User:Csaal/Crowdsourcing

=Crowdsourcing=

Overview
We have already learned about human intelligence and tasks that can only be accomplished by humans. Crowdsourcing is only one step further into this topic. Crowdsourcing is a mechanism for undertaking human intelligence tasks. The following chapter is going to give you a deeper insight into what crowdsourcing is, its functionality and what its advantages and disadvantages are. You will also be provided with some examples of today's use of this mechanism.

Introduction
Since Web 2.0 the web is not only a place to get information but we also started to contribute our own content. Millions of people interact in websites like YouTube, Flickr and Facebook. The web has shown us that a crowd of individual people is able to create quality content together. Take Wikipedia for example. It is a free encyclopedia created and edited by users all around the world. Why not make use of todays network like that to accomplish other specific tasks? This is what crowdsourcing means to do.

What is crowdsourcing?
Crowdsourcing is a relatively new term that is only a few years old but became a buzzword in today's business and is a widely used mechanism. You might already be familiar with the term "outsourcing". Outsourcing is the process of giving individual tasks to a third-party. This process became a popular process in the early 21st century because of its advantages. It reduced the need of hiring specialized staff and made it possible to save expenses. Jeff Howe and Mark Robinson realized that this process of outsourcing was happening on the web too. Businesses would outsource their tasks by finding individuals via the web. In 2006 Jeff Howe coined the term 'crowdsourcing' in one of his articles. . He noticed that businesses do not only outsource their tasks but forward them to a large crowd of people thus the term crowdsourcing. Jeff Howe wrote an accompanying blog post about crowdsourcing with an even more detailed definition of the term:
 * "Simply defined, crowdsourcing represents the act of a company or institution taking a function once performed by employees and outsourcing it to an undefined (and generally large) network of people in the form of an open call. This can take the form of peer-production (when the job is performed collaboratively), but is also often undertaken by sole individuals. The crucial prerequisite is the use of the open call format and the large network of potential laborers."

The concept of crowdsourcing is not new but it takes advantage of the networked world. With today's online community it has never been so easy and efficient to reach a broad audience to make proper use of this open call format. The term crowdsourcing is used for a lot of different forms of strategies that all have one aspect in common: They depend on the input of the crowd. Due to its blurry limits on what a crowdsourcing task is, varying definitions came to be. 2012 Enrique Estellés-Arolas and Fernando González Ladrón-De-Guevara set themselves to find one integrated definition. After studying 40 different definitions they proposed the following one:
 * "Crowdsourcing is a type of participative online activity in which an individual, an institution, a non-profit organization, or company proposes to a group of individuals of varying knowledge, heterogeneity, and number, via a flexible open call, the voluntary undertaking of a task. The undertaking of the task, of variable complexity and modularity, and in which the crowd should participate bringing their work, money, knowledge and/or experience, always entails mutual benefit. The user will receive the satisfaction of a given type of need, be it economic, social recognition, self-esteem, or the development of individual skills, while the crowdsourcer will obtain and utilize to their advantage that what the user has brought to the venture, whose form will depend on the type of activity undertaken."

Both definitions show that the essential components of crowdsourcing are an open call by the crowdsourcer, to make the task known to people and a crowd of individuals to contribute to that task. The open call refers to the lack of a mechanism to select the participants. This means that, in principle, everyone can answer to the call.

Crowdsourcing is a term that has been conceived only a few years ago but that doesn't mean this mechanism has not been in existence before that. A well known crowdsourcing task that has been done over years before this term came to be are beta invitations to certain platforms or video games. Developers had to test their work. The company made use of the crowd, the beta-users to do that.

Wisdom of Crowds
The underlying concept of crowdsourcing is the so called "wisdom of the crowd". This is a concept that is essential to understand why crowdsourcing works. The question is why would you chose a crowd of individuals to solve a sometimes very complex task instead of someone specialized on the topic. The Wisdom of Crowds is also a book written by James Surowiecki. He argues that a decision based on a group of individuals is often better than one made by a single member of that group. An example to substantiate his proposition was performed by us in one of our classroom sessions:
 * Each of us students had to make an individual guess about our professor's weight. A lot of us were far off the right value. Merging all our individual guesses to one value we got a proper guess though.

Examples like this one underline what Surowiecki says about the wisdom of crowds:
 * "[…] ask a hundred people to answer a question or solve a problem, and the average answer will often be at least as good as the answer of the smartest member. With most things, the average is mediocrity. With decision making, it’s often excellence. You could say it’s as if we’ve been programmed to be collectively smart."

