Business

Why founders change strategy before they have enough evidence

A founder can call it agility when what is really happening is discomfort with uncertainty. The problem is not changing strategy. It is changing the experiment before the business has had a chance to answer.

Soma Pattern10 min read

There is a point in almost every early business where nothing is obviously wrong and nothing is obviously working.

The product exists.

A few people like it.

Somebody paid.

Traffic is inconsistent.

One post performs.

The next five disappear.

A sales conversation feels promising and then goes quiet.

You change the landing page.

You adjust the price.

You narrow the audience.

You rewrite the positioning.

You test another channel.

Three weeks later, you are looking at a different business.

Founders usually call this iteration.

Sometimes it is.

Sometimes it is an experiment that never stayed still long enough to produce an answer.

That distinction matters because startups do need to change direction.

The question is not whether to pivot.

The question is whether you are changing strategy because the evidence changed, or because uncertainty became uncomfortable.

Startup advice creates a strange contradiction

Founders are taught two things at once.

Move fast.

Do not give up too early.

Listen to the market.

Stay committed to the vision.

Pivot when the evidence tells you to.

Ignore noise.

This sounds sophisticated until you are the one trying to decide whether three quiet weeks mean the positioning is wrong, the distribution is weak, the offer needs time, the audience is wrong, or absolutely nothing meaningful has happened yet.

There is no clean dashboard for this.

Especially in an early startup, the evidence is often sparse.

Entrepreneurship research has long treated uncertainty as a defining feature of venture creation. You are not simply choosing between known options. You are often trying to learn whether an option is viable in the first place.

That means changing strategy is not only a decision.

It also changes the experiment.

Every strategy change resets part of the evidence

Suppose a founder launches a product.

Week one brings little traffic.

They conclude the positioning is unclear.

They rewrite the homepage.

Week two still feels quiet.

They lower the price.

Week three produces a few signups but no purchases.

They change the target audience.

Week four improves slightly.

Now they do not know why.

  • Was the original positioning wrong?
  • Was the new positioning better?
  • Was price the issue?
  • Did a different audience respond?
  • Did traffic simply increase?
  • Was four weeks far too little time to infer anything?

The founder has been active the entire time.

The business has been changing.

But the learning system is weak because too many variables moved.

This is one reason experimental thinking matters so much in entrepreneurship.

In a randomized controlled study of 116 Italian startups, researchers trained one group of founders to formulate explicit predictions and rigorously test their assumptions. Those founders performed better and were more precise about which ideas to continue, change, or abandon.

A much larger replication published in 2024 examined 759 firms across four randomized trials.

The interesting result was not that scientific founders never pivoted.

They were more likely to terminate weak ideas.

They were also more likely to pivot.

But they tended to make a small number of focused pivots rather than either never changing course or repeatedly changing strategy.

That is a much better definition of agility.

Not movement.

Information.

A pivot is useful when it teaches you something

Startup culture sometimes treats speed as inherently intelligent.

But speed can mean two completely different things.

One founder moves quickly through a sequence of clear experiments.

Another founder moves quickly through a sequence of emotional reactions.

From the outside, both businesses are changing every week.

Internally, they are doing opposite things.

The first founder is narrowing uncertainty.

The second may be generating more of it.

A useful experiment starts with some version of:

I think this is happening because of this.

If that is true, changing this should produce this kind of result.

I will keep these other things reasonably stable.

Then I will look at what happened.

Without that structure, the founder can easily turn every result into a story.

Traffic is low.

The product is wrong.

Traffic improves.

The new headline worked.

A prospect disappears.

Pricing is too high.

Someone buys immediately.

Pricing is fine.

The problem is that every interpretation is being made after the event.

There is no clear prediction to compare with reality.

Founders are not neutral observers of their own companies

This is where business advice becomes much less comfortable.

Founders often talk about data as though data arrives in a psychological vacuum.

It does not.

The same signal can mean very different things depending on the person looking at it.

