Don’t Scale a Business Model You Haven’t Proven

At its peak, WeWork was valued at $47 billion. When it filed for bankruptcy in November 2023, it listed more than $18 billion in debt against about $15 billion in assets, according to NPR’s reporting on the petition. Its stock had fallen 99.2% that year.

In between, the company grew fast. It opened more than 700 locations in nearly 40 countries. That growth never fixed the core problem: WeWork signed long-term leases and rented the space out short-term, and the numbers on each desk never worked well enough to carry the expansion.

“WeWork claimed to have reinvented the business without ever reinventing it,” NYU Stern professor Aswath Damodaran said at the time.

WeWork is an extreme case, but the mistake behind it is common. New research in Harvard Business Review argues that it’s the default mistake of ambitious companies.

Growth First vs. Foundation First

In the November–December issue of HBR, Benjamin Hallen of the University of Washington, Ron Tidhar of OpenAI and Kathleen Eisenhardt of Stanford set out two ways to build a company. They call them “growth first” and “foundation first.”

Founders are usually told to choose the first: grab the market, then work out the economics. The authors’ research finds the opposite. Rushing to scale locks in a flawed business model with weak unit economics, and those flaws get harder to fix as the company grows. Startups that chase growth while ignoring unit economics are more likely to fail.

The companies that succeed test their revenue and cost drivers first. They simplify their business models and hold off on growth until the foundation is sound. When they do grow, the authors write, growth becomes replication rather than discovery. The company repeats something it already knows works, rather than paying to find out whether it works.

What the Winners Did Differently

The HBR article draws on a decade-long study that Hallen and his co-authors published in Organization Science, which UW’s Foster School summarized this year. The researchers followed six early-stage online fashion companies in matched pairs. Every firm was well funded, had experienced founders and showed early traction. In each pair, only one became a mature company.

Talent, timing and technology didn’t explain the difference. Focus did.

The successful firms spent their early years taking each transaction apart: revenue, costs, logistics, pricing and customer behavior. They held back growth on purpose, even under pressure from investors. One firm, which the study calls StyleStar, avoided paid marketing, capped sign-ups with waitlists and ran constant experiments. “The first year was just about proving this works,” its CEO told the researchers. It later went public, about six years after launch, at a valuation of more than a billion dollars.

Its counterpart, StyleStuck, launched fast, set aggressive sales targets and spent