Original picture by Julija under a Creative Commons license
During the Age of Separation we shielded ourselves from strangers by reducing all access to goods and services to money. No personal economic relationships are important because we can always "pay someone else to do it" wrote Charles Eisenstein in the book Sacred Economics.
In contrast, the Age of Reunion that we are entering is all about sharing: sharing with friends, sharing with coworkers, sharing with neighbors, sharing with complete strangers, sharing for free, sharing with a payment, etc. but sharing is not without its risks as the Airbnb incidents exposed. Can I trust you?
“For pretty much anything related to sharing resources, thinking through trust and reputation is a critical first step —particularly as it relates to user acquisition. If these companies don’t make their communities feel safe, they won’t have communities anymore.” -
Craig Shapiro, Collaborative Lab
"Reputation capital is becoming so important that it will act as a secondary currency, one that claims "you can trust me". It is shaping up as the cornerstone of the 21st-century economy It's the ancient power of word-of-mouth meeting the modern forces of the networked world." – Rachel Botsman
First level of trust: closed networks
Not everybody is ready to share with complete strangers. A number of projects address this issue by allowing users to create closed networks where matching their needs and building trust are easier to manage. For example:
- For sharing a ride (Amovens in Spain, Zimride in US, etc.) allow companies, universities, and even events to create local closed groups.
- NeighborGoods allow individuals to start a closed sharing group for as little as $6 a month.
Second level of trust: The social networks component
A lot of people are already building digital profiles on social networks like Facebook and Twitter. An additional reputation mechanism implemented by various projects is to connect users’ profiles with their profiles on social networks, so you can know more about who it is exactly you’re about to rent your house to.
MovoMovo (P2P car rental in Spain) has introduced their own "Social Trust Rank", an algorithm of initial social reputation. It works based on your Facebook account and you can add your Twitter or blog too. For example, newly created or very low activity accounts will be penalized with a low “Social Trust Rank.”
Feedback systems and peer-police
eBay has proved how the trust we typically form face-to-face can be built and assigned online by creating the grandfather of reputation systems, the Feedback Forum. After any transaction, buyers and sellers can rate each other using a simple points system (1,0,-1). Once users reach a certain number of points, they receive a star attached to their screen name, indicating their trustworthiness. The "silver shooting star" indicates the highest rating.
With around 98 to 99 percent of trades receiving positive rating the eBay approach seems to work and has been copied and adapted to a lot of other projects. One of the most common complaints with eBay is the use of nicknames instead of real names, and this is where the use of social network profiles can be a handy extra in the reputation system.
CouchSurfing, with more than 3 million users from all over the world, mainly works with a reference system. People can vouch for each other (under certain strict conditions) and even become verified (after you make a financial contribution to support CouchSurfing). In addition, the more complete the profile, the more likely it is the user will be able to get a couch. For those who offer a couch there is no obligation to accept couching requests, but it is expected that the user will answer (yes/no/maybe) the received requests. Your profile shows the percentage of requests you have answered.
Similar approaches and measures has been adopted by many other projects as, for example: Airbnb, TaskRabbits or Communitae (P2P lending site in Spain).
"Who is the person coming to stay in your home? Who is the person you are renting from, for that matter? When you have a good peer-to-peer policing system, a system of trust, you see things like eBay … I have never gotten into a dirty or damaged Zipcar. If it did not work that way, collaborative consumption would never work. … Peer-to-peer policing works actually very, very well" – Roo Rogers
Being able to create a sense of community among the users of the system also provides
Typical characteristics of reputation system
One of keys to success for collaborative consumption and sharing initiatives is that the hassle factor of engaging in such activities has been reduced to a minimum through the use of technology. Therefore, the reputation system needs to be complete, but simple enough for regular use.
Highlight five of the Rachel Botsman list of issues to consider when designing a peer-to-peer reputation system:
- Competition: We love being at the top of the heap. Publish your user rankings to create healthy competition among peers.
- Quality: Celebrate and reward users who take the time to contribute quality feedback; they should become the benchmark for others.
- Sticky ratings: Pick a primary scoring system (stars, ticks, tiers, thumbs, badges, numerical ratings) and give the ratings sticky names, such as “Power Seller”.
- People like me: We like to know, and tend to value, what our friends and people
- like us think of other people. Integrate “inner-circle” vouching mechanisms (for example, went to the same school, work in the same office) into your reputation system.
- Peer-police: An open reputation system must be peer-policed but if things do go wrong, your organization needs to be on hand quickly to offer support, resolve disputes and weed out the vandals and abusers.
"Reputation is a summary of one's past actions within ... a specific community, presented in a manner that can help other community members to make decisions... whether and how to relate to the individual" – Chrysanthos Dellarocas
One can not forget that business practices, trust and reputation do not work the same way for different people from different cultures. Qifang, the Chinese peer-to-peer lending platform, which caters to students, has lowered default rates by cleverly leveraging cultural norms by requesting borrowers to provide family details, so they'll feel pressure not to shame the family name.
