5 Business Failures Caused by Misreading Market Signals

There is a common piece of startup advice: validate your idea before you build it.

It sounds sensible. The problem is that validation can be surprisingly easy. If you look hard enough, you can usually find evidence that supports an idea. A growing market. People complaining about existing products. A survey where 70% of respondents say they'd be interested. A Reddit thread asking why somebody hasn't built exactly what you're proposing.

None of that necessarily means you have a viable business.

Some of the most famous business failures weren't based on completely ridiculous ideas. In fact, many were built around market signals that were absolutely real.

The mistake was interpreting those signals incorrectly.

Here are five examples.

1. Pets.com: a huge market doesn't necessarily mean a good opportunity

Pets.com has become shorthand for the excesses of the dot-com bubble.

On paper, though, the idea wasn't particularly stupid. People spend enormous amounts of money on their pets. Pet food is a repeat purchase. Bags of dog food and cat litter are bulky and inconvenient to carry home. The internet offered consumers a more convenient way to shop.

Those are all genuine market signals. If you were trying to validate the idea, you could have built a very convincing case for it.

The problem was elsewhere.

Pet food is heavy and relatively low margin. Shipping it directly to individual customers is expensive. Consumers already had a perfectly acceptable way of buying it from supermarkets and pet shops, often alongside other purchases they were making anyway.

Pets.com therefore faced a much harder question than whether people wanted pet supplies delivered to their homes:

Could it acquire and serve those customers profitably?

The answer turned out to be much less attractive than the size of the pet market suggested.

That's an important distinction. Market demand and market opportunity aren't the same thing.

2. Webvan: being right about the future isn't enough

Webvan is an even more interesting example because its fundamental prediction was correct. People did want to buy groceries online. Today, online grocery shopping is completely normal.

Webvan saw that future early. It raised hundreds of millions of dollars and invested heavily in sophisticated automated warehouses and delivery infrastructure.

Then it collapsed.

The problem wasn't necessarily the underlying consumer trend. It was the assumptions surrounding how quickly that behaviour would develop and how economically it could be served. Webvan built expensive infrastructure and expanded rapidly before proving that sufficient numbers of customers would order frequently enough, at sufficient margins, to support it.

In other words, it identified a genuine market signal and then extrapolated too much from it.

This is something founders still do.

Finding evidence that consumers are moving towards a particular behaviour doesn't automatically tell you when that behaviour will become mainstream, how much they'll pay, or whether your particular way of serving them will be profitable.

Sometimes the market really is coming. It just isn't coming quickly enough for your business.

3. Quibi: people doing something doesn't mean they'll pay you for it

Quibi launched in 2020 with around $1.75 billion in funding and some of the biggest names in Hollywood behind it.

Its central insight looked extremely compelling: people were increasingly watching video on their phones, and short-form content was exploding in popularity.

Both observations were correct.

The leap came afterwards. Quibi assumed there was a substantial market for professionally produced, short-form video delivered through a standalone paid subscription.

That proposition had much weaker evidence behind it. Consumers already had YouTube, Instagram, TikTok and other platforms providing effectively unlimited amounts of mobile video. Quibi wasn't simply competing for people's interest in short-form content. It was asking them to change their existing behaviour and pay for another subscription.

Those are very different things.

This distinction appears constantly in startup ideas: "People do X" therefore "people will pay for my version of X."

The first statement can have overwhelming evidence behind it while the second has almost none.

Quibi lasted around six months.

4. Juicero: a real trend doesn't mean there's a real problem

Juicero is one of the stranger Silicon Valley stories.

The company raised more than $100 million to develop a sophisticated connected juice press that squeezed proprietary packs of fruit and vegetables.

Again, there were plenty of positive market signals available. Consumers were increasingly health conscious. Cold-pressed juice was popular. Premium kitchen appliances had a market. Subscription-based consumables could produce recurring revenue. Consumers were embracing connected devices.

Put those together and you can create a very impressive market opportunity slide.

Unfortunately, there was another signal that mattered rather more: people could squeeze the Juicero packs by hand.

