GrowthCraft

How to Make Better Decisions When You Don’t Have Enough Data

One of the hardest parts of being a startup founder is making decisions before you have enough information to feel confident about them.

Should you build the feature? Change the pricing? Hire someone? Focus on a different customer? Spend money on marketing? Keep pursuing the current idea or change direction?

Established companies can often answer these questions with years of customer data, historical performance, market research, and large teams of specialists. Early-stage startups usually cannot.

That creates an uncomfortable reality for first-time founders: you have to make important decisions with incomplete information.

The goal, however, is not to somehow eliminate uncertainty. You cannot. The goal is to develop a process for making reasonable decisions, testing what you believe, learning quickly, and changing course when the evidence tells you to.

That is one of the most important disciplines a founder can develop.

The Founder Decision Traps

When founders do not have enough data, they tend to fall into a few predictable traps.

The first is making a decision based entirely on instinct. Founder intuition has value. You probably understand the problem you are trying to solve better than most people. But intuition is still a hypothesis. It should not automatically be treated as evidence.

The second trap is looking for information that confirms what you already believe. If you think customers will pay $99 per month, it is easy to focus on the person who says, “That sounds reasonable,” while ignoring the five people who say they would never pay it.

The third is asking for too much information before acting. This is where analysis paralysis begins. The founder keeps researching, interviewing, comparing competitors, building spreadsheets, and collecting opinions because making a decision feels risky.

The fourth trap is confusing activity with learning. You can conduct 50 customer interviews and still learn very little if you are asking vague questions or looking for compliments instead of evidence.

Y Combinator makes a similar point in its guidance for founders: early-stage companies need to maintain a direct connection with users and continually use what they learn to improve the product.

The problem is not that you have too little information.

The problem is that you may not have a process for turning limited information into better decisions.

You May Not Need More Data. You May Need Better Questions.

When founders feel uncertain, their first instinct is often to collect more information.

Instead, start by asking a better question:

What would I need to know to make this decision?

Suppose you are deciding whether to build an advanced reporting feature.

You could spend weeks researching competitors, surveying customers, studying market reports, and analyzing potential revenue.

Or you could identify the core assumption:

“We believe our target customers will use this reporting feature frequently enough that it will increase retention or willingness to pay.”

Now you have something you can test.

Talk to existing users. Ask how they currently solve the reporting problem. Look at how frequently they use related functionality. Create a mockup. Put the proposed feature in front of customers. Ask for a commitment, not just an opinion.

The decision becomes much easier because you have converted a vague question into a specific hypothesis.

Strategyzer’s approach to business testing is built around this idea. Before running an experiment, founders should identify the assumptions that need to be true for the business idea to work, then determine which assumptions are most important and least supported by evidence.

Use Assumptions Instead of Pretending You Know

An assumption is not necessarily a bad thing.

Every startup is built on assumptions.

You assume a particular customer has a problem. You assume the problem is important enough to solve. You assume your solution addresses it. You assume customers will pay. You assume you can acquire customers at a reasonable cost. You assume the product can be built and delivered.

The mistake is not having assumptions.

The mistake is forgetting that they are assumptions.

A useful founder habit is to write important beliefs as statements beginning with:

“We believe that…”

For example:

“We believe that small professional services firms will pay $500 per month for automated reporting.”

“We believe that founders will spend 30 minutes per week reviewing a startup performance dashboard.”

“We believe that customers who use this feature twice per week will be more likely to remain customers.”

This simple exercise changes the conversation. You are no longer arguing about whether an idea is good. You are identifying something that can potentially be proven or disproven.

Strategyzer recommends making hypotheses testable, precise, and discrete so that experiments produce useful evidence.

Prioritize the Assumptions That Could Hurt You Most

Not every unknown deserves your attention.

Some assumptions are minor. Others could kill the business.

Imagine you are building a software product for accountants.

You may have 20 unanswered questions about the business. What should the dashboard look like? Which integrations should you build? What colors should the interface use? Should you offer three pricing tiers?

Those questions may matter eventually.

But one question matters more:

Will accountants actually pay for this solution?

If the answer is no, the other decisions are largely irrelevant.

A useful framework is to evaluate each major assumption according to two dimensions:

How important is this assumption to the business?

How much evidence do we currently have?

The assumptions that are both highly important and poorly supported should receive the most attention.

This is essentially the logic behind assumption mapping, which Strategyzer uses to help teams identify high-risk, low-evidence hypotheses before committing significant resources.

