Experience Planner

On the final post of this series on reputation systems, I’ll talk about how online reputation can work as an incentive for participation. I’ll provide examples of online communities that use reputation elements to promote positive behavior and engagement. Let me start with a brief story.

Last year, I traveled to Tennessee for the first time to attend the IA Summit. As a full-time grad student, I couldn’t afford staying at the hotel that held the conference. So, I went to Yelp to look for reviews about cheaper places to stay in Memphis. I ended up finding a hostel in Midtown, which had positive ratings and a reasonable price. It turned out that the Pilgrim House was exactly what I was looking for: cheap, neat and well-located. After the trip, I left a review for it on Yelp. I also sent a “Thank you” compliment to the reviewer who convinced me to stay there.

Yelp is a catalog of local businesses, designed to help people find the best services. There are over seven million reviews on the site. Someone who is looking for a hotel or a restaurant, for example, can access Yelp to look for tips provided by the community. Users can assign ratings to the reviews and they can also offer compliments to the reviewers. This is a way to stimulate participation from both sides: the reviewer, who is encouraged to keep the good quality of the reviews; and also the reader, who wants to show gratitude for finding good information. When you’re navigating profile pages on Yelp, you can see the compliments each member has received.

Communities, such as Yelp, use reputation to promote positive engagement. Reputation can be used for multiples purposes, not only to narrow choices ─ as on Amazon and BookMooch ─ but also to increase participation and to incentivize certain types of behavior. As mentioned on Reputation Systems are Everywhere, we should think of online reputation as a cycle: “quality contributions attract more attention, which begets more reward, which inspires more quality contributions”.

Here’s another interesting example: Yahoo Answers. Yahoo Answers is a Q&A site. Any user who signs up for an account can ask and/or answer questions. Users can build their reputation by accumulating points. In order to get points, they have to perform different tasks on the site. The most rewarding way to participate is by providing a good answer, which could be selected as the best one (user gets more points). Yahoo Answers benefit from reputation because as users compete trying to provide the best answer, the site naturally filters the good from the ordinary content. Similarly, users benefit from the status and authority they get by building a positive reputation with points.

This post ends the series on reputation systems. The key takeways this time are: (a) reputation can promote positive behavior and engagement within communities and (b) there are multiple types of reputation systems; some of them are more collaborative (Yelp, compliments), while others are more competitive (Yahoo Answers, points). The challenge for designers is in finding the type of system that best fits each community.

One final tip: the book Building Web Reputation Systems is now available.

(This post is part of my research for the Information Architecture and Knowledge Management academic program at Kent State University).

If there’s an old book in your house and you don’t know what to do with it, here’s a tip: try using BookMooch. BookMooch is a non-profit site for book lovers who enjoy exchanging books. The site was built with a very simple idea. Users have profile pages where they can list the books that they are willing to give away. Anyone can sign up for an account and registration is free. Once you have an account, you are able to see the inventory of other members and mooch their books. Every time you mooch a book from somebody else, you “pay” with one point. And every time someone mooches a book from you, you get one point. Books travel from sender to moocher via regular mail.

Unlike the e-commerce sites mentioned on my previous post, BookMooch doesn’t have a central authority to control the transactions. The site is based exclusively on mutual trust. Therefore, it’s important to display clues about past mooches, so everyone can see who is playing fair and who is not. Nobody wants to use a point and mooch a book that will never arrive. Each user has a profile page that lists basic information on past activities, such as overall feedback score, books mooched and books given. Members can use that information to decide which members they should mooch from.

Reputation is a fundamental part of the experience on the site. Needless to say, BookMoochers with low feedback score are less likely to exchange books than those with higher scores. Other than score, the profile pages display additional clues about the community members. Their list of books given and mooched as well as their inventory and wishlist can indicate reading habits and preferences. Those records are open to any other user on the site, and they offer more information to assist BookMoochers during the transactions.

When a member puts Huckleberry Finn on her wishlist, she is not only increasing her chances of getting books. She is also letting everyone else know a little bit more about herself, as a Mark Twain reader. In the long run, the list of books given and mooched could serve as a way to cluster users with similar preferences. As noted by Bryce Glass on his post Reputation is Identity, on the Web “visible histories reveal shared affinities and allow users with common interests to find each other”. BookMooch doesn’t rely exclusively on user feedback to build reputation. The site also tracks its members’ regular activities and makes them visible to those who are interested.

This is an interesting approach for two reasons: (a) it takes advantage of the natural actions taken by users, instead of requiring extra efforts (e.g. filling out forms to describe preferences) and (b) it offers a granular perspective of each member, which is something particularly important for new users with no transaction history yet. Transparency is an important tool used on BookMooch and it helps BookMoochers make better decisions.

On my next post, I’ll go further and talk about reputation as an incentive for participation within online communities.

(This post is part of my research for the Information Architecture and Knowledge Management academic program at Kent State University).

