ratings. It should be noted here that the average YouTube video rating was common across all users. Five red stars on Netflix means a movie or TV show is perfect for you. The rating you see next to every movie or TV show (on Netflix) is average among your “like-minded people “, not all users, as on YouTube.
How often do people rate?
Previously at Uber, it was almost mandatory to rate your driver before moving on to your next ride. Now this is not necessary, as users miss the rating quite often. In 2014, Uber (San Francisco) sent out a guide to its drivers explaining the principle of a driver rating system. Additionally, if the driver s rating was 4.6 or lower, Uber might consider deactivating their account.
“Deactivating the accounts of drivers who consistently provide a bad experience ensures that Uber continues to be known for its quality.”
Uber drivers depend on a good rating for their earnings. Just like the attendance of restaurants depends on their rating. The earning of many products and services depends on high ratings.
Do all people understand ratings the same way?
Probably not. To better answer this question, you need to understand the purpose of rating systems, especially in modern digital products.
What is a ranking system in the UX world?
A rating system is a possible investment by your users in your product (digital or otherwise). Once users understand the benefits of this investment, they are more likely to interact with your rating system. In this way, they will help with product growth by encouraging good actions and punishing bad ones.
Last year we conducted a survey for one of our client s projects. Over 50% of respondents said they would not watch a movie with a score below 7 on IMDB!
Interestingly, the LAR (minimum acceptable rating) for a show or series is 8 (not 7 like movies). “I almost never give a movie more than 8 points because it has to be a perfect movie,” said one respondent. While a 7 is acceptable, a 9 is too rare and exceptional for films. Those same users will give the same movie a thumbs up on Netflix.
Platforms, their use, rating systems and their possible engagement outcomes seem to play a large psychological role in how a user rates content. Colors, labels and the effect of the rating itself are also major influencing factors.
About 80% of respondents in the same survey say they call Uber at least once a week, with more than 30% driving every day. Almost 50% of these passengers will give up the ride if the driver has a rating below 4.5 (out of 5)!
The same users view the listings on Zomato at least once a month. The minimum acceptable rating for a restaurant drops to 3.8 (out of 5)! Perhaps the frequency of use and relevance in daily life also influences our perception of ratings.
“I give 4 stars when something meets my expectations. And I give 5 when it exceeded expectations! “
Likes on Instagram (or Facebook) and retweets on Twitter are also rating systems, in fact, real binary systems. They value two extremes – love or hate!
Facebook, Twitter and Instagram don t care about hate events. They value and promote the experience that users love. Binary systems make it easier to get ratings because users only have to act when they like something on these platforms.
Group norms and subordination
In 1936 Muzafer Sheriff conducted a classic experiment. The participants were placed in a dark room and asked to observe a small point of light 15 feet away and rate how much it had moved. The experiment showed that the responses of the participants tested individually differed greatly. While the participants tested in groups of three converged on an overall rating. Sheriff s experiment showed that instead of making individual judgments, people always tend to obey the group.
An Instagram user who saw that a certain post on 9gag already has 559,031 likes agreed to add another like. Similar subordination holds for ratings, stars, votes and other systems if the sum (or average) of all ratings is shown before the rating act. Perhaps this factor is one of the key factors behind the concept of virality in social networks.
Cotton for grading
The binary equivalent of the “Like” button on Medium, the “Recommend” button (heart icon), was redesigned sometime in mid-2017 to “Cotton”. A user can clap an article up to 50 times, with 0-50 claps rating its attractiveness (or quality …), which is perhaps the equivalent of a star rating system.
While you can like your own Facebook posts, Medium doesn t want you to “clap” yourself.
One intriguing difference between star ratings and pops on Medium should be noted – the visual unavailability of possible counts. The star rating asks for an estimate within 5 points. While the number of claps asks to estimate up to 50.
Many fear that this change in the rating system in Medium will lead to an increase in the “rating rate” of the platform. For an article that brought about 2,000 recommendations, even 20,000 claps now seem less.
Delving Deeper into UX Rankings!
Some rating systems suffer from an often overlooked error – averaging concepts…
For example, in addition to a critical number of ratings, a 5-star rating for a service (or product) will be averaged to a certain number, say 4.3. From now on, in most scenarios, given that ratings continue to rise in good numbers, it will take a significant number of extreme ratings (1 or 5) to increase the average rating from 4.3 to 4.4 or downgrade to 4.2. With a 5-star system, the rating is 4.3, after N number of ratings, it becomes the average rating!
New ratings may not have a significant impact on this rating, not allowing you to see real feedback.
The user experience of the Uber Driver app forces the driver to rate each passenger at the end of the ride. However, passengers do not need to rate the driver. Likewise, Zomato and Amazon made it optional for users to add a post-purchase review. In fact, Amazon allows you to leave a review for a product even if it was not purchased on the platform if the lowest acceptable rating of the requested product drops to 2.
“I know there is a lot of bullshit in the reviews on Amazon, so sometimes I bought things with a 2 or 3 star rating and I was fine with it.”
The user s personality, mood, environment, urgency, potential satisfaction (and its conditional value) all of these factors are of great importance for the assessment. We also found that users quickly rate applications on mobile phones due to the ease of use. 74% of users prefer to rate on mobile phones over other devices.
While some users choose not to rate apps or related services, when you need to write a review as well.
The future of rating systems
The series “Black Mirror” in one of its episodes – “Dive” (the first episode of the 3rd season) tried to reflect the psychology of rating systems. In this episode, users were able to rate all of their interactive and personal interactions on a five-star rating system.
Everything from status in society to access to certain services and employment opportunities have been a factor in a person s current ranking.
We have already become a generation of critics and received the ranks of managers (who are paid little). We observe every movement of the waiter in the restaurant, evaluate the quality of the ringing that the spoon makes when he puts it on our table. We evaluate the level of politeness when he agrees with the choice of the dish, and then move on to the next “draft” for criticism.
At the same time, the use of a digital product has reached new heights, and the concepts of user experience and gamification are defined by psychology. Therefore, rating systems inevitably score points. I hope this gives you an idea of how they work.
We hope this article will help you improve the UX of your mobile app or web platform and increase user engagement.
This post was originally written Himanshu khanna, CEO and Chief Designer at Sparklin.com. If you liked the article, have an opinion, or want to have a deeper discussion on UX, branding, marketing, or anything else related to design, write to us on Twitter!