For an early-stage founder, this can become a simple weekly exercise. Ask yourself:

“What do we currently believe that, if proven wrong, would materially change what we are doing?”

That is probably where your next experiment belongs.

Avoid Analysis Paralysis

Analysis paralysis often disguises itself as responsible leadership.

You tell yourself that you are “doing research.”

You are “waiting for more information.”

You are “making sure we get it right.”

But startups operate under uncertainty. Waiting for perfect information can be more dangerous than making a reasonable decision with incomplete information.

The better question is:

Can I make this decision reversible?

If the answer is yes, move faster.

Testing a landing page is reversible. Interviewing 10 customers is reversible. Trying a different pricing page is reversible. Running a small advertising experiment is reversible.

Signing a five-year contract, hiring 30 employees, spending hundreds of thousands of dollars, or building a product architecture that is difficult to change is much less reversible.

This distinction can dramatically improve decision-making.

When the cost of being wrong is low, make the decision quickly and learn.

When the cost of being wrong is high, slow down and gather stronger evidence.

Build Fast Experiments Instead of Large Research Projects

One of the best ways to make decisions with limited data is to create your own data.

You do not necessarily need a large research project.

You need a small experiment designed to answer one important question.

For example, if you believe customers will pay $200 per month for a service, you could spend three months building it.

Or you could test the assumption first.

Talk to 10 potential customers. Present the offer. Ask them about their current spending and alternatives. Then ask whether they would be willing to move forward under a defined set of conditions.

You may discover that the price is wrong.

You may discover that the problem is not painful enough.

You may discover that the customer segment is wrong.

Or you may discover that you were right.

All four outcomes are useful.

The important thing is that you learned something before committing significant resources.

Strategyzer recommends using small experiments to test critical hypotheses and emphasizes that the experiment should be connected directly to the assumption being tested.

Think in Learning Loops

A strong startup does not operate like this:

Decide → Build → Hope

It operates more like this:

Assume → Test → Measure → Learn → Decide → Repeat

This is a learning loop.

The decision you make today does not have to be perfect. It needs to create the opportunity to learn something that improves your next decision.

For example:

You believe a particular customer segment is your best market.

You interview customers and discover that the problem exists, but it is not urgent.

You adjust the positioning.

You run another test.

Customers respond more positively, but pricing remains an issue.

You test pricing.

Now you have a better understanding of the market than you had three weeks earlier.

The startup is becoming smarter through repeated cycles.

Y Combinator has similarly described startup execution as a process of forming hypotheses, testing them, drawing conclusions, and repeating the cycle.

This is why early-stage startups should value speed of learning, not simply speed of execution.

Know What Counts as Evidence

Not all information deserves equal weight.

A customer saying, “I love this idea,” is interesting.

A customer giving you a credit card is stronger evidence.

A customer using the product repeatedly is stronger evidence still.

A customer paying, continuing to use it, and referring someone else is powerful evidence.

This does not mean qualitative feedback is unimportant. Early-stage founders often have too little quantitative data to rely exclusively on metrics. Conversations can reveal motivations, objections, frustrations, and problems that analytics cannot explain.

But you should understand the difference between what someone says they will do and what they actually do.

When possible, design your experiments around behavior.

Instead of asking, “Would you use this?”

Ask, “How do you solve this problem today?”

Instead of asking, “Would you pay $100 for this?”

Ask, “What are you currently spending to solve this problem?”

Instead of asking, “Do you like the feature?”

Ask, “How often would this change what you currently do?”

Y Combinator’s guidance on customer conversations similarly emphasizes asking about real experiences and past behavior rather than relying heavily on hypothetical questions.

Create a Decision Framework

When you are facing a difficult decision, write down five things:

1. The decision.
What exactly are you deciding?

2. The assumption.
What must be true for your preferred decision to work?

3. The evidence.
What do you actually know today, and what are you simply assuming?

4. The test.
What is the fastest reasonable experiment that could increase your confidence?

5. The threshold.
What result would cause you to continue, modify the idea, or stop?

That final question is particularly important.

If you do not define what would change your mind before running the experiment, it is easy to reinterpret the results afterward.

For example:

“We will continue pursuing this customer segment if at least five of the next 10 qualified prospects agree to a paid pilot.”

Now the result has meaning.

If you get eight, you have encouraging evidence.

If you get two, you have a reason to reconsider.

If you get five, you have a more complicated decision that requires additional testing.

The important thing is that you decided in advance what the evidence would mean.

