Bold (23 page)

That's when Ahn and his colleagues hit on another dual-use idea, realizing they could teach people new languages at the same time as these people were translating the web. “At any one time,” says Ahn, “there are about 1.2 billion people out there trying to learn a foreign language. Stuff needs to be translated, so why can't we get the people who are learning a foreign language to translate this stuff for us?”

This was the birth of Duolingo, both a language education website and a translation game that really works. “It's been super-successful,” says Ahn. “People on Duolingo learn a foreign language as well as they do with any other language translation program. But because they're translating real content (say,
New York Times
articles),
it's inherently interesting—people are motivated by the quality of the content. And because we're using multiple sources for every translation, the results we get are as accurate as those done by professional language translators.”

And far, far cheaper.

For example, right now, only 20 percent of Wikipedia exists in Spanish. If you were going to go out and hire translators, the cost of translating the remaining 80 percent would be roughly $50 million. It would also take years and years and years to finish the job. Duolingo, meanwhile, can do it for free, in about five weeks, with about 100,000 users. And as of today, the site has about 300,000 users.

So, while the point of the Freelancer and the Tongal case studies were to explore two different crowdsourcing platforms that offer today's entrepreneur astounding leverage, the point of reCAPTCHA and Duolingo is the inverse—an example of the kind of crowdsourcing platform a bold entrepreneur might be interested in creating, the kind that both makes money and betters the world at the same time.

As you can see from our case studies, crowdsourcing is a diverse and growing field, with more novel applications being dreamed up every day. So, before we dive into lessons learned, to give you a better sense of what's going on, I've broken this section into four of the most common uses for crowdsourcing and provided a short explanation for each.

Tasks are work. Crowdsourcing tasks means getting someone somewhere to do the work for you. In most cases, you pay only if you like the result. In other cases, you can specify what you're willing to pay per task or let the crowd marketplace compete for your business. In this taxonomy, tasks come in two basic flavors: micro and macro.

Microtasks are bite-size, well-defined chunks of work that can either independently solve a small problem or, when combined with many other microtasks (for example, reCAPTCHA), collectively solve much larger problems. This means that one of the most important questions to answer when approaching crowdsourcing is whether the work can be broken down into smaller, simpler units. If so, what is the simplest microtask that can be defined and distributed? For example, a while back I wanted to determine whether
T
ime
magazine cover articles have gotten more negative during the past sixty years. At first I had my executive assistant start in 1945 and group articles into positive, neutral, or negative categories. After a day's work, she had barely put a dent in the problem. That's when I decided to turn to the crowd. By offering $0.05 per categorization, I got the entire 65 years' worth of issues, roughly 3,000 in total, done for under $200.

I used Amazon's site Mechanical Turk (
www.mturk.com
) to get those magazine covers analyzed. While MTURK isn't all that useful for more complicated jobs, it is where to go to get simple, quick tasks done fast. Aggregation and classification jobs tend to be popular uses. Aggregate photographs of red trucks, for example, or write product descriptions, or perform sentiment analysis exercises on thousands of Tweets. Requesters (you) post tasks known as HITs (human intelligence tasks) while workers (called providers) browse among existing tasks and complete them for a monetary payment.
¹⁶

Another microtask site that I've previously relied upon (and with great result) is Fiverr (
www.fiverr.com
), an online marketplace offering microtasks starting at $5. Typical services include voiceovers, animations, crafts, promotional videos, and art. Offerings can get wacky, for example: “I will print and hand out 500 flyers for you in Toronto, Canada, for $5” or “I will draw you in my caricature style for $5.” You can scan the site for services of interest to you, or make a request. I've had my hand-drawn sketches turned into great digital art for five dollars.
¹⁷

In contrast to the above microtasks, macrotasks are jobs that (a) can't be broken down, (b) can be done independently, (c) require some type of specific skillset or thought process, (d)
are additive and dependent on already completed work for the task, and (e) take a discrete, fixed amount of time to complete. There are a number of different companies that allow you to crowdsource macrotasks, with the aforementioned
Freelancer.com
being the largest. It's also important to remember what
Freelancer.com
founder Matt Barrie said about the diversity of work being done: “This isn't just plumbing and pest control. We've got PhDs on the site. I've seen both quantum physics and aerospace jobs handled perfectly.”
¹⁸

Even better,
Freelancer.com
is only one of a myriad of macrotask sites. For a detailed list of the latest sites with examples of how to use them, please see
www.AbundanceHub.com
.

An asset is anything that provides value to you and your business—that is to say, applications, websites, videos, software, designs, algorithms, marketing materials, physical goods, machinery, and technical plans. To understand how to crowdsource assets, I've broken things into two different categories:
creative
and
operational assets.

Creative assets
include a wide variety of design-based assets such as logos, videos, website designs, CAD models, marketing plans, and advertising plans. We've already mentioned two of my favorite creative asset development sites: Tongal (
www.tongal.com
), which can make you a TV or Internet ad in weeks instead of months and at a cost that is usually about a tenth the industry average, and 99Designs (
www.99designs.com
), which provides crowdsourced graphic design (logos, apps, web pages, infographics, blogs, and more). I've used 99designs repeatedly and have found that contests usually yield between 25 and 400 entries, depending on purse size. Even better, if you don't like any of them, 99Designs will provide a full refund.

Operational assets
, meanwhile, are those things required for business to run effectively. For example, if you are running a software business, these assets include the algorithms powering your software, your
database architecture and server implementation, technical designs, models, and frameworks that organize deal flow and customer acquisition strategies, and so on. A number of companies allow you to crowdsource the creation of operational assets. In fact, doing so is one of the keys to becoming a data-driven, exponential organization.

A great example of this is TopCoder (
www.topcoder.com
). You've probably heard about hackathons—those mysterious tournaments where coders compete to see who can hack together the best piece of software in a weekend. Well, with TopCoder, now you can have over 600,000 developers, designers, and data scientists hacking away to create solutions just for you. In fields like software and algorithm development, where there are many ways to solve a problem, having multiple submissions lets you compare performance metrics and choose the best one.

Or take Gigwalk, a crowdsourced information-gathering platform that pays a small denomination to incentivize the crowd (i.e., anyone who has the Gigwalk app) to perform a simple task at a particular place and time. “Crowdsourced platforms are being quickly adopted in the retail and consumer products industry,” says Marcus Shingles, a principal with Deloitte Consulting.

Retailers and consumer products manufacturers have a challenging time obtaining vital information about how their products are being sold, merchandised, and priced. Is the product in stock? Is the price correct relative to the in-store promotion? Is the display and promotional signage in place? What are competitor products selling for? These are all very key causal data variables that can be converted into insights that a retailer and manufacturer can use to optimize inventory, promotions, and overall sales. It's not always cost effective to have store employees monitor this data, nor is it efficient to send the manufacturer's sales reps into each store to check. Crowdsourced platforms utilize the everyday consumer instead. The consumer gets paid a nominal fee, say five dollars, for taking five minutes to take a photo of a shelf, and the retailer and manufacturer only pay for the five minutes of time needed to collect a specific data point. In our
pilots, on average, the crowd was able to deliver this data in less than an hour, and for five to eight dollars per request, across thousands of posted tasks. That data is then brought back into the retailer's and manufacturer's operational systems, in which data visualization techniques are used to make sense of the information collected so timely decisions can be made.
¹⁹

Of course, in this age of exponentials, we are generating more data than ever before. Unfortunately, not everyone knows how to tease out valuable insights from this deluge. Enter companies like Kaggle (
www.kaggle.com
) and TopCoder (
www.topcoder.com
), both of which are crowdsourcing, data-mining competition platforms that allow you to define your goal/desired insight, set a monetary prize, upload your data, and watch as hordes of data scientists (tens of thousands, to be exact) figure out the best way to sort through it. The best algorithm wins. The reward levels vary from kudos or zero dollars to hundreds of thousands of dollars from bigger companies.

And for exponential entrepreneurs, not relying on the advantages of data is no longer an option. Deloitte Consulting's Chief Innovation Officer Andrew Vaz explains: “As Big Data overwhelms traditional computing and analytical tools, the combination of A.I. and Big Data will create an insights ‘arms race,' where competitive advantage will be dominated by individuals and organizations that capitalize on these emerging technologies.”
²⁰

Insights are invaluable to your business. They can shape the goals and operations of the entire company, dramatically improve and optimize performance, and provide you with counterintuitive ideas or hidden data for a strategic advantage over competitors. When it comes to crowdsourcing insights, there are two main variants:
testing
and
discovery
.

Testing-based insights
often come from examining existing assumptions and current best practices. These include surveys, A/B testing, representative sampling, customer feedback, case study abstractions, and focus groups, among others. When running a test, focus on asking one specific question and using the data or resources at your disposal to frame the question appropriately.

For example, if you are in the world of software development, you know that testing your wares can be incredibly tedious, difficult, and time-consuming. There is no room for error, so you have to have as many people as possible hunting for bugs before launch. No problem. uTest (
www.utest.com
) provides a massive community of “professional testers” who run functional, usability, localization, load, and security tests on your code. By leveraging the crowd and data from previous tests, they optimize and simplify the process, making it less expensive, while offering reduced churn rates, better functionality, and quicker time to market.

Creatives are also getting in on the insight game. Take ReverbNation (
www.reverbnation.com
), a music distribution, publishing, and crowdsourced testing platform. Say you're an aspiring musician. You've produced a few songs, but before spending money on paid advertising or management, you want to see whether anyone actually likes your music. Now you can have songs rated and reviewed long before you actually go to market.

The other side of this insight equation is the crowdsourcing of
discovery-based insights.
This can mean a few things. You can ask the crowd for their interpretation of a particular problem or question. For example, Genius asks the crowd to annotate song lyrics. Kaggle outlines a problem that its community solves with original algorithms. Or perhaps more simply, you can provide a platform for the crowd to come up with their own ideas and inventions, as Quirky does with its invention network and Threadless does with its T-shirt design competitions.

Discovery-based insights can be as simple as asking the crowd for answers and paying attention as the best solutions, designs, and inventions
bubble their way to the top. As a personal example, in the spring of 2014, in preparation for rereleasing
Abundance
as a paperback, I asked the crowd to help me discover new evidence of abundance that I could include in the appendix. People emailed me their data, charts, and graphics to [email protected] (new submissions are always welcome). The response was extraordinary and yielded considerable evidence of our continued forward progress.

So there you have a top-level overview of the crowdsourcing space. Since this is an area that will change rapidly as new players enter and new AI capabilities come online, I wanted to point you to a few industry websites worth visiting that will keep the taxonomy fresh and up-to-date.

AbundanceHub.com
:
This is the place I'll be posting my own experiences, providing updates on my lessons learned (successes, experiments, and failures) and working with entrepreneurs interested in creating wealth while creating a world of abundance. AbundanceHub content is driven by the Abundance360 community, my mastermind group of entrepreneurs whom I've committed to coach over a twenty-five-year period.