This week we will be looking at the wealth of companies and projects that use crowdsourcing as part of their business. Here’s a list of companies and projects that you could research:
5 Calls, 99Designs, Airbnb, Apple HealthKit/ResearchKit, Benefunder, Change.org, CivilServant, Clickworker, CloudFactory, Couchsurfing, Coursera, CrowdAI.org, CrowdMed, Crowds On Demand, Daemo, Dataturks, Defined Crowd, Duolingo, EatWith, eBay buyer/seller ratings, edX, EyeWire, Facewatch, Field Agent, Figure Eight, Flattr, Freelancer, Genius, GoFundMe, Good Judgment Open, Google Crowdsource app, Iceland’s Crowdsourced Constitution, The iLabour Project, IndieGoGo, Indivisible, Innocentive, Instacart, Invisible Boyfriend, iStockPhoto, Intrade, Iowa Election Markets, Jovoto, Kaggle, Kickstarter, Kiva, leadGenius, Lending Club, Lyft, Microworkers, Mighty AI, Netflix prize, New York Times, OpenStreetMap, Orchestra, PatientsLikeMe, Patreon, Polis, Postmates, PredictIt, PredictWise, Premise, Prolific, PublicStuff, Quora, Qualtrics, Samasource, Scale, Scistarter, Sunshine, TaskRabbit, Threadless, Thumbtack, Topcoder, TurkOpticon, TurkPrime, TurkServer, Uber, Upwork, Ushahidi, vTaiwan, Waze, We the People, Wikipedia’s AI helper, XPRIZE, Yelp, Zensors, Zooniverse.
This assignment has two deliverables:
This assignment should be done in pairs.
First, please sign up for a company or project. Please do not pick a company that another team has already signed up for. If there’s a company that you’d like to research that isn’t on the list, you are welcome to ask the instructor or the TA to ask if it is OK.
You should independently research the company, and then fill in your answers about it in the Homework 2 questionnaire on Gradescope. Your short answers to these questions on Tuesday, February 5, 2019. Remember to save your survey answers on your own computer, so that you can have a copy to use when you do your video profile.
You will prepare a short presentation, optionally a 5 to 7 minute video presentation, about your company. Your presentation should address the following questions:
Your presentation is due on Tuesday, February 12, 2019. To turn in your video, please upload it to Vimeo. Give us the link to your presentation or video using the Homework 2 presentation form. Make sure that your video is publicly viewable or that you give us a password to view it on the web form.
If you have managed to make it this far in life without having to sign up for accounts for things on the internet, here are more detailed instructions:
What is the difference between Users and Contributors?
Many of you seemed confused about what constitutes a user vs. a contributor. In many crowdsourcing companies, this line is blurry- or nonexistent! Many companies you looked at fit what can be thought of as a “data-mining” model (e.g. Yelp, Foursquare), in which the primary service being provided depends on using data and modeling observed patterns of behavior– e.g. to target ads or to recommend products. In these cases, the users are the contributors– everyone who participates provides data, and everyone uses everyone else’s data. This is an awesome crowdsourcing model because it is (ideally) self-sustaining and self-incentivizing. The more a person contributes (by providing more data about themselves), the better product they receive (e.g. better recommendations).
How does my company do Aggregation?
Many of the companies we look at fit into the “match-making” or “marketplace” model (e.g. Uber, Airbnb, Etsy). This is a very common case in which the company is simply working to match supply with demand, where the suppliers (e.g. drivers, in the case of Uber) and the consumers (riders) are distributed all over, and may have trouble finding one another otherwise. Here, rating systems etc. are one piece of the platform that might require aggregating, but arguably the more interesting aggregation problem is how to match a supplier with a consumer. Is this through preference-based recommendation systems (as in the data-mining model discussed above), through location-based matching, through bidding/price? Making these matches is non-trival, but crucial to making the company function.
Does this update a previous service or business model, or is it completely new?
For this question, more than a third of you said “completely new.” Try to think broadly when you are considering where these crowdsourcing companies fit into the economy as a whole. (Think Silicon Valley’s favorite buzzword: “disruption.”) Most of these services being provided are not completely new, but they are directly competing with a service that was traditionally supplied by a brick-and-mortar company in a more centralized manner. Uber competes with taxi companies, Airbnb with hotel chains, Coursera with higher education universities. This is part of what makes crowdsourcing so exciting!
Is getting venture capital funding is a business model?
No. Smart investors don’t invest in a company who says their plan for generating revenue is to “get money from investors.” When in doubt about the company’s business plan, it is probably “ads”. :-)
How to make a group submission on Gradescope?
One group member can submit answers just as you do in previous homeworks. Then at the submitted page where you could see all your answers and points distribution, there is an option on the top right for you to edit group member. You can add your partner there with his or her Gradescope Id inside this course. Both of you with share the same submission and get the same grading. The submission is required to be made by only one team member.
