A learning launch is a carefully designed experiment or prototype designed to test the key underlying value-generating assumptions of a potential new-growth initiative. In contrast to a full new-product roll-out, a learning launch is a learning experiment conducted quickly and inexpensively to gather market-driven data to determine whether a growth idea has enough merit to warrant detailed investigation with the commitment of further time, people and resources.

The learning-launch concept is based on the Darden Growth Leader Research Project and research in the area of successful serial entrepreneurs. From these, we have learned that the key to success is an entrepreneurial mindset of action and not analysis; speed and not long study times; engaging potential customers in the creation, design and testing of new ideas as soon as possible; and learning as you go, iterating constantly for improvement.

The purpose of a learning launch is to learn. If you learn valuable information, the learning launch is a success, whether the particular idea is validated or not. This attitude of learning overcomes the stigma associated with success or failure that limits innovation and growth in many businesses.

To be successful, learning launches must quickly determine the validity of key underlying value-generating assumptions. And learning launches should be designed as a low-risk or affordable-loss way of testing.


Where do learning launches fit strategically? Companies should undertake a portfolio of growth initiatives and manage that portfolio across business units based on timelines designed to produce new sources of revenue, which can be scaled across a large geographical area or customer base. The purpose of growth initiatives is to create new S Curves of income to replace existing S Curves as their useful lives peak. In order to do this, companies need to generate growth ideas and test them in a disciplined manner to determine which ideas warrant investment.

Learning launches are a methodology for testing ideas to determine which qualify as business opportunities. Because they should occur quickly and be low cost, learning launches have also been described as “placing small bets fast.” Unfortunately, some executives have interpreted that phrase as advocating the pursuit of small revenue opportunities, which is not the case. Most good growth companies are looking for growth opportunities to be scaled and result in needle-movers. “Small bets fast” is a testing procedure, however, and not a strategy decision.

The Scientific Method

A learning launch is a business-world application of the scientific method; it is a way of thinking that relies on generating and testing hypotheses. It starts with hypothesis-generating by asking a what-if question, which allows you to identify a new possibility to act on. That possibility is your hypothesis — basically an educated guess about something you think is likely to be a good idea. That’s the creative part. Then to test your hypothesis, you specify what scientists call the boundary conditions. In other words, you ask: Under what conditions would that hypothesis, in fact, be a good idea? Or — worded differently — what would need to be true in order for my idea to be a good one? So, you bring to the surface the assumptions that lie under your hypothesis, and then you go find the appropriate data with which to test them. That is the analytic part.

If it turns out that your assumptions are not true, and your hypothesis is disproved, then you should take the new data and go back and use it to restate the hypothesis better. And then you test that new and improved hypothesis. It is an iterative process that you cycle through continuously, learning something new each time that allows you to develop a better hypothesis for the next pass. The process is what we refer to as intelligently opportunistic. In other words, it lets you rigorously examine your current best guess, but not in a way that forces you to accept or reject it. It instead allows new and previously unanticipated improvements to your idea to emerge. This is learning as you go.

This hypothesis-driven approach is based on the idea of learning, rather than knowing. As a result, it is particularly useful under conditions of uncertainty in which we don’t know as much as we’d like to.

Does all of this sound familiar? Of course it does — a new business idea is, in fact, just a hypothesis. And a learning launch is just a process for testing and improving that hypothesis.

That’s the rub. Hypotheses about the future are a lot trickier to deal with than those about the present; in fact, they can never really be proven to be true. As a strategist or a growth leader, the best you can do is to marshal compelling evidence that supports your argument that your view or your “story” about the future is likely to be worth pursuing. It has to be testable with today’s data. That means that you have to tie the future hypothesis back to some phenomena that are already operating today. Then you are going to ask your audience to leap with you to tomorrow. Their ability to take that leap will be a function of how compelling you can make your story about tomorrow on the basis of what we know today. But only the passage of time will prove whether you are right or wrong.

Key Stages

Stage 1: Immerse yourself in the data you’ve got, looking for patterns and possibilities.

Start the process by looking at what you already know, based on observing customers, analyzing the industry value chain, and assessing internal capabilities and resources. You don’t want to do any more real-world experimenting than you need to; you want first to capitalize on what you’ve got under your nose. So begin by taking a strategic look at what you already know that is likely to be important to a new venture.

Stage 2: Generate a new business idea — an educated guess about an opportunity you see.

Based on what you’ve learned in Stage 1, you now generate a hypothesis in the form of a new business idea with a particular focus on specifying the value proposition to a particular customer, the execution strategy, and likely competitor reaction. Call this your business idea.

Stage 3: Articulate the assumptions on which the new business idea’s success is based.

Having identified the new hypothetical business idea, turn to dealing with the uncertainty piece: what you don’t know. Because your new business is an idea about the future and can’t be tested directly, you’ll need to articulate the critical underlying assumptions that it rests on relative to customers, capabilities and competitors. In this stage, it is essential to separate out the unknowns that matter — the critical assumptions on which the new business rests — from the ones that don’t.

Stage 4: Design and conduct a learning launch to test assumptions.

Now design the hypothesis-testing experiment: the series of small moves that will allow you to gather data to test whether or not these assumptions turn out to be true as you launch the product. Critical here is getting a prototype in the market with prospective customers as quickly as possible.

Stage 5: Revise or kill the idea based on your findings.

Based on the outcome of your test, you will either continue to pursue the new initiative, refining and improving it as you go, or kill it if results suggest that, contrary to your initial projections, it is unlikely to succeed.

Stage 6: Scale the business idea when appropriate.

After you have completed enough research through the learning launch to assure yourself of the sound fundamentals of the new idea, you will move to staging a roll-out of the idea.

At this stage, you will have deciphered the recipe for success and are confident about your ability to accelerate your efforts and change your focus from learning to growing volume rapidly. How you do that, however, is the subject of a different note.

The preceding is excerpted from Darden Professors Jeanne M. Liedtka and Edward D. Hess’ technical note Designing Learning Launches (Darden Business Publishing).