Failure analysis

They raised billions and still failed. The common cause was not technology.

Quibi, Juicero, Webvan, Better Place. None of them struggled to raise money, and all of them hired good engineers. What they all lacked was the same thing: a question nobody asked.

14 July 2026 · 9 min read

When we hear that a startup failed, we usually reach for money or talent as the explanation: they ran out of funding, the team fell apart, the technology was not ready. Yet in most of the most expensive failures, none of that is true. The money was there, the team was there, the product worked. The only thing missing was enough people who wanted it.

CB Insights has spent years collecting the post-mortems that founders write when they shut down, and ranking the reasons they give. The percentages shift between versions of that list, but the top entry does not move: no market need. Team problems, running out of cash and competition all come after it.

Four expensive lessons

Quibi — roughly $1.75 billion, six months

Founded by Jeffrey Katzenberg and Meg Whitman, the short-form video platform launched with Hollywood's biggest names and around $1.75 billion in funding behind it. The idea: people will pay to watch ten-minute, high-budget shows on their phones while waiting in line.

It launched in April 2020 and the shutdown was announced in October of the same year — roughly six months. Nobody had tested the central assumption hard enough: when people watch video on a phone, they are already inside a free and unlimited alternative (TikTok, YouTube). Why would they pay a subscription for the same moment? On top of that, the product initially would not even let you cast what you were watching to a television.

The Wall Street Journal — Quibi shutdown

Juicero — roughly $120 million, undone by one video

Backed by some of Silicon Valley's best-known investors, Juicero built a wifi-connected press that squeezed juice from proprietary packs and sold for $699 at launch. The engineering was genuinely impressive: the device scanned each pack's QR code and checked whether it was still fresh.

In April 2017, Bloomberg reporters squeezed the same packs by hand and got nearly the same amount of juice in about the same time. The company shut down a few months after the video went out. The unasked question was simple: does opening this pack really require a $699 machine?

Bloomberg — “Silicon Valley's $400 Juicer May Be Feeling the Squeeze”

Webvan — roughly $800 million, two years

Webvan went public in 1999 as an online grocery delivery service and began building enormous automated warehouses. In hindsight the idea was right — people did want to order groceries online.

It filed for bankruptcy in 2001. The idea was not wrong; the assumption about scale and timing was. The company built infrastructure sized for a demand it had never verified existed yet. The companies that made the same idea work twenty years later built the demand first and the warehouse second.

CNET — Webvan files for bankruptcy

Better Place — roughly $850 million, five years

Rather than have electric cars wait to charge, Better Place proposed swapping their batteries at a station, and built station networks in Israel and Denmark. The vision was compelling and it raised a great deal of money.

It was liquidated in 2013. The model depended on carmakers agreeing on a single standard battery format — an assumption that was never validated, and never came true. By the time the network existed, only one car model carried that battery.

Reuters — Better Place to be liquidated

The pattern all four share

What these four companies have in common is not a bad idea. Quite the opposite: every one of them was convincing on paper — convincing enough to persuade some of the most experienced investors in the world. What they share is that they tested their most critical assumption AFTER building.

Quibi: the assumption that “people will pay a subscription for short video” was tested after $1.75 billion in production commitments.
Juicero: the assumption that “this pack requires a machine” was disproved by a journalist's bare hands, after the device shipped.
Webvan: the assumption that “demand at this volume exists today” was tested after the warehouses were built.
Better Place: the assumption that “carmakers will converge on one standard” collapsed after the station network was built.

So how would this have been prevented?

None of these cases lacked intelligence or effort. What was missing was a handful of boring questions asked before construction began: Who solves this problem today, with which product, and how much do they pay for it? Where did the people who already tried this stumble? Which single assumption is this idea standing on, and is that assumption verified right now?

The answers are usually already out there: in competitors' user reviews, in the post-mortems of companies that closed, in funding announcements, on complaint platforms. The problem is not that the information is missing. It is scattered, and nobody wants to spend weeks assembling it.

That assembly is exactly what Premisey does: you describe your idea, it researches your market, competitors and revenue signals with live web search, ties every finding to its source, and shows separately what is verified, what is an estimate and what is still unknown. The goal is not to talk you out of your idea — it is to make sure you ask the right question before you spend months on the wrong one.

Ask the same questions about your own idea.

Describe your idea and we will research its market, competitors and revenue signals, with sources attached. The first analysis is free.

Research my idea