For this principle to work it is important that the participants make individual choices and aren't influenced by other people's choices. We did this in our classroom session example by writing down all our guesses on paper instead of saying them out loud. The web can provide a way to aggregate many independent ideas without running into the problem of agreement among participants. Our world wide network also makes it possible to provide a context where Surowiecki thinks the principle of the wisdom of the crowd applies:
 * "There are four key qualities that make a crowd smart. It needs to be diverse, so that people are bringing different pieces of information to the table. It needs to be decentralized, so that no one at the top is dictating the crowd’s answer. It needs a way of summarizing people’s opinions into one collective verdict. And the people in the crowd need to be independent, so that they pay attention mostly to their own information, and not worrying about what everyone around them thinks."

Surowiecki thinks that given a context that satisfies these four qualities – diversity, decentralization, collective verdict and independence – then the wisdom of the crowd applies and the mechanism of crowdsourcing is possible.

Collective intelligence
Another underlying concept which goes hand in hand with the wisdom of the crowd is collective intelligence. Pierre Levy describes collective intelligence as follows:
 * "It is a form of universally distributed intelligence, constantly enhanced, coordinated in real time, and resulting in the effective mobilization of skills. I'll add the following indispensable characteristic to this definition: The basis and goal of collective intelligence is mutual recognition and enrichment of individuals rather than the cult of fetishized or hypostatized communities."

Collective intelligence has been talked about for decades. With today's technologies, especially the web, people take advantage of this concept. With the help of the Internet, people are able to connect world wide, create a world wide network and thus create a network of collective intelligence. No one in this network knows everything, but each individual knows something and as a whole group therefor holds a more detailed, comprehensive knowledge.

It's not only the crowd as a whole and its collective intelligence which make crowdsourcing work. Crowdsourcing can be effective when it comes to solving complex tasks because of the lack of certain knowledge. Take a small group of people that are the best in their field, they all might think alike. The lack of knowledge in this certain field might be the key quality to come up with a thought that brings a breakthrough.

What motivates crowdsourcers and the crowd?
After having learned what crowdsourcing is, the question arises why a business owner would decided to crowdsource a project and what motivates the crowd to participate in that. In this chapter you will get further information about the motivations of both parties and what can lead to the decision to participate in this crowdsourcing process. For crowdsourcers it is vital to consider a form of motivation for the participants. It is important that both parties can gain something from the project or otherwise it might fail.

Motivations to crowdsource a project
One of the main reasons why a decision to crowdsource has been made, is saving expenses. Every business owner, every project manager wants to realize his project as cheap as possible. Crowdsourcing builds a foundation for that. Thanks to the world wide network it is possible to gain access to cheap labor and information. Through the web we have access to huge amounts of data but it would be time-consuming and hence expensive to gather them on our own or search for the exact data we need. Crowdsourcing can help with that. By making an open call and asking for a crowd to help with this task, we will not be troubled by the information overload of the web.

This world wide network also provides access to a wider array of talent. Companies do not need to hire experts or train their employees for special tasks that will cost them additional money. Through crowdsourcing they are not only able to find the people they need, but they can also find them faster and for a lesser cost.

Crowdsourcing is also interesting when in comes to research. Researcher might need large groups of people to collect the amount of data they need. With the help of crowdsourcing they can gather the information they need easily across the web from a large group of people. This is especially helpful when they want to gather information that they cannot access locally by performing user studies on-site.

Motivations to participate in a crowdsourcing project
An essential aspect of crowdsourcing is the crowd itself. Without the crowd that answers an open call this kind of outsourcing would not be possible. The question is, what motivates people to be part of this crowd that answers open calls. By participating the user wants to gain something. Depending on the kind of crowdsourcing task there are different profits for the user.

Some crowdsourcing tasks come in form of a game and some users playing this game might not even be aware of this game being a crowdsourcing project. What motivates the user in this case is having fun. By gathering information in a playful way, the project can ensure the users' enjoyment and hence guarantee the participation of a larger crowd.

Another motivation might be being part of a community. By having a Facebook account users are part of its community and have a way of social contact. Being part of this community makes them want to use Facebook's Like button just like everyone else does.

Then there are tasks that come with a certain payoff. By paying the participants it is easier to find a large enough crowd that wants to help out with the project than wanting to get help for free. The motivation of the crowd is dependent on the amount of payoff.

