28 Jun Reputation Inflation: The Problem With Fake Star Ratings
When shopping on Amazon do you ignore the items that have no reviews? I do. I’m not willing to be the person who risks getting a crappy product (if it arrives at all). The same is true for products with only five star reviews, especially if they are devoid of useful information, short, or full of grammatical errors. Reviews like, “great product,” “It arrived quick,” and “It is best,” are immediately suspect. Part of the problem is reputation inflation with higher ratings that are motivated by something other than the quality of the product.
The same is true of negative reviews, or reputation deflation. Whether from unrealistic expectations or an intent to purposely cause damage, there are unwarranted negative reviews too. Reputation inflation is more than useless. It masks problems, ruins data and feeds a culture of fear.
Can a 4.6 Get a Lyft Driver Fired?
Lyft drivers say that if they receive too many reviews below 4.7 stars they could be terminated. According to this BuzzFeed article and interview, drivers even put signs in their vehicles for passengers to read that makes this clear. This is an attempt to get better reviews, and thereby keep their jobs.
That article was from 2107. Fast forward to 2023 and a report from Rideshare Drivers United (RDU) and Asian Americans Advancing Justice – Asian Law Caucus (ALC). Their data was based on surveys of over 800 Uber and Lyft drivers in California, two thirds of which were ‘deactivated’ (essentially app-speak for ‘fired’). It revealed many disturbing issues, but for the purpose of this article it highlighted problems with the review system.
Poor ratings, just like inflated ratings, can be due to personal biases not the quality of the service or product. For example, 50% of drivers surveyed who had experienced racism had a customer file a complaint against them.
In a workplace like the gig economy, there’s no manager or other representative to witness the work. The deactivations are fueled primarily by reviews and complaints.
There’s also no one to intervene. This means drivers and riders are potentially subject to poor treatment without any recourse other than the rating system. It also means that some people are afraid to leave honest negative reviews. This is partly because riders are rated too and afraid of not getting picked up.
Uber and Lyft aren’t finding out about details that could enable them to improve problems because they don’t know what they are. They’ve created a model that avoids the issues rather than provides opportunities to solve them.
Reputation Inflation is Everywhere
The problem goes beyond this industry. It’s everywhere. People are afraid to make a mistake, lose their jobs or be denied valuable services. This way of thinking, and subsequent reputation inflation/deflation, poisons the data.
There needs to be room for error without such extreme fear, because no business can improve what it doesn’t even know is wrong. That’s a recipe for a crash and burn much larger than the initial problem.
There are also people who never give 5 stars because they don’t believe anyone deserves a perfect rating. And that’s part of the point. There are perfect moments, but not across-the-board, perfect for everyone, products or services. It’s unrealistic. To portray that erodes trust because it’s not believable.
Less Than Perfect is Bad?
Across many industries anything less than a five star review is perceived as bad. This is part of what fuels the rating inflation problem. Customer service agents are encouraged, even pressured, to receive five star CSAT ratings. They can be penalized for less and let customers know about this when asking for reviews.
But how useful is a guilt ridden or bullied five star rating? If it isn’t genuine it isn’t useful data. When the only feedback given is perfect feedback, and it isn’t really authentic, there’s no way to make data driven decisions to improve or evolve.
There is value in errors and how they impact growth. As Ty Givens of CX Collective says about her team, “I foster growth and development by providing room for fresh errors. In my view, making new mistakes is indicative of learning. As long as they slip up in innovative ways, progress is being made.”
A balanced view of performance and experience is necessary to identify areas that are working and those that need improvement. When pressure comes from the top down, and those reviewing or being reviewed only get perks through five star reviews, they’ll find a way to get them.
Consider shifting away from rewarding empty perfect numbers. Create a culture that functions from a perspective of satisfaction instead of fear. Honor positive experiences and foster real growth.
How AI Can Help
AI can’t stop people from making reviews based on fear or anger. As AI learns from human data there is a concern that the AI will learn to engage in rating inflation too. That’s why you want multiple points of data the AI tracks: customer sentiment over time, analytics during interactions to allow for pivoting, and surveys (without over-sampling the same customers). This gives you the checks and balances for clear information. That way you can make data-driven decisions not fear driven decisions based on a five star system alone.
CSAT is not the end all be all of measurement. There needs to be a well-rounded view. It can’t be a static number stuck in the past or based on extreme outliers (over the top responses, revenge, idiosyncrasies). It needs to be used proactively. That’s where CSAT.AI comes in.
CSAT.AI has multiple ways to address metrics from the patented predictive tool with phrase based prompts, to the generative AI features that parse out additional metrics and training enhancements. If you want an improved experience, not just the illusion of one, CSAT.AI is the way to go.