Five Questions Before You Build

The odds are public

The U.S. Bureau of Labor Statistics follows new business establishments over time. Of those born in March 2013, 34.7 percent were still in operation in March 2023. After the first year the survival rate was 79.6 percent, and after five years it was 50.6 percent. The largest single drop came in the first year after opening, at 20.4 percentage points.

These figures describe establishments, not firms. The Bureau measures age and survival at the establishment level, so a chain opening a new branch counts as a new opening. We read the numbers as a broad guide, not a forecast for any one company. They tell us that building something people keep using is hard. They do not tell us what to build. That choice is where the early questions matter most, because changing course costs little before work starts and a great deal afterwards.

Five questions

We think five plain questions, asked before any work begins, cost little and can change what is built.

First, who has the problem, and how do we know? Naming a person and a situation is harder than naming a market, and it is more useful. Second, what do they do today instead? Every problem is already being handled somehow, even if badly. Third, what would make them choose something new over that habit? Fourth, what is the smallest piece of work that would test the answer? Fifth, what result would tell us to stop?

None of these needs a spreadsheet or a budget. All of them need honest answers, and the honest answer is sometimes "we do not know yet." We think that answer is worth writing down, because it marks exactly where the next piece of evidence is needed.

What the numbers do not say

We think a closed business has not always failed in the way the word suggests. Owners retire, move on or change direction for reasons that no establishment count captures. We also do not claim that asking five questions changes anyone's odds. We have no evidence on that, and we would not invent any. The questions are a discipline, not a guarantee, and we hold them to the same standard of evidence as everything else we publish.

Why we ask them

We ask these questions of every idea we consider, including our own. MaxxForge is designed to serve outside founders as well as the portfolio, and we treat the portfolio as a worked example, not the limit.

If you think a sixth question belongs on the list, or that we have misread the data, we would like to hear it.