By participating in a crowdsourcing project users might also gain a certain kind of status. They might for example get acknowledgment for their attribution in a funding project.

Two other reasons for a person to be interested in participating is a job related matter. On the one hand he might gain new skills or improve his old ones. Through participating, people can also demonstrate their skills and hence may get noticed by possible employers.

Crowdsourcing tasks may attract users because the independence of time and place. Participants can usually work at home and can decide for themselves when to work on the project.

Types of Crowdsourcing
When it comes to crowdsourcing it is not only a question of whether to crowdsource or not. The more important question is which strategy works best for the given problem. As mentioned before there are multiple ways to define the term crowdsourcing and many different strategies that are being classified as such. There is a variety of attributes which distinguish between different tasks. The tasks can differ in their size. It can be a rather small task that is accomplished in a short amount of time or a rather large task that needs days or even months of work. They can also differ in the kind of knowledge that is needed to accomplish the task. Some tasks may require a certain niche knowledge, others a more common one. Tasks can also be distinguished by the effect they have on the crowd or the kind of motivations they work with.

There are multiple approaches on how to categorize these crowdsourcing tasks. Jeff Howe, author of crowdsourcing.com, based his approach on the way the process of crowdsourcing works. He came up with a list of four different types of crowdsourcing.
 * Crowd creation
 * This type of crowdsourcing might be the most well known one. The crowd works together to create something. Funder's of this kind of task ask people to be part in creating a new product. An example that will be further explained in detail is Threadless.com, a company that sells T-shirts. These T-shirts are designed by the crowd. With this kind of task being a creative process, there is no need of a certain knowledge and everyone can participate, which make these so well known.


 * Crowd voting
 * As the term suggests this kind of task includes a voting by the crowd. In this case the crowd does not participate in creating something but making a decision about a product or an idea. Crowd voting is a popular form of crowdsourcing because it is easy for the crowdsourcer to gather information. They can rely on only a few resources. It is also an easy task for the crowd. This results in a high level of participation which also benefits the crowdsourcer. A well known example for this kind of task is Facebook's Like! button. People indicate that they like a certain post and these votes are being used to display posts in a certain order.


 * Crowd wisdom
 * This category describes the kind of tasks that make use of the collective intelligence of the crowd. The crowdsourcing tasks' intention is to solve complex problems by utilizing the knowledge of a large group of people. These problems can be of scientific nature, concerning marketing strategies, or relate to search engines and recommendation systems. An example you often come across in online shops are recommendations like "People also bought...".


 * Crowd funding
 * Crowd funding platforms, like Kickstarter, help individuals or groups gain money to realize a project that otherwise would not come into effect. The crowd, by making donations, puts together a fund that can make these projects happen. The motivation to fund a project is the wish to see the project come to fruition. People support their friends and fans support people or projects they admire. Kickstarter adds an additional motivator by adding rewards depending on the amount of money the supporter gave.

David Alan Grier wrote a slightly different classification which consists of five different types that are based on their features and usage.
 * Crowdcontests - similar tasks to crowd voting
 * Macrotasks - larger tasks that take more time and result in bigger rewards
 * Microtasks - simple tasks in exchange of financial reward. They are often part of a larger project
 * Crowdfunding - see above
 * Self-organized crowds - the crowd organizes itself into teams and solves a problem. The team with the best solution wins

Skipso Labs put together a more comprehensive list of types that list Jeff Howe's classification as only one part of it. The article differentiates tasks by these four variables:
 * 1) What kind of task is performed?
 * 2) What is the kind of motivation to participate?
 * 3) How does the process of the problem solving work?
 * 4) What kind of problems are being solved?

This list gives a widespread overview off all the different strategies on how to crowdsource. The four given variables make it easy to choose the one that fits the needs of a certain project.

All these different approaches show that it is not an easy task to categorize crowdsourcing strategies. Many strategies overlap and have similarities, others are easier to extinguish.

Crowdsourcing Plattforms
When McDonalds asks their customers to design their own burger or Threadless asks to submit designs for their T-shirts they always have a place where to place this call to make a crowd aware of the crowdsourcing task. Other interested parties do not have their own resources to get a crowd's attention. This is what crowdsourcing platforms are for. They provide a place for crowdsourcers to make their open call. It is not only a place to make these calls though. They provide a marketplace where these calls can be made with a certain type of interface that helps making these calls. Those platforms also often come with a crowd. People who want to participate are registered and can use these platforms to find tasks they are interested in and which fit their skills. Another advantage of such platforms is that they often only cover a certain kind of task, depending on topics or knowledge.

An example is TopCoder, which concentrates solely on programming tasks.

This way it is easier for people to find tasks they want to participant in. Open calls will not go unnoticed by an overload of other tasks, which deal with topics that gain more interest. There are many different platforms. You will find more examples in further in this chapter.

Besides these platforms there is another one called CrowdFlower. This is a so called meta platform. CrowdFlower comes without its own marketplace but publishes registered tasks in third-party marketplaces. This way crowdsourcers do not have to search for the right place to make their call but can use this meta platform to make sure their calls get placed on one or more fitting platforms.

A critical view on crowdsourcing
As seen in the section about motivations above, this mechanism comes with a lot of advantages. Crowdsourcing can provide quick and cheap answers, can give first hand insight into customer needs, comes with a global pool of resources and is independent of money and time when volunteers are used. Crowdsourcing opens the possibility to heterogeneity of participants and can be a motivator for people to learn new skills. The question is whether these advantages are reason enough to overlook the disadvantages crowdsourcing brings with it.

A bothering aspect about crowdsourcing is the payoff for the crowd. Paid crowdsourcing tasks can violate minimum wage laws. Not only do people have different opinions on if these laws should apply to crowdsourcing participants but it also comes with a certain other problem. Crowdsourcing works worldwide and is not limited to a country. Minimum wages are different in each country though. This means that crowdsourcers might actually have to limit the crowd to a certain country.

Another disadvantage in crowdsourcing is also due to its payoff. On one hand payoffs might be a motivational factor to secure good results. On the other hand this can lead to the opposite, faulty results. Crowds who get paid per task might want to gain as much money as possible and therefore work rather fast than doing it well. The possibility of faulty results causes additional time and work. The question about how to filter faulty results and the filtering task itself might cause expenses that negate the originally saved expenses.

Another problem is that the project might fail. This can arise for different reasons. On the one hand the project might lack interest. On open call placed on the wrong platform, at the wrong time or missing motivational factors might cause this. On the other hand a project might get attention but lack in usable results. This can be caused by the before mentioned problem of payoff but also by misunderstandings regarding the content of the task given with the instructions.

Cause for these misunderstandings can be the lack of communication. Crowdsourcing tasks often lack interactions between client and crowd. If a project fails, it causes a throwback. Not only is the time spent on the project lost but it might be necessary to reconsider the whole crowdsourcing task. One might even consider that crowdsourcing does not work in this case.

One reason to use crowdsourcing is saving time and money. One has to consider for each task if this is really the case. As already listed, filtering results causes additional work. There are other tasks that can cause this additional workload too. Especially with microtasks, the management of the results might not be worth it. One might have to consider if bringing these results from smaller projects together to one bigger result is worth the effort.

Two more critical points regard the nature of the crowd. For one the crowd should be unbiased, especially in crowd votes. The crowd is often benefited from the outcome of such tasks and is therefore not unbiased. Crowds are also described as being a random group of individual people but in reality they might not always be that random. Not everyone has the same possibilities of access to be part of these crowds.

Considering to crowdsource a project is not an easy task. Experiments on the value of crowdsourcing like the one from Poetz and Schreier show its success. Research on this success like the one from Walter and Back cite what effects different approaches can have. One has to consider many aspects to make sure that it can be accomplished and that the resulting advantages surpass possible disadvantages. Even though crowdsourcing comes with a lot of challenges it is still widely used.

Examples
In the following section you will be given different example of crowdsouring tasks and platforms which does not only show how well spread this mechanism is but also its success.

Threadless
Threadless.com is an online website that mostly sells T-shirts but also other items like art prints. It was founded by Jake Nickell and Jacob DeHart in November 2000 after Nickell won a t-shirt designing competition against DeHart. Threadless makes use of crowdsourcing when it comes to choosing new designs for their shirts. Artists from all over the world are able to submit their designs and enter the online contest. It's then for the crowd to decide which design they would like to see in print. Every design is up for public vote over a seven day time span. The open call in this case is „Score designs“. Threadless openly asks every registered member to be part in this contest.

