You are here
Home > e-commerce > Top 3 reasons why it’s time to kill the stars! – Part 1

Top 3 reasons why it’s time to kill the stars! – Part 1

Would it surprise you if I say that the very first system for rating originated in 1820?

In 1820 exclamation points was used to indicate works of art of special value. This was the first form of rating. Murray’s Handbooks for Travelers and then the Baedeker Guides (starting in 1844) borrowed this system, using stars instead of exclamation points, first for points of interest, and later for hotels.

The earlier rating system was used to inform users about the availability of a service or to set standard. Let’s take star rating system for Hotels as example.

Initially, a hotel rating system indicated the amenities available, like AC, swimming pool, restaurant etc. This was simple to arrive at and simple to assign stars. The critics / experts were involved in rating. The ratings were represented with stars, mostly

For the last 170+ years the star rating system has undergone many changes and has also remained the same.

Later came the rating system in which parameters beyond amenities, like ease of access, location, room view, etc, got included. Again, the critics / experts did the rating and stars were used to represent the rating.

Then came the internet, e-commerce, online services, apps etc.  In addition to the rating,  reviews came into existence.  And stars represented complete experience. A bad room service, a missing laundry, a good billing system everything added to the good or bad rating & reviews. And again, the rating & reviews were represented by stars.

Depending on the nature of service / business the rating system gathers user experience.  Ola captures driver issues, driving issues, etc when a user does not give a 5 star rating. Same goes with Swiggy or Zomato, they capture delayed delivery and delivery person related information.

There is no denying that a lot of innovation has gone into the star rating system.  The average rating was advanced to a weighted average. Machine learning algorithms were introduced. The rating parameters have changed, the people who rate are now users, the reviews are considered into the rating system, yet, the consolidated rating is represented by 5 stars.

Here are top 3 reasons to kill the star.

1.   It is no longer relevant.

Earlier days when we see a 3 star rated hotel one knows for sure that it does not have a swimming pool. Today if we see a 3 star on Makemytrip or Tripadvisor we do not know why.  (It could be a 4 star hotel with a 3 star rating ?)  To know why we must spend time on the reviews.

The stars by itself cannot be a complete representation.

2.  It does not fit the scope

Even though we have moved towards an experience based rating system, the representation is still primitive. Even today most rating systems use the average method to represent the ratings.

Hotel (A) gets 3 stars for its amenities & food & 5 stars for proximity to city.

Hotel (B) gets 3 stars for being away from city and 5 stars for amenities and food.

Both hotels may end up with same number of stars as it is based on average method. If a weighted average is used then we may see a small variation.

Earlier the rating had a renewal process, now most online ratings live forever. 5 star ratings given by users 2 years back may not have same relevance today.

Since the user who rate are in huge numbers, specially, in online market places the relevance of rating gets diluted due to the age of rating & average method system used for rating.

Companies like Amazon had made changes to its rating algorithm, to include age of rating, use weighted average, etc.    Amazon & Flipkart also collects feedback/ rating (seller) & reviews (product) separately and represents the cumulative rating in stars.

 But is this good enough? What about the other players?

3.   It is deceptive

Consumers make a buying decision based on rating and reviews. Even with all issues with the rating system consumers trust the stars to buy products. If you look at the ratings system currently.

–        Only some services like the food ordering, cab booking etc collect ratings during the next order. This is possible as the apps / services are used frequently.

–        For other services like hotel booking, online purchase, the feedback / reviews are obtained through e-mails or popups post service usage or product purchase. The seller requests for a rating, the website requests for rating, but how many would stop and submit ratings & write reviews? I am sure the percentage is very less.

–        But when the service / product is bad the customer will surely return to submit a review. General human tendency is to vent out the negative emotions like anger / frustration immediately.

 Hence the ratings & reviews is mostly emotional representation of one’s experience.

–        Generally, the total number of ratings and reviews are displayed without considering the total number of purchases.   If the product has 50000 purchases and 5000 customers have given star ratings of which 3000 are 1 star do we take the product is bad or the experience is bad? What about those 45000 customers who have not rated?

–        The fact is approx. 2-5% of buyers rate products. The seller feedback / reviews may be higher but not more than 15%.

–        All the ratings / feedback & reviews is averaged to show the star rating.

So is this rating accurate enough to make buying decisions? The star representation is skewed. Hence it is not the right representation and it is deceptive.

Don’t you think it is time to kill the stars. It’s time for the large players in e-commerce space to formulate new methods to arrive at a new pattern to represent the ratings for products & services.

What are your views?

Can you visualize some mechanism which will help the rating system become more accurate?

We will continue with this discussion in the next part.  I will touch upon the other forms of rating systems existent and its pros & cons. I will also do some loud thinking around the possible improvisations into a rating system.

Keep a watch.. See ya..

It is fair to share..