How do online rating systems affect participants in the sharing economy? Chen Jin, a post-doctoral researcher funded by the Mack Institute, shares preliminary findings from his research about the impact of a unilateral, or one-way, rating system on technology platforms, service providers, and customers. He recently spoke with Knowledge@Wharton about his paper, “Do Ratings Cut Both Ways? Impact of Bilateral Ratings on Platforms.”
An edited transcript of the conversation follows.
Listen to the podcast:
This podcast is part of a series produced by the Mack Institute and Knowledge@Wharton. Explore the rest of the series here, or find even more K@W podcasts on iTunes.
Knowledge@Wharton: Can you tell us about your work? What are you trying to study?
Chen Jin: Our paper is essentially about trying to understand the impact of rating systems [reviews] on online marketplaces. Digital innovations, and especially the rapid development of the mobile internet technology, as well as smartphones, have really changed people’s everyday lives.
Take, for example, sharing economy companies like Uber and Airbnb. These companies provide platforms that connect millions of online individual customers to millions of offline individual service providers through very simple applications on their smartphones. On the demand side of these platforms, customers now are able to order at any time and any place. On the supply side, the individual service providers now can decide when to work and how long to work on the platform.
These new technologies remove physical obstacles between two parties. On the other hand, a psychological obstacle, namely generating trust between the two parties, is still there. Notice that the concept of developing trust is not new, but it is playing a more important role in today’s economy.
The reason is because transactions that happen on these platforms are between two complete strangers. Because of the high liquidity [turnover] of both parties, people — the customers and the service providers — are less likely to encounter each other more than once. For example, you are not likely to stay at the same accommodation listed on the Airbnb website a second time because probably you want a different place to visit on your next vacation. Even if you take Uber to the office every day, still it is not likely that you will meet the same driver.
“By allowing customers to rate the service providers, it makes service providers more motivated to exert extra effort.”
So because of this high liquidity on both sides, and also entering costs are almost negligible on the supply side, this creates big issues in establishing trust. Notice that it is very difficult for the customer to judge how serious the service provider feels about his business and his willingness to continue to provide the service in the long run. The rating system of reviews has proven to be an effective tool to establish trust between two parties.
It is natural that people want to have reassurance from other users so that they can make sure that they don’t waste their money on bad products or bad service. It is the rating system that makes the crowds available, that brings complete strangers together to conduct business activities. Our research aims to understand how these rating systems affect the three parties of the game, which are the platforms, service providers and customers.
Knowledge@Wharton: What are your paper’s key takeaways, and were there any conclusions that surprised you?
Jin: We are still in the process of analyzing our model, but we think we have some interesting results to share with you. In the first stage, we considered a platform without a rating system versus a unilateral rating system, in the sense that only the customers are able to submit their ratings to the service providers. In the next stage, we are going to analyze a bilateral rating system where both parties can rate each other after their transactions.
What we found is that the rating systems will have a different impact on the three parties of the game — the service providers, the customers and the platform. Let’s talk about this one by one.
First, from the perspective of the platforms, we found that it is always in their best interest to implement these rating systems. The underlying intuition behind it is because by allowing customers to rate the service providers, it makes service providers more motivated to exert extra effort, because now their effort is more transparent and observable through the ratings. So this is going to make the customers feel a little bit happy because they get some extra benefits. This in turn will make the platforms more flexible in re-designing its pricing strategy. So for example, right now the platforms are able to increase the price a little bit, and also increase the fee charged by the service providers a little bit. That’s going to boost the revenue of their platform.
When we turn to the service providers, one would tend to think that their circumstances should get worse after the implementation of the one-way rating system, and this is also what we thought initially. First, because they have to work harder since their effort is now observable, and second, the platforms now [can charge them more money.]
So it sounds like the service provider would incur a double loss, and it is true that for the profit per order they do get a decrease, but this is only part of the story. We now know that on average the service providers are making more of an effort. So it’s going to attract more customers to use this platform to book a service. That’s indeed what we found. Under some circumstances, the transaction volume got an increase that exceeded the loss of the profit per order. In that case, the service provider is actually getting better due to the fact that they are being rated.
The last part is the customer. It is natural to think that after the implementation of the rating systems, it should attract more customers to join this platform. However, we found that it is actually not always the case. The reason is it really depends on the pricing strategy of the platform. Notice that the platform cares about two things. One is the profit per order; the other is the transaction volume. And there is usually tension between these two factors.
“It is always in the best interest of the platform to implement such a rating system.”
If you want to have a higher profit margin per order, usually you cannot at the same time achieve the goal of having a high transaction volume, right? So what we found is that if the customer’s valuation about the service itself is low enough, then the transaction volume plays a more important role than the profit per order to the platforms. So what the platform will do is adopt a less aggressive pricing strategy such that it’s going to induce all the servers on the platform to exert some level of extra effort. That’s going to make customers feel happy so more customers are going to be involved in these platforms, and increase the transaction volume.
On the other hand, if customers already highly value the service, meaning they are already very satisfied with the service even without any extra effort by the service provider, the profit per order plays a more dominant role than the transaction volume. In this case, the platform will adopt a very aggressive strategy in the sense that it is going to increase the fees charged to the service provider.
That’s going to make only some service providers exert extra effort. The remaining service providers will produce the usual, basic service, without the extra effort. These service providers just barely make a living on the platform. So in this case, the transaction volume on these platforms could actually be lower than the platform without a rating system. Those are our main findings.
Knowledge@Wharton: What are some practical applications of your research?
Jin: Based on our analysis, we know that it is always in the best interest of the platform to implement such a rating system. Second, in terms of its pricing strategies, the platform should conduct some marketing research to gather information about how customers feel about the service itself. For example, they need to know the customers’ willingness to pay for this service. After they know this information … the platform can either adopt [an aggressive pricing scheme or moderate pricing strategy.]
Knowledge@Wharton: How will you follow up your research?
Jin: This is still ongoing research. In the next stage, we are going to focus on analyzing the bilateral rating systems where the servers can also rate the customers, [and compare it with the unilateral rating system].
In this case, both the customers and the service providers are going to interact with each other for some period of time. It could be as short as a ride, like taking Uber, or as long as a couple of months, like booking a room on Airbnb. Here, the service provider also faces risks. For example, the customer could damage the asset of the server, such as the car or the house, which will affect the servers’ ability to provide the service in the future. In these type of platforms, the two-way rating system is needed and we will analyze this system in the next stage.