Lean Experiment Techniques
There are many techniques to consider when designing an effective lean startup experiment, so we decided to capture a few of our favorites below. Mix-and-match each example to create a powerful experimental learning vehicle for your next concept. Even better, see if you can recognize these lean startup experiment examples next time you’re using your favorite application or website. Good luck!
Remember, the most effective experiments measure real behavior, while also delivering real value to the customer.
Manually perform tasks related to delivering the value of your product or service. Eventually you will automate and optimize the process you perform manually, but it’s likely you will move faster by simply performing tasks manually.
Example: P-2-P Payments platform (name is confidential). The founding team created a peer-to-peer payment workflow via a web “responsive” application, where user can request payments from friend and family. This workflow was presented to the user as functional website. In reality, the founding team manually processed all requests by hand, meaning they manually sent email notifications, tracked payment requests and captured billing information.
Mechanical Turk (or Wizard of Oz)
For offerings where a complex backend is required, such as databases, algorithms or complex engineering, consider simulating this backend with real people or existing apps. It’s likely you can mimic the “engine” behind the scenes. Even if this takes longer for the customer to receive an answer, you will avoid wasting precious time building features the customer does not want. This is especially good for support centers.
Example: See it in action at a call center – Why don’t you Just tell me which movie.
Imposter Judo (Boomerang)
If a similar ideas to yours already exist in market, you can use these as a quick and simple way to gather feedback. You might ask customers to sign-up and give you feedback on a competitors website, or repackage an existing product. This is especially effective when selling physical products, or showing early static mockups.
Example: In the early days of Zappos, the founders simply purchased shoes as needed from local shoe retailers, instead of stocking their own inventory. This allowed Zappos to test their idea fast and cheap, before investing ion their own inventory.
Sell a physical version of the product, even if your final product will be digital. This is especially effective for information or data-based products such as customer lists or how-to guides.
Example: When selling a digital information product the experiment team gathered early versions of this data “by hand”, then produced a printed report for the test customers. This printed report provided real value, and was used to gather feedback. Eventually, the team created a digital version of this report.
In order to test whether your customers will actually purchase your products, and for how much, you can deploy a simulated “purchase now” experience. This may take the form of a simple e-commerce check-out, or perhaps a letter of intent request. Once the user begins the purchase process, you can simply respond with a “out of stock” message, or other elegant way to proceed such as simply not billing the customer at all.
Example: You will encounter this strategy the team designed a website for selling a subscription based physical product. To gauge interest, the team quickly created a web based purchase form simulating the check-out process, so the customer believed they were purchasing the product. However, no billing occurred since the form did not connect to payment processing…
Refers to any experiment where you actually make your experience more difficult to use, in an effort to gauge real interest. The higher the hurdle, the more validity you can attribute to your results.
Example: Design a sign-up process with multiple, difficult to answer questions, which adds friction to the sign-up process. This way you only capture the most passionate customers for your early testing. Over time, you can reduce the friction in order to improve your funnel.
Record a “real life” scenario using your product, where you edit the video to create the illusion your product is real. Then, use the video as part of your landing page or marketing message. This is especially useful for products where the behavior can be easily simulated (see other experiment types).