Ten days without a sale might mean:

  • We need more traffic.
  • We have not spoken to enough customers.
  • The offer is not resonating.
  • The test is still too small.

Or:

This is failing.

That last interpretation may be true.

But it can also come from somewhere inside the founder.

This matters because founders do not merely collect evidence.

  • They decide what counts as evidence.
  • They decide when there is enough evidence.
  • They decide which result deserves attention.
  • They decide when to keep going.
  • They decide when to stop.

For a solo founder, those decisions are often concentrated in one person.

There is no second founder automatically saying:

You changed this three days ago. We do not know yet.

Or:

You have been explaining away the same customer feedback for six months.

That concentration is one reason solo founding is such an interesting context.

Research published in 2026 using Y Combinator and Crunchbase data found that solo founders can face disadvantages relative to founding teams, although broader or deeper founder experience can reduce some of that gap.

The paper points to the resources, knowledge, skills and connections that cofounders can contribute.

This does not mean a solo founder needs a cofounder.

It does mean independence removes some natural friction.

You can change anything whenever you want.

That freedom is powerful.

It also means nobody has to agree before the experiment is reset.

The opposite problem exists too

Some founders change too quickly.

Others stay far too long.

That distinction is important.

Research on early business model experimentation found that founders can resist changing core parts of their business even after receiving negative feedback.

In particular, they showed more resistance to changing the value proposition than some other elements of the business model.

Entrepreneurial experience, startup mentoring, and larger team size appeared to help founders become more receptive to changing course.

The researchers proposed that these factors broaden the founder’s perspective.

So founders can distort uncertainty in both directions.

One founder interprets weak evidence as a reason to abandon the strategy.

Another interprets stronger negative evidence as a reason to defend it.

The interesting question is not:

Am I persistent or am I adaptable?

It is:

What kind of evidence changes my mind?

And perhaps even more revealing:

Does the amount of evidence I require change depending on whether I already want to stay or leave?

Sometimes the strategy is changing because the feeling changed

Imagine two weeks with almost identical business data.

During the first week, the founder feels optimistic.

A customer sends a positive email.

A respected person likes the idea.

A post performs well.

The founder sees the same modest metrics and thinks:

Early signs. Keep going.

Two weeks later, nobody replies to three outreach messages.

A competitor announces something impressive.

Revenue has not moved.

The founder looks at roughly the same level of evidence and thinks:

We need to reposition.

Nothing fundamental happened to the business.

Something changed in the founder’s experience of the business.

That is not irrelevant.

Founders should pay attention to intuition.

But intuition and evidence should not quietly become the same category.

This is where a recurring internal pattern can become a business pattern.

A founder might repeatedly make the prediction:

If this were really working, I would know by now.

That prediction creates urgency.

Urgency creates intervention.

The intervention changes the strategy.

The strategy change interrupts the accumulation of evidence.

The lack of evidence returns.

Then the original prediction feels confirmed.

If this were working, I would know by now.

The founder may experience the loop as business analysis.

From a distance, it looks more like a pattern.

This is the part Soma is interested in

Soma Pattern does not try to tell a founder which business strategy is correct.

It cannot know that from introspection alone.

What it can do is make a different layer visible.

Across repeated conversations, a founder might start to see:

  • uncertainty appears
  • then a particular prediction appears
  • then urgency rises
  • then a business decision follows
  • then the evidence window resets

If that sequence appears once, it may mean nothing.

If it appears across pricing, positioning, distribution, hiring, product decisions, or client work, it becomes more interesting.

That is the premise behind Business Pattern Intelligence.

It connects recurring internal patterns with business decisions, actions, outcomes, and unresolved evidence over time.

Not so that every business problem becomes psychological.

Quite the opposite.

The point is to separate them more clearly.

  • Is the offer actually weak?
  • Did the test fail?
  • Is the acquisition channel wrong?
  • Or did uncertainty trigger a familiar need to change something before the business had produced enough information?

Unknown should be allowed to remain unknown.

That is harder than it sounds.

One of the most useful founder skills may be leaving a question open

Founders are rewarded for answers.