Below are the questions that you will be asked to answer about the company or project that you are profiling. Please turn in your answers using the Homework 2 survey on Gradescope.
We recommend that you save your survey answers in a file on your own computer, typing them directly into Gradescope, so that you’ll have a copy after you press the submit button.
The written survey is worth approximately 10% of your overall grade in the course. Please answer the survey questions thoughtfully. Extra credit would worth approximately 5% of the overall grade in the course.
Video profile: http://player.vimeo.com/video/154886545?password=nets213"Who were the founders? Peter Weijmarshausen, Robert Schouwenburg and Marleen Vogelaar
When was it started? Company was started in 2007-2008 in the Netherlands
Does it have any interesting origin story? It started as a spinoff from Royal Philips Electronics in the Netherlands in 2007.
What services does Shapeways provide? Shapeways provides a 3D printing service in many materials and finishes. Anyone can upload a 3D model to the website and someone may order to print that object. Users can also sell those designs and find designers to collaborate with to create products. Shapeways will print those objects for a fee on top of the price they set. Recently they have partnered with businesses offering 3D scanning services in order to make 3D scanning technology accessible to their customers. They also offer tutorials and tools to help customers design 3D models.
Who uses the services? A Jewelry designer who uploads my models on Shapeways' marketplace. People can buy his/her designs in different printed materials and he/she can charge any price for a design. For each sale he/she makes his/her listed price minus the quoted print cost from Shapeways.
Designers and entrepreneurs design and sell products while anyone interested in non mass-manufactured products buy from the marketplace. These tend to be products that are not yet offered by any online retailer and solve a very specific issue for the user.
The incentive scheme would still work, because it would mean that there would be more buyers and, therefore, more money for the sellers. Quality may decrease, but they could potentially offer different printing qualities.
Video profile: http://player.vimeo.com/video/154945640?password=welcome"Who were the founders? Leigh Drogen
When was it started? 2011
Does it have any interesting origin story? Leigh Drogen was previously an analyst at a quantitative hedge fund that used many strategies around earnings events. Every quarter, Leigh would see the same thing - companies would beat their earnings estimates given by Wall Street almost 100% of the time. Upon further inspection, he saw that Wall Street analysts are incentivized to low ball their estimates for companies due to the relationships that they have, and he saw the opportunity to make the estimates a more fair representation of what will truly happen. He quit his job at the hedge fund and started Estimize, with the drive to make estimates more democratic and transparent for everyone.
What services does Estimize provide? Estimize provides multiple services through earnings estimates. The first is the sheer raw data of estimates. Estimize collects estimates for earnings per share (EPS) and revenue for nearly every US Exchange listed company. The raw data that the company collects comes from individuals who sign up for the company's website. These individuals can be professionals, students, independent investors, research analysts, etc. Then, Estimize generates a consensus, which is an average of every valid estimate, into one singular number. To sell the data, Estimize provides 3 spreadsheets- the first includes a list of every user on estimize along with their biographical data, the second includes every estimate that has been made on the website, and the third includes every release, or every earnings event that has occured, with both the estimize consensus for EPS and revenue, as well as the Wall Street estimates.
The second main service that Estimize provides is a stock screener. On this screener, individuals are able to screen US listed stocks for specific analytics in their earnings estimates. They are able to search for analytics such as % of the time Estimize is more accurate than the Wall Street consensus, or % of the time company beats Estimize consensus. This stock screener is unlike any other stock screener already available, because no other data sets exist like Estimize's, so it is novel.
Despite all of these incentives, when polled, the contributors cited wanting to build a more accurate data set as the
Secondly, Estimize runs many more complex algorithms to further ensure the quality of its data. Estimize runs algorithms that detect various details, such as the amount of time spent on the website before making the estimate, how accurate the user has been on the past, what other parts of Estimize has the user recently clicked to before submitting this estimate, did the user only change the default estimate by a little before submitting, etc. All of this results in a simple boolean, true or false, as to whether the estimate was flagged. An analysis of their data shows that nearly 1 in 10,000 estimates needs to be flagged, so users typically do not submit false estimates. Of those that are reported, nearly 999/1000 are due to mechanical errors by humans - someone submits $15 instead of $1.50 for earnings per share. Barely any estimates are truly submitted with the intent of gaming the system.
Graphs display individuals' estimates on the website, along with a list that shows every estimate made.
2) Users contribute (as said above) through the front end of the website, or through software that directly hooks to their financial terminal (although this is about 2% of users).
3) Users get to see what the Estimize consensus is for any given stock, along with being able to build a track record for their history. They are also able to win prizes or money.
Other ways the company generates revenue is through the screener, which allows individuals to manage lists of stocks with given characteristics around earnings (see previous answer), and this screener is sold for a monthly fee depending on the assets the individual or company manages.
Finally, Estimize generates revenue through its leagues feature - companies pay Estimize both in cash as well as through prizes to sponsor a league and have their name appear on the page.