Deciding what your product should do next: How to "Build Measure Learn" in 8 Steps

Eric Ries first introduced the idea of making incremental, small changes to your product to get feedback faster -- with the 3-step “build-measure-learn” cycle -- before investing too much money into the idea, in his book The Lean Startup.

Here’s a more specific, 8-step iterative process, that you can use at any stage of your product development to help you do this -- or what you can do to not break your checkout process.

  1. List all ideas you’re considering exploring and specify what a 1% (out of 100%) simplified solution would look like

  2. Prioritize the list to narrow it down to 1 idea, based on: business impact, time to deliverable, data availability and how interconnected this project is to the others

  • Example: Beginning to charge users for (an aspect of) your product

  1. Define what success metric will look like for Step 2 (see Phase 2 for more details)

  • Example (cont'd): % unique users that would buy product at $10/month

  1. Calculate what the answer to Step 3 looks like now (0 is OK here) -- or specify what target we’d like to hit in what time frame, to see if the experiment succeeded [Lean Analytics book]

  • Example (cont'd): We’re currently at 0% because we haven’t charged anyone yet. We want to get to 10 users by the end of the week.

  1. Build an experiment to test that hypothesis”. Gary Livingston suggests testing the hypothesis by evaluating pre-orders/waitlist for the feature/product before building the feature/product:

    1. Create a landing page to sell the product to your ideal customers, as if it exists.

    2. Take them through the complete checkout process, which includes them placing the order.

    3. Once the order has been “placed”, don’t save any of this information -- and tell them so.

    4. Then ask them to join the waitlist for the offer and record the waitlist opt-in.

  1. Roll-out to a small % of the ideal audience see what you’ve built in Step 5

  • Example (cont'd): While the landing page is available to everyone, the marketing announcement is sent to 1% of the target audience.

  1. Measure and monitor results of Step 6

      • Please note: In practice, the majority of companies tend to make the offer available to everyone -- and may not have monitoring in place to see if something went wrong. We’ve all been there: trying to buy something online, but can’t actually complete the process because of broken links -- and go elsewhere.

  2. Learn from experiment to decide next steps: