Mechanical Turk is a great system for outsourcing repetitive, tedious tasks. People use it (and other similar services) for all sorts of things: filling out surveys, getting research done, data analysis & acquisition, etc. You submit your task, indicate how many times you want it done and how much you’ll pay, and people do the work for you. There are over 138,000 tasks available on Mechanical Turk right now. What makes it particularly interesting for startups is that human power has become super inexpensive, practical and easy to use.
CardMunch (which was acquired by LinkedIn) is a service for business card scanning. You scan in a card and shortly thereafter you get the data back to approve, and the new contact is saved in your contact list. The key is that CardMunch is using humans to look at the cards and enter in the data. They built their entire business around crowdsourcing; their core value proposition is the human intervention.
For Lean Startups this is extremely useful. If you can “mechanical turk” something, why code it? You might not use Mechanical Turk specifically, you might be doing the work manually behind the scenes. Aardvark (which sold to Google for ~$50M) is a Q&A platform. The founders were answering questions at the outset to test out the user experience, help them learn and build traction.
Developers often leap to a technical solution to every problem, but if you take a rigorous Lean Startup approach and follow the customer development process, you put yourself in a position to genuinely think about the simplest possible solution to the problem you’ve identified. Customers don’t care how you get things done – just that you get it done and solve their pain. If you start with the solution it’s almost always going to come out too complex. If you start with the problem (and subsequently realize that the customer doesn’t care if you’ve got crazy ass algorithms behind the scenes with software that took two years to build), you have a chance at building the smallest solution possible. And that’s likely going to save you a lot of time getting your solution into customers’ hands and validating if you’re on the right track. You might even discover a new target market or business model. We’re now seeing more and more startups tackling things like translation and editing through crowdsourcing (or “mechanical turking it”), as opposed to coming up with super-fantastical technology.
When just starting out, give some thought to what you can do manually, even if it’s just temporary. Don’t rush to code and build features. You should be looking for a solution to a problem that people care a ton about. You shouldn’t be looking for an excuse to build software or write code.