  • What is the strategy?
  • Who is the customer?
  • What is your positioning?
  • Why is growth slow?
  • What will you do next?

But early companies often contain questions that cannot be answered yet.

The temptation is to replace uncertainty with interpretation.

Nobody converted.

Therefore the price is wrong.

People clicked but did not buy.

Therefore the product is weak.

A competitor launched.

Therefore the market is moving away from us.

Our first ten users love it.

Therefore we found product market fit.

Each statement turns incomplete evidence into certainty.

A more disciplined founder might say:

We have a signal. We do not yet know what it means.

That sentence sounds less impressive.

It is often more intelligent.

Before changing strategy, ask what the current experiment was meant to teach you

There is a simple question that can expose a surprising amount of chaos.

What are we currently trying to learn?

Not:

What are we trying to achieve?

Of course you want revenue.

Of course you want customers.

The question is narrower.

What assumption is the current strategy testing?

Maybe:

  • People understand the value once they reach the product page.
  • This audience has the problem strongly enough to pay.
  • The price is acceptable if the value proposition is understood.
  • The product is good enough and distribution is the current bottleneck.

Then ask:

  • What result would weaken that belief?
  • What result would strengthen it?
  • How much evidence would I need?
  • What am I keeping stable while I test it?

Without those answers, a pivot can become little more than movement away from discomfort.

The founder and the business need separate evidence

This is perhaps the most useful distinction.

There is evidence about the business.

  • Customers bought.
  • Customers did not buy.
  • Retention improved.
  • Activation dropped.
  • The same objection appeared in seven sales calls.
  • A channel produced no qualified traffic after a meaningful test.

Then there is evidence about the founder.

  • Silence makes me want to change direction.
  • Visibility makes me reduce my ambition.
  • A positive week makes me expand too quickly.
  • Negative feedback makes me immediately question the whole offer.
  • I keep changing several things at once when I feel behind.

Both sets of evidence matter.

They should not be confused.

A founder does not become more rational by pretending to have no psychology.

They become more precise by noticing when psychology is participating in the interpretation.

That is also why Soma’s broader methodology treats patterns as hypotheses that should strengthen, weaken, or change when new evidence appears.

A founder pattern should not become another fixed story.

It should remain falsifiable too.

The goal is not to become slower

None of this is an argument for waiting indefinitely.

Startups die by waiting as well as by thrashing.

The research on scientific entrepreneurial decision making is interesting precisely because the more disciplined founders were not simply more conservative.

They terminated weak ideas earlier.

They changed direction when the evidence supported it.

What changed was the quality of the decision process.

So the alternative to premature strategy switching is not stubbornness.

It is better experiments.

Better thresholds.

Clearer predictions.

And enough continuity to know whether the evidence actually changed.

Before the next pivot

If you are considering changing your pricing, positioning, audience, product, or acquisition strategy, try answering five questions first.

  1. What exactly is not working? Describe the observation before explaining it.
  2. What did I originally predict would happen? If there was no prediction, admit that too.
  3. What evidence do I have? Separate observed facts from interpretations.
  4. What is still unknown? Do not force an answer simply because the uncertainty is uncomfortable.
  5. If I change this now, what will I no longer be able to learn?

That final question is usually missing.

Every pivot opens a new experiment.

It can also close the old one before it produced an answer.

Sometimes the founder needs to stop changing the business long enough to observe themselves

A startup is already an uncertain system.

A solo startup can become even more difficult to read because the founder is often product leader, marketer, salesperson, analyst, strategist, and emotional weather system at the same time.

The answer is not to distrust yourself.

It is to create enough separation that you can see what is happening.

  • What did the market do?
  • What did I predict?
  • What did I feel?
  • What did I change?
  • What happened next?

Then let the pattern accumulate.

Over time, the more interesting question may stop being:

Why is my strategy not working?

And become:

What keeps happening just before I decide the strategy is not working?

That question does not replace business analysis.

It makes the analysis cleaner.

Research referenced

Related insights

Browse all insights