A Bloomberg demonstration showing that the expensive machine wasn't actually necessary became symbolic of the company's underlying problem. Juicero had built an extraordinarily sophisticated solution without establishing that the problem it solved was sufficiently important.

That's another distinction that gets lost in idea validation. Consumers can like your proposition. They can understand its benefits. They can even tell you they would use it. None of those necessarily mean the problem matters enough for them to change their behaviour or hand over their money.

5. Segway: "that's amazing" isn't the same as "I'll buy one"

Few products demonstrate the difference between interest and purchase intent better than the Segway.

Before its launch, the hype was extraordinary. Venture capitalist John Doerr reportedly predicted it would become bigger than the internet. People genuinely were fascinated by it. That fascination was real.

But there is an enormous difference between:

"That's incredible."

and:

"I will spend several thousand pounds on that and use it to travel around every day."

The Segway encountered practical problems around price, regulation, where it could actually be ridden and whether it solved a sufficiently important transportation problem.

Production eventually ended in 2020 without anything approaching the transformation of urban transport that had originally been predicted.

If you'd surveyed people after demonstrating a Segway to them, you might have received extremely positive feedback. That wouldn't necessarily have told you very much; interest isn't intent, and intent isn't behaviour.

---

There is a pattern running through all five examples.

Pets.com had evidence of demand.

Webvan identified a genuine behavioural shift.

Quibi correctly spotted explosive growth in mobile video.

Juicero sat at the intersection of several genuine consumer trends.

Segway generated enormous consumer interest.

The signals weren't necessarily wrong.

The interpretation was.

And that's why I think the conventional idea of "startup validation" can be misleading. Ask somebody whether they think your idea sounds interesting and they'll often say yes. Search for statistics supporting your target market and you'll probably find them. Look for people complaining about existing products and you'll almost certainly find those too.

You can accumulate a surprisingly impressive collection of evidence without ever seriously challenging the assumptions your business depends upon.

The better question isn't: "Can I validate this idea?";

It's: "What does the available evidence actually support?"

Those sound similar. They aren't.

This distinction is one of the reasons I built SignalCraft.

SignalCraft isn't intended to tell founders whether their idea is "good" or "bad", and it deliberately doesn't try to produce the kind of reassuring validation score that has become common among AI startup tools.

Instead, it looks for market evidence: customer behaviour, complaints, workarounds, existing alternatives, competitors, willingness-to-pay signals and contradictory evidence.

The objective isn't to prove that an idea will succeed. No research tool can do that.

It's to identify what the available evidence supports, what it doesn't support, and where important assumptions remain unproven. For something like Pets.com, for example, "people spend billions on their pets" isn't particularly useful intelligence. We already know people buy pet food.

The interesting questions are further down:

Why would someone switch from their existing purchasing behaviour?

How important is home delivery to them?

What alternatives already solve that problem?

Are customers willing to pay enough to cover the additional cost of fulfilment?

Does the proposed revenue model actually work with the economics of the product?

What evidence contradicts the opportunity?

And, perhaps most importantly:

What would have to be true for this business to work that we don't currently have evidence for?

Those questions won't guarantee that a startup succeeds. They might, however, stop someone spending six months building the wrong thing.

There is already plenty of encouragement available to people building businesses. Friends will tell you your idea sounds interesting. AI will happily explain why your target market is worth billions. Search engines will find statistics showing that your industry is growing.

None of those things are useless; none of them are due diligence either.

Sometimes the most valuable piece of market intelligence isn't another signal supporting your idea. It's the awkward piece of evidence that doesn't fit.

Pets.com didn't need someone to tell it that Americans loved their pets.

Quibi didn't need another statistic showing that mobile video consumption was increasing.

Webvan didn't need proof that buying groceries online was more convenient.

They needed to understand whether the behaviour, economics and competitive environment actually supported the businesses they were building. That's the difference between looking for validation and looking for intelligence.

And it's a much more useful question to ask before you build.

The lesson for validation

These failures show why finding a real market signal is not enough; interpretation matters. The business idea validation guide sets out a framework for testing competing explanations, while the companion piece on business ideas that succeeded by reading market signals differently shows the opposite side of the problem.