A few years ago, I listened to a very interesting podcast called The Architecture of Participation. It was a panel discussion that took place at the Web 2.0 Conference in San Francisco, 2004. During the talk, the Chief Technology Officer from Amazon, Allan Vermeulen, described his team’s approach for using ratings and reviews on the site:

“So our choices [on Amazon] are: either we can hire a bunch of people who sort of know how to get things done and that can write a bunch of guides and so on, on the net; or what we can do is build a technology so that everybody else out there who actually know this stuff better than we do, can go ahead and write about it and build these guides and push them out to people.”

This podcast brings up a very important fact: most online reputation systems are built with user-generated content. Amazon does not rely on experts to assign ratings to its products. Instead, the site uses a platform that is capable of capturing users’ inputs in many ways. Ideally, Amazon’s reputation system becomes more effective as more users contribute to the site. By combining multiple and diverse opinions, Amazon is able to filter the entire catalog, narrowing the list of options for each customer. As I mentioned on my previous post, ratings-and-reviews can provide guidance and therefore, improve decision making.

It is important to note that just because reputation systems allow users to send feedback it does not mean that users will actually do it. In an interesting study about ecommerce, Jared Spool compared the number of reviews for the book Harry Potter 7 on different sites, including Amazon, Target, Walmart and Barnes & Noble. One month after the book’s release, Amazon had the largest number of reviews – 1805 -  in contrast with Target, which had the smallest number -  just 3. Both sites use the same platform to offer product recommendations, despite the huge difference in the amount of user feedback captured by each.

Even though ratings-and-reviews are becoming increasingly popular and expected, designers should remember that this functionality does not rely on technology alone. Users have to be motivated in order to participate, and each community fosters participation in a different way. As noted in the example above, some communities are more active than others. And because user feedback is the foundation of most reputation systems, an early challenge for sites like Amazon or Target is in creating engagement.

Still, there are alternative methods for capturing user feedback. On my next post, I’ll talk about a smart approach for assigning reputation values to entities.

(This post is part of my research for the Information Architecture and Knowledge Management academic program at Kent State University).

On my previous post, I provided an overview of online reputation systems with some examples. Here, I’ll explain why they are important.

As we know, online shopping can be very convenient. Sites such as Amazon offer an incredibly big variety of items. A search for “laptop” on Amazon retrieves more than 450,000 results. For any item, a customer can choose the best price, compare it with similar products and create bundles to get more for less. From one hand, this is a good thing and that’s what makes Amazon so attractive to customers. But it also poses a new problem: with so many options out there, how do I choose the best one?

Unlike a brick-and-mortar store, in which customers have the ability to see and sometimes even test the products before buying them, the Internet lacks some of the crucial elements a real shopping experience has. Whenever customers go to a store inside a shopping mall for example, they are able to make immediate judgments from their interaction with other people and the environment. Customers can hold the products, ask questions, evaluate if the staff is friendly, look around and see if the store is well-organized and observe the other people that buy things there.

All those physical cues are not present on the digital space. For that reason, e-commerce sites try to minimize the problem with good design, online assistance, discounts on shipping and so on. Reputation systems are also part of that strategy and they are built to facilitate the online shopping process. Amazon provides ratings and reviews for products. This is one way of helping customers overcome 2 problems: (a) the overwhelming number of available options for products and (b) the lack of personal interactions.

In a nutshell, reputation systems can improve decision making by aggregating user feedback and presenting it in a way that allows people to narrow their options. On my next post, I’ll talk about a little bit more about user feedback.

(This post is part of my research for the Information Architecture and Knowledge Management academic program at Kent State University).

Online reputation systems have become part of the experience of many websites, helping users make smarter decisions and stimulating them to participate on a regular basis. A popular example of online reputation is the star-rating system that tells you if a book is worth reading based on users’ ratings. Amazon, for instance, uses this feature to assist customers during their shopping process. Those who are unsure about whether to buy book A or book B can look at the number of stars for each book, and then make an educated guess on which one is the best.

There are different types of reputation systems, and some of them have people as entities that can build reputation. BookMooch – an online community that promotes book swapping – uses a feedback score to help users evaluate the reliability of their peers. On the other hand, Yelp – a website that displays reviews for services, such as hotels, bars and restaurants – lets its users give badges to those who provide useful information.

Even though online reputation systems can be very useful, they may also bring new challenges to the sites that use them. An early obstacle for any reputation system is in capturing user feedback. Just because a website allows users to provide ratings and reviews, it does not mean that users will actually do it. Another obstacle is in figuring out which reputation model is the most suitable for the product or service that is being offered. Because there is no universal reputation system that will work for all sites, the choice of which one to use depends on the type of community, the site’s purpose and the technology constraints of each project.

On my future posts, I’ll try to highlight the importance of online reputation in an information-overloaded space. I’ll also pinpoint the effects that may arise from interaction under such community-controlled rules. The mail goal here is to provide a quick report of issues to be considered when designing your reputation systems.

(This post is part of my research for the Information Architecture and Knowledge Management academic program at Kent State University).