Crowdsourcing, Volunteers, and the Sharing Economy - S01E05
This week, Scott Moore joins the cast…
In this episode we talk about the emerging idea of the social economy and how it interacts with some commons: specifically pre-existing crowdsourcing and volunteer-driven applications.
—- Links —-
Professor Arun Sundararajan: The Sharing Economy [video]
Rides: http://rideshare.511.org/ http://en.wikipedia.org/wiki/Slugging/ https://www.uber.com/ http://www.lyft.me/
Crowdsource Control: Researchers show how easy it is to manipulate online opinions [article]
Dan Ariely - Predictably Irrational: The Hidden Forces That Shape Our Decisions [book]
Panos Ipeirotis - A Computer Scientist in a Business School [crowdsourcing research blog]
Stack Overflow [Technical Q&A site with top-quality content, socially generated]
Farmer, Glass - Building Web Reputation Systems [Chapter 5 - web version]
—- Transcript —-
Randy: Welcome to Episode 5 of the Social Media Clarity Podcast.
I’m Randy.
Marc: I’m Marc Smith.
Scott: I’m Scott Moore.
Randy: This week we’re going to talk about the sharing economy and the commons.
First, a little bit of news. Bryce Glass will no longer be joining us regularly on the podcast. We hope to have him back soon as a guest.
In today’s news, the California Public Utilities Commission, at the end of July, came to an agreement with the ride-sharing companies. If they meet five conditions, they would have a special new way to operate legally. Those conditions are $1 million per occurrence liability insurance, criminal background checks, in-person screening and vehicle inspections. They must have a zero tolerance policy for drug and alcohol use and a five-star rating system for community reviews of drivers and passengers.
Today, we’ll speak with Scott Moore who’s got a long history with online communities and user generated or crowdsourced content.
Before we talk to Scott, Marc, talk a little bit about the science behind the commons.
Marc: Sociology has paid a lot of attention to the ways that groups of people come together and produce a resource that would not be available if each of them worked individually. This is known as collective action. When collective action results in a resource that can be used by anybody, that’s called the commons.
The tragedy of the commons is when the commons falls apart because too many people either take from it or not enough people give to it. The internet is the home of many collective goods and many commons. If you’ve ever read Wikipedia, you’ve already benefited from the contributions of other people and you get to read from Wikipedia without regard to how often you’ve made a contribution.
There are many great examples of success stories of these commonses, but they don’t all succeed and quite a few of them fail. Not even every Wiki page is equal. Now there are new kinds of collective action. We’re now seeing the rise of the sharing economy. Rather than owning a car, you use a web service and a car is sent to you. Somebody drives you where you’re going, and then you get to offer them a donation. Other kinds of resources like sharing back bedrooms are emerging as a hotel alternative. These are all examples of the crowd providing a service.
The challenge of course is getting people to continue to contribute with high quality and to have people consume the resource and moderate their consumption, which leads us to Scott who’s been at the front lines managing groups of people who are making contributions. Scott, can you tell us what’s new with the sharing economy? Why now? What’s making sharing such a big phenomenon?
Scott: What is new is the scalability and the flexibility. The idea that one person in a large city now has the ability to rent out a spare room to anybody in the world.
Randy: I think there’s one more thing that’s significantly changed. Before it was called the sharing economy, it was called crowdsourcing, and most of it was done on a volunteer basis, tapping into altruism or egotism. If I’m contributing to Wikipedia, part of my motivation is that my opinion is going to affect millions of other people.
One of the big shifts we see when we start talking about the sharing economy is the flow of cash.
Marc: Is the serpent now in the sharing garden? Is the introduction of payment going to displace the willingness of people to make charitable contributions? Is it always the case that pennies are better than pixels?
Let’s use the automobile sharing economy companies and compare them. Specifically, how does it contrast with previously existing ride sharing models? We used to have this thing called carpooling and ride sharing. Can you talk a little about the difference between Uber, Lyft, and currently existing California Bay Area ride sharing practices?
Scott: Let’s actually start with the ride sharing aspects, the organic, largely volunteer, no central coordination practices. In the Bay Area, it’s called the casual carpool. Across the country, it is also called slugging. Usually at transportation hubs, bus stops or train stops, people will gather. Cars will come by looking for someone to pick up as a ride share.
The history of it goes back at least as far as World War II. Back then fuel was a premium. It became popular again in the ‘70s, during the oil crisis. Another factor is being in a carpool, there are special high occupancy vehicle lanes and there might be a reduction in the toll for crossing the bridge.
There’s no central organization that organizes the riders and the drivers. Transportation systems do advertise and help map them out, but if one of these people who are participating or one of the organizations who are participating were to suddenly stop participating, it wouldn't stop the ride share.
Uber is a coordination of existing … they might limousines, but they’re generally called liveries. They provide a dispatching service, a very scalable service and it’s very easy to use. It's done through the mobile. The payments are done through mobile. The drivers are professional drivers. The drivers carry certain kinds of licenses. Their cars are some kind of livery car.
Lyft is the different idea that anybody could be a driver. The idea of who is driver and who is rider could shift at any given point. If I sign up as a driver, then I can get a notification from Lyft that somebody needs a ride. I go to that location, I can give that person a ride. The pool of drivers is potentially much larger. It’s also much more ad hoc. The Lyft is much more along the lines of the commons that Marc described, a large number of people who are putting in a small amount into the commons of driving, and then there’s a large number of people who are accessing that pool.
Randy: One way it might be easy to think about it is in Uber, the customers sit in the backseat. In carpooling, people tend to sit in the front seat because it’s kind of a social thing. Lyft is trying to be more like carpooling and less like Uber.