Know When to Change Direction

Changing direction is not necessarily failure.

Sometimes the evidence tells you that your original assumption was wrong.

That is valuable.

A founder should become concerned when the same assumption repeatedly fails and the team keeps finding explanations for why the evidence “doesn’t count.”

That is confirmation bias disguised as persistence.

Changing direction becomes more reasonable when you see patterns such as customers consistently describing a different problem than the one you are solving, repeated difficulty getting customers to pay, engagement that disappears after initial use, or a customer segment that responds much more strongly than your original target.

A pivot does not always mean abandoning the entire company.

Sometimes it means changing the customer.

Sometimes it means changing the problem.

Sometimes it means changing the pricing model.

Sometimes it means changing the delivery method.

Sometimes it means removing features instead of adding them.

The goal is not to remain committed to your first idea.

The goal is to remain committed to solving a meaningful problem and building a viable business.

How GrowthCraft Helps Founders Make Better Decisions

This is an area where GrowthCraft can serve as a valuable resource for first-time founders.

Early-stage founders do not always need another generic business course. Often, they need experienced people who can challenge their assumptions, ask better questions, and provide perspective when they are too close to the problem.

GrowthCraft’s community and mentorship model is designed around helping early-stage founders work through practical business challenges rather than simply giving them information.

That distinction matters.

A founder can read about customer validation, experimentation, financial planning, leadership, or business strategy. The harder part is applying those concepts to the specific situation in front of them.

GrowthCraft provides a place for founders to work through those questions with advisors, peers, workshops, and practical conversations. GrowthCraft

The value is not having someone make the decision for you.

It is having people who can help you think through the decision more clearly.

A Simple Weekly Decision Practice for Founders

Set aside 30 minutes each week to review the decisions currently facing your company.

Choose the one that has the greatest potential impact.

Write down what you believe, what you know, what you do not know, and what would change your mind.

Then ask:

What is the smallest experiment I can run this week that will give me better evidence?

Run it.

Record what happened.

Then make the next decision.

Over time, this creates something more valuable than a collection of answers.

It creates a company that learns.

And for an early-stage startup, that may be one of the most important capabilities you can develop.

You will rarely have enough data.

You can, however, build a better process for making decisions with the data you have, identifying what you do not know, testing your assumptions, and learning faster than the uncertainty around you changes.

That is what good startup decision-making looks like.

Frequently Asked Questions

How do startup founders make decisions without enough data?

Start by identifying the assumption behind the decision. Determine how important that assumption is, how much evidence you have, and what small experiment could provide better evidence. The goal is not certainty. It is making a reasonable decision while creating a path toward better information.

What should founders do when they are stuck in analysis paralysis?

Separate reversible decisions from irreversible ones. If a decision is inexpensive and easy to change, make it quickly and learn from the result. For higher-risk decisions, define the specific information you need before acting rather than collecting data indefinitely.

How can a startup test an idea without spending a lot of money?

Start with the smallest experiment capable of testing the most important assumption. That might involve customer interviews, a landing page, a prototype, a manual service, a paid pilot, or a simple pricing test. The best first experiment is often much smaller than the product you ultimately intend to build.

When should a startup change direction?

Consider changing direction when repeated experiments consistently contradict a critical assumption. Look for patterns rather than one-off negative results. A change in customer segment, problem, pricing, product, or business model may be enough. The goal is to respond to evidence rather than becoming attached to the original plan.

What is the most important decision-making habit for a first-time founder?

Learn to distinguish between what you know, what you believe, and what you need to test. That simple distinction prevents assumptions from becoming accepted as facts and creates a more disciplined approach to uncertainty.

References and Further Reading

• Y Combinator, The Scientific Method for Startups: Explains the role of hypotheses, testing, measurement, and repeated learning in startup development. Y Combinator: The Scientific Method for Startups
• Y Combinator, Startup School: Covers startup fundamentals including evaluating ideas, talking to users, building MVPs, and getting first customers. Y Combinator Startup School
• Strategyzer, How to Test Your Idea: Provides a practical framework for identifying critical assumptions and testing them through experiments. Strategyzer: How to Test Your Idea
• Strategyzer, Validate Your Ideas with the Test Card: Explains how to define a hypothesis, design a test, establish measurements, and determine success criteria. Strategyzer: Validate Your Ideas with the Test Card
• Strategyzer, Assumptions Mapping: Provides a framework for prioritizing high-risk assumptions with limited evidence. Strategyzer: Assumptions Mapping