Fake it until you are ready to make it.

It is a variant of the concierge model. The basic premise is that customers care about the finished story of using your product or service, not the product or service by itself.

In the Wizard of Oz there's a guy behind the curtains who is controlling all the machinery. But you never get to see him. We can similarly deliver value to customers by cobbling together a bunch of stuff in the middle where the customer never gets to see it. We do something similar in the concierge. The difference here is the interface to the customer: in the concierge model you provide a customer facing high touch interaction while in the wizard of Oz model the interface gives the appearance of automation.

e.g. Aardvark built a social search engine that was able to root search request to actual people that you trusted versus relying on a keyword based search engine. The premise was that people would be able to give you more relevant answers to certain types of questions such as where to go biking around the Golden Gate Bridge.

They envisioned building a complex software based solution using NLP algorithms and social graphs. But they didn't start there. Their initial MVP was a live landing page in which when you entered a question in a box, it was routed to actual human beings and mechanical Turk which is a service offered by Amazon where you pay people to answer questions for you.

Even though the founders had a lot of experience in building complex search engines, they wanted to first test if people would even use a service like this and on the other hand if people could actually deliver superior search results than a keyword based search engine. In the process, they further learned about the nuances of how people were asking questions, what topics they were asking questions about and what types of responses drove most value.

So on the surface, this hiding of service internals looks a lot like the concierge model. But the critical difference, again, is that in the concierge model there is a visible human interface to the service while in the Wizard of Oz model there isn't. As their learning grew, the Aadvark founders also began to automate the service internals and several steps replacing them with software. While they got pretty far with this automation, at some point they were acquired by Google.

e.g. Another great example of the Wizard of Oz case study is Zappos which was the biggest online shoe retailer and was acquired by Amazon for 1.2 billion dollars in 2009. The founder of the company had a vision of delivering a superior customer experience. And he believed that would ultimately help him build differentiation and a succesful business. But he didn't start by stocking up shoes and building a complex e-commerce backend. Instead, he went to local retail shoe stores and took pictures of their shoes, which he put on his website. Once orders started to flow in, he bought the shoes from the same retail stores and shipped them manually under the Zappos brand. Again, not scalable in the beginning. But once he was able to test the riskiest part of his business, which was a superior buying experience, he then started automating the process bit by bit.