This is an example of crowdsourcing that shows how successful itself is and also the motivations for both the company and the participants. Today, Threadless sells millions of shirts each year, an indicator of the success of this crowdsourcing business. For this competition to start, Threadless needs artists to provide their designs. Their motivation is that they get paid once a design goes into print. This kind of creative process is what Jeff Howe calls „Crowd Creation“. Once a design is submitted, every registered member can vote for that design to get printed. The users' motivation to participate in this, is to be able to buy the design they like. This on the other hand means that Threadless knows that people are going to buy the shirts. For them it's a warranty that the shirt will sell.

As you can see with this example, both the business owner and the crowd benefit in different ways from this crowdsourcing task.

Amazon's Mechanical Turk
Amazon's Mechanical Turk is a marketplace for so called HITs, Human Intelligence Tasks. This platform provides a place for both the crowd, the so called Turk Worker, and the crowdsourcers, the so called Turk Requesters. Workers can use this platform to search for tasks that fit their interest. Requesters can make their calls to ask Workers to complete their HITs.

An advantage of amazon's system is that they offer to categorize workers by qualification. Workers can gain these qualifications by completing tests, some can be gained by requesting them. Others are only available through certain performances regarding completed HITs. Requester can then specify if a worker needs to meet a certain qualification to be able to work on their HIT. This way requesters can make sure that their HIT is being worked on by someone qualified for this certain job.

InnoCentiv
InnoCentive is the place to go when it comes to business, social, policy, scientific, or technical related problems. It was founded in 2001 by Jill Panetta, Jeff Hensley, Darren Carroll and Alpheus Bingham and became the "global leader in crowdsourcing innovation problems" They "specialize in connecting solution Seekers – commercial enterprises, public sector agencies, and nonprofit organizations – to problem Solvers to help solve their most pressing problems and innovate faster and better."

InnoCentive is a global marketplace with over 300,000 participants from around 200 countries. They state to have a success rate of 85%.

Further Examples

 * A concert film of Neil Young's acoustic concert at Carnegie Hall. The video has been stitched together from different videos from the attending crowd.
 * CrowdSPRING is an online marketplace that provides its costumers with all things design specific, like logos and graphic designs. The customer has to submit what he wants designed, the price he is willing to pay and the deadline. The crowd will then submit ideas from which the customer can choose his favorite or get his money back.
 * In order provide site guidelines that fit the tune with local cultures Facebook made use of crowdsourcing too. Facebook user helped translating Facebook for free.
 * The Google Image Labeler was a game that was used to help improve google's image search. Players had to provide tags for certain images to get points and google could make use of those tags.
 * ReCaptcha extends the known captcha interface that provides the user with a blurry image and demands an input to protect a website from spam. ReCaptcha added a second image which was used to help digitze text from books.

For more examples see a list of projects on wikipedia. Also check the crowdsourcing example from the lecture (47:23)

Conclusion
Outsourcing tasks is something that has been done for years. Doing this via the web is not a recent idea but the term crowdsoucring only came to be in 2006 and was coined by Jeff Howe. Since then, this mechanism is getting more and more attention over the years and is a well known and often used method in today's business.

In this chapter we learned the advantages that come with crowdsourcing tasks. It can provide quick and cheap answers, can give first hand insight into customer needs and comes with a global pool of resources. Crowdsourcing also comes with its challenges like quality assurance, separation between bias and error, or the time needed to obtain results. Examples have shown us that we come across this mechanism in our networked life multiple times and that it is a favored way to collect data and solve problems due to its success.

Being a relatively new term, crowdsourcing still leads to discussions about its nature, like how to categorize its types and legal matters like minimum wages.

Quizzes
{What are reasons to consider crowdsourcing for a project?} + Being able to reach a broad array of talent - The project is guaranteed to be successful + To save expenses

{What motivates users to participate in a crowdsourcing project?} - It is well paid + One can train his skills + The possibility of prestige or status + To be able to work independently

{What is needed for a crowdsourcing project?} - A crowd of people with uniform knowledge + On open call has to be made + The task has to entail mutual benefits + A crowd of variable size

{What are advantages of crowdsourcing?} - Quality of data is guaranteed + No need to train experts internally - Entails close interaction between client and crowd + It is location independent - The ability to gather large amount of data

{Which of these statements fit the principle of "the wisdom of the crowd"?} + The group needs to be diverse - In every group is at least one person who knows the answer which guarantees the success of the group + The group as a whole can make a better decision than a single member of this group - The principle holds in every context - Centralization of the crowd is necessary + Each individual has to make a choice independently