Diet for a large Crowd: what to feed a hungry herd of workers
Okay, so you've caught some of the buzz and you're ready to do some real Crowdsourcing. Not the volunteer kind of 'vote-on-your-favorite-idea' polling or "invent-our-next-sneaker" contests that might be termed crowdsourcing-lite. Here I'm talking about big barn-raising, roll-up-your-sleeves Crowdsourcing projects that heave under the weight of their own data, require many humans, and compensate the crowd for the results.
Before leaping into the throng, you should determine whether your project will actually benefit from division of labor and scale. The project should either be too large for an in-house team to do (e.g. 1000 graphics/listings/locations) or require many people performing the same task to achieve a varied result (e.g. 1000 variations on a logo treatment/taglines/domain names).
Here are the basic Crowdsourcing ingredients for labor outsourcing: tasks that can be equally divided, described clearly, take a predicable amount of time to complete, and can be submitted (or verified) electronically.
Labor-crowds prefer a simply portioned data-rich diet so not every project is naturally Crowd-friendly. Complex entrées with lots of ingredients don't go down so smoothly and can lead to poor results. Consider creating a technical manual or designing a database. These kinds of projects benefit from the consistency of a primary chef d'cuisine. However, many of the building-block components of these projects can be accomplished faster by tapping the Crowd.
To date, the mainstay of LaaS (Labor as a Service) crowdsourcing has been focused around aggregating tables of data--collecting research, verifying, correcting, tagging, evaluating--anything that fits neatly into a spreadsheet. This kind of work is simple to divide amongst many people: one row of data per task. Many crowdsourcing platforms, including JobSpooler, support this kind of data collection. It is well-suited to building lists or reviewing large quantities of existing data.
The second most common use is transcription and translation. Humans do a much better job at interpreting language than computers do but even the fastest typists can't keep up with hundreds of people all working on small chunks of a large document or media file. While I would not endorse the use crowdsourcing to translate "Catcher in the Rye", a crowd-powered translation of many business documents or interviews is quite adequate.
Beyond these basic tasks, there are numerous other lesser-known categories of work that have been run successfully on our platform. By combining data collection with geo-location filters on the JobSpooler platform, field-data collection has been performed more efficiently. Web content (articles, blogs, tweets, ad copy) plays well to a crowd provided that the crowd has been pre-qualified and there are clear guidelines on the content. Presentation media particularly template-oriented variations such as display ads, buttons, podcasts, videos, charts and graphs can be created on a human assembly line using a crowdsourcing platform.
Small blocks of programming code (scripts, functions, algorithms, modules) make good bite-sized projects that can be worked on in parallel without slowing down core development. By contrast, not every problem can be solved with a clever algorithm. Some needle-in-a-haystack problems need the power of human intelligence and interpretation. When trying to find relevancy and context in a mountain of data (does this paragraph have positive or negative association with the product? Are these facts and footnotes substantiated?) We are a long way from having machines make good decisions requiring nuance while humans do it routinely.
As the quantity and diversity of essential business data continues to expand, it becomes necessary to harness large teams of humans and computers to create, interpret, process and respond in more efficient ways. Much like mechanical assembly lines transformed productivity in the 20th century, virtual assembly-lines will become essential to the knowledge economy in the current century.