Scott: Exactly.
Randy: That’s where it causes some interesting challenges. The reason we covered the news item was it causes trouble with government regulation. Uber is happy to be lumped in closely with previously existing livery drivers. Lyft is afraid to.
Marc: If we want people to contribute their driving services, we may need to pay them so it makes sense to add money to the system, but sometimes money can be a problem. How can money be a problem in a peer production or a crowdsourcing environment?
Scott: For example, the casual carpool does not have an official direct fee payment. There is an understanding that $1 to the driver is a nice thing to do and so in that case it’s a donation, it’s a goodwill gesture.
A couple of years ago, if you’re in California, the bridge tolls used to be free for carpools, and that rule changed and there’s now a charge. Some of the incentive for carpooling seems to have gone down. It’s put a little bit of a social pressure on the rider to actually give the $1 “donation” or perhaps even more. You still won’t be refused a ride if you don’t give a $1, but there is a bit more resentment. This is an external price environment change.
In the book Predictably Irrational there is a story about a daycare center that had a problem with parents arriving late in order to pick up their children. As a form of incentive to get them to arrive earlier, they would charge them a small fee. It was something like maybe $3 for every 5 minutes that the parent was late. The desire was that it would reduce the number of parents being late to pick up their children.
The opposite effect occurred. It actually caused more parents to come late. The social contract was broken and a value was placed on how much 5 minutes was worth. Parents would say, “Well, all right, if I’m 15 minutes late, I’m willing to pay $10 extra so that I can be 15 minutes late.”
When they went back and they took away the fee, thinking that maybe the behavior would go back, it didn’t. The parents who had been charged the fee prior continued at the same lateness that they were before.
Randy: Because once you set a price on this, people then continue to think of this in the economic sense, not in the social sense.
Scott: Yeah.
Randy: Paying people can actually be demotivating. Sometimes the amount of the payment isn’t enough to really motivate and the existence of this now substandard payment actually demotivates. Whereas if the task or the good was not paid for previously, at least you could derive from that some kind of social satisfaction, at least I’m the guy who answers all the questions or provides all of these resources. Once you begin to pay me, I start to make economic comparisons and I think well, this isn’t really worth my time anymore. I was going to do it for free, but I won't do it for an insufficient fee.
Randy: Look at StackOverflow.
You couldn’t afford to pay those people for their time to answer those questions. It makes no sense to introduce money to that system and it will cause it to fundamentally break. Google built their first Q&A site, which was a pay site and it was a miserable failure.
Scott: I think one of the key stories that’s been lost in history is the AOL volunteer “Guide” lawsuit. Over 15 years ago, AOL had a lot of volunteers. Their entire chat system was based on volunteer moderators. This was before the web. These services charged by the hour and so as an incentive for the moderators, they would receive free access to AOL. This added up. When I moderate for AOL, I’m potentially saving hundreds and hundreds of dollars a month in service fees that I don’t have to pay.
Then the environment changed. Flat-rate ISPs were taking AOL’s business. AOL started shifting to a flat rate. Now instead of hundreds of dollars that my time was worth in volunteering, it was now worth about $19 a month. It’s a huge reduction in the perception of the value of my work, yet the actual work that was being asked of them didn’t change. It was at this point plus having them do the same thing that paid employees are doing triggered a couple of lawsuits and a Department of Labor investigation which led to fines on AOL, and it really damaged the volunteer services that they were able to offer.
Randy: When I worked in a Q&A center, we looked into crowdsourcing. They were an all volunteer organization, but they were looking to pay. If people don’t really want to contribute, if they’re only in it to make the money, they maximize their revenue for the minimum effort. The number one problem with crowdsourcing, things like question and answers or any kind of small task on the internet is quality management. It’s virtually impossible to get quality without paying significantly more for the work.
Scott: One of the crowdsourcing companies, and I had a private conversation in which we were talking about crowdsourcing, and the issue of quality crowdsourcing of ratings or categorizations. I was working on a project where we were trying to do both. They admitted that categorizations are much easier, but generating content in a crowdsourced way was a nut they hadn’t quite cracked. It wasn’t even just the quality. Creation was a problem. They then didn’t even have a way of efficiently grading the quality. They couldn’t even tell if they were creating good quality without having to then crowdsource the whole quality check.
Randy: You would have to pay twice, once in the generation and once in the quality then you need to recursively do that one more time on the quality check.
Scott: Yeah, it’s crowdsourcing all the way down.
Randy: It’s crowdsourcing all the way down.
For this week’s tip, if you’re considering paying users for crowdsourcing, first, read Dan Ariely's Predictably Irrational.
I will share one story from that book that maps on to everything we discussed. The research he did with volunteerism, in one of his experiments, he had college students perform a menial task on a computer. I believe it was dragging one shape into a hole and he said “do that as many times as you can”.
He found that people did it more when they were asked to do it and were not given money as incentive: that money was a disincentive. But there was a specific case I wanted to share and I think it gets at a lot of what Scott was talking about. He did an experiment where you just got asked to “do it”. You could also receive a small candy bar or you get paid. What they found is that the candy bar was approximately equal to being asked, maybe even a little higher. But then there was a change: he ran the experiment again and he put a price sticker on the same candy bar and immediately, the candy bar’s incentive dropped to be the same as the cash amount. That perception of being paid figures into the calculation.
I’d like to thank Scott and Marc for today’s episode.
Scott: It’s been great.
Marc: You couldn’t have paid me to do this.