Lean Startups: Part 2 - The Magic Test

As discussed in Part 1 of Lean Startups, learning quickly is imperative in bringing an idea to a productive and successful product. An important piece of the learn, measure, build loop is learning from the very start of any idea. The question is, how do you test your idea and learn from potential response? Enter: The Magic Test.

The Magic Test--a recipe designed and supported by Eric Ries--entails attempting to sell your idea as "magic". Before jumping into discussions about features and functionality, think of your product as solving one or two major pain points magically. You don't need to tell people how you're going to solve their problems just yet--at least not until you know that people are interested in that problem even being solved. Simply put, if people won't buy magic to fix their problem, then they won't buy your product.

Here's my suggested process on setting up and running a magic test.

The Process

  1. Set up a landing page. You can design/develop your own, or you can use services like LaunchRock, Unbounce, My Beta List, or Prefinery. Each of these tools provide different options around analytics, A/B testing, and compatibility with third-party tools. Based on these elements, choose the tool that fits your needs best.
  2. Sell your magic. Fill your landing page with minimal, succinct, and powerful content such as: "Hate paying your bills? Then don't do it anymore! Sign up to be on our beta list." The key here is to identify the problem you're attempting to solve, making it clear that you'll solve it (not how, just that you will), and then giving them incentive to "sign up".
  3. Get traffic. There are several marketing channels that can drive traffic to your site. But, one of the best drivers of unbiased traffic in your target market is PPC (pay-per-click) driven traffic. Set aside a budget for PPC by doing some keyword research (using tools like the Google Keyword Tool) and pull the trigger. While PPC is often the most effective, it still might not be the appropriate channel if you're attempting to penetrate a new market or developing a service that no one might be searching for. In this case, utilizing an advertising content network might be better suited for driving traffic to your landing page.
  4. Learn and measure. Use whatever analytics tools you've integrated to evaluate the percentage of visitors who actually sign up. If this is a high percentage, then great, move forward with the problem set/idea. If the percentage is low, you have something to think about. This could lead to pivoting your idea, it doesn't mean you have to scrap it.
  5. Rinse, wash, repeat. The great piece of this method is that it can be repeated to test specific components of your idea--such as features and price points--rather than just the general idea itself. As you repeat the above steps, get more specific. How are you going to solve the problem? How much is it going to cost? Integrating A/B testing and analyzing differences in percentages provide even greater insight into how you should design and price the product. Now you're no longer selling magic; you're selling your product.

Some examples of startups who have used similar testing methods include unseat.me, Fetchnotes, and Graphicly.

In the end, you will be taking a leap of faith regardless of how much testing happens. But why not minimize the distance you have to leap? Don't spend 6 months developing a product only to learn no one cares. Start learning now--test your magic.

Zach is a senior digital strategist, combining client-focused business acumen with creative digital ideas. He helps bring on new clients and ensures their success, including Discovery Channel, ESPN, Dick's Sporting Goods, and POLITICO.

More posts by Zach