Reputation banks
What's missing in collaborative consumption today? Roo Rogers answer was clear:
"We still need great, centralized databases of reputation. If we’re good Zipsters, and good eBay tradesman, our rep should follow us. People should be able to see our record, tied to a fixed identity, and do business with us more confidently."
Facebook also has the potential to become the arbiter of online trust. "The incentive to be a good player in that ecosystem goes up dramatically when it's associated with my real identity," says Carl Sjogreen, manager of Facebook's platform product team., "because if someone leaves a bad review of me on AirBnB, that will carry with me to the rest of the web.”
These are new territories and there are still a lot of unanswered questions: How to balance different sources of activity? Does all “sharing history” count the same? Should your offline citizen behavior be part of it? Should governments be involved at all?
Prepare for the worst, so users can trust
Of course, renting out one's home or car to a stranger comes with risks. "My sense is that the enthusiasm in some spaces may be outstripping logic," Eric Clemons says, and also warns that the security measures associated with some of these sites may not be keeping pace.
Case in point: Airbnb recently received negative press when a few consumers went public with stories of their homes being robbed or damaged by renters found through the site.
Airbnb responded with a new policy where it will cover loss or damage from guests up to $50,000 among a renewed set of safety measures. The case created a ripple effect in many other projects (specially the ones focused on P2P lending) that reviewed their own policies, but sometimes unrealistic expectations are put on companies.
"As with any peer to peer marketplace, it's impossible to remove risk completely. It's our job as sharing platforms to make sure our members are aware of the risks and have the information needed in order to manage the risk effectively." -- NeighborGoods.net
Car rental such as Zipcar or Connect has also been subject to scrutiny by the press with a focus on insurance issues. Legislation and habits on this topic differ from country to country, for instance, SocialCar (Spanish P2P rental) has a tailor-made insurance policy with a leading insurance company. Car owners have to cancel their existing insurance policy and sign with the one proposed by SocialCar, there is simply no other way to operate according to Spanish regulation.
The biggest challenge remains, and it’s to present all legal requirements in a way clear and simple enough for the average user to understand.
So, do you think you can trust me?
Comments
cameron tonkinwise wrote on 09.28.11, 1:56pm:
It is very important to distinguish between generic trust, which is virtually meaningless, and trusting someone to do something particular. I would trust some people to do some things but would not trust those same people to do other things. This is the problem with attempts to determine general trustworthiness through aggregated reputation systems. It also misses that part of trusting someone to do something is not just an assessment of their general intention, but their capacity - do they actually have the skill to do what is required? I might be really trustworthy according to the leading reputation systems, but that has little to do with my ability to handle your specialized equipment.
It is also crucial to differentiate more clearly between
1) trusting a person (to do X),
2) trusting the other people who vouch for or vet that person (there is academic literature about how few people give negative couchsurfing ratings, for fear of reciprocal negatives)
3) trusting the system that does the valuations (how transparent are their algorithms/systems/data collection)
4) trusting the people who control the system of valuation (e.g., to not on-sell aggregated data)
In addition to sharing systems designing for worst-case scenarios, it is important to also enable
a) trialing in low risk ways to build trust of the system and people
b) create error-friendly systems that allow for mistakes, misunderstandings
Rob wrote on 09.28.11, 2:28pm:
The days of inherent trust over the internet are gone. I now need to have gotten something out of person X that builds trust over time in order to conduct bigger and larger transactions... typically not monetary.
Juho Makkonen wrote on 09.29.11, 2:27am:
A good post that summarizes well many main points of reputation systems in collaborative consumption (and many other online) sites. The three-layer trust model that is presented (closed networks, social network component and feedback systems) is in the very core of the "start-your-own-sharing-site" -system that we are building with Kassi.
However, there is one point that I disagree on. I think that Rachel Botsman's idea of competing with reputation is fundamentally flawed and does not lead to the best results.
The best book on online reputation systems is probably written by former Yahoo! researchers Bryce Glass and Randy Farmer. It's called Building Web Reputation Systems and it addresses all these issues thoroughly and thoughtfully, based on years of hands-on experience with huge online communities. One of the biggest takeaways of the book is summarized in this post "Leaderboards considered harmful": http://buildingreputation.com/writings/2008/12/leaderboards_considered_h...
Systems like leaderboards should only be used in sites that are competitive by nature, like gaming sites. In collaborative sites, however, people want to achieve common goals together, and putting them against each other is the worst thing you can do. Even if you manage to get the metrics of the desired behavior right (which can be surprisingly hard), the users might then start gaming the system. Think of original Amazon's top reviewers list: the people on the top got there by using false methods (like posting reviews of books they have not read). And the leaderboards might also discourage new users: those people have done so much and I have done so little, I'll never get to the top, so why even try? When people are compared, they often take it way too seriously, and forget the actual meaning of the site, which is sharing with each other.
Show the users' positive feedback averages, and give them visible rewards and badges when they do a good job. But don't try to compare them with each other too much, that's against the true nature of collaborative consumption.
--
Juho is a Co-Founder and CEO of Sharetribe, a platform for creating custom peer-to-peer marketplaces. http://www.sharetribe.com
Paul M. Davis wrote on 09.29.11, 8:40am:
Good points, Juho. I also find many of the proposed reputation systems troubling. Many of them prioritize the sort of accomplishments and metrics that are primarily relevant in the workplace, which are not necessarily effective representations of an individual's trustworthiness — say, valuing someone's networking prowess over real connections made with other individuals. It suggests a further mixing of our professional and personal selves that makes me very uneasy, and perpetuates a false impression of the values that make an individual trustworthy.
A truly relevant reputation system may be impossible — what is more valuable, inflating some meaningless Klout score by playing Twitter in the way the "social media expert" shysters tell you to, leaving a rote happy birthday message on a colleague's Facebook wall, or flying across the country to console a friend who has lost a family member? Clearly the latter, but there's no way to effectively track or quantify the behavior that makes an individual truly trustworthy.
As you mention, positive feedback averages work great in existing services, not quite sure why we're reinventing the wheel here. Judge people on the metrics that are relevant to the task or service at hand.
Dimitris Tzortzis wrote on 09.29.11, 4:44pm:
+1 to Juho and Paul.
Some more thoughts...
It is very troubling to me the idea of a company (or companies) managing our reputation and trust online. I believe it goes completely against the true P2P connections that sharing networks want to accomplish to have an external company vouching for you. It's a trap that will keep us in the "Age of Separation". Those connections we'll be making won't be true. You won't have to trust the other party. You can trust TrustCloud. If something bad happens in this transaction maybe you can be compensated as well by TrustCloud. Now, how's that different from typical commerce?
Also, centralized reputation/trust systems may lead to a bad situation for sharing. In 5-10 years there will be many more sharing networks than now. Let's imagine lots of people use them and that we have a trust currency that we can carry around on the web to use on P2P transactions. Immediately, that marginalises those who don't have that currency. People who have lived their lives offline, who have not participated in an organised sharing network online. Sure I have plenty of good references in Couchsurfing but what about my mother? When she tries to go online and barter for her homemade marmelade, what is she going to show about herself? How is this woman? Can she be trusted? She's not on facebook, or trustcloud. She hasn't even tweeted yet.
We're supposed to use the sharing networks so we can have human connections and be liberating ourselves. From corporations or anything else that's holding us back. So we can focus on what we need, on the greater good, on living better etc. If we're putting a service between ourselves, it's not really liberating, is it? It's probably another bond, another need we will be having. We won't be able to have a connection without checking the other party first. Those who aren't online will be excluded, marginalised. It sounds like tyranny to me. Am I exaggerating?
(Not to mention concerns about that company like, what's their business model? Who are their clients?)
----
By the way, I saw yesterday this documentary by the BBC, "The Trap: What Happened to Our Dream of Freedom". Part one has some very interesting presentation about how we got to the "Age of Separation" in the first place and how it got hard to be able to trust each other. The author describes how the rise of game theory established the idea of human beings being selfish and that numbers and metrics was the only thing that could be trusted to give accurate predictions about society, economy, human behaviour etc.
Rings a bell?
Graeme Smith wrote on 10.03.11, 7:16pm:
At one point, I had begun to believe the "Jobless" stats, then I heard rumors that people like me were not being counted anymore. Then later, I heard that the median income in my community was almost exactly twice the average income... Then in a fit of censureship the Government insisted that all census forms be done on the short form.
It was at that time, I knew, that Government didn't want to be gainsaid by its own figures anymore we were to consume the numbers as if they were right, even if there was no statistical support for them.
So does anybody REALLY trust the metrics anymore?
Barbara Pantuso wrote on 10.09.11, 8:58am:
Great points, Cameron. Trust isn't a simple "yes" or "no" and it's measured on very personal and specific terms, as you mention. It cannot be compared to a credit score. As such, we need flexible (and personalized) systems to measure that trust. While I appreciate the value of algorithms and aggregated trust systems, I would also love to be able to rank and review what I personally deem as trustworthy for whatever particular scenario.
Really interesting post!
Usually it's more about creating a "bubble" where you can share in confidence, and also digital identities help a lot more than the anonymous internet of some years ago. But even then, the discussion forums had a simple form of meritocracy: the more you participate, the more you are a part of the community, and the less likely you are to do something that could stain your reputation and break relationships.
The examples you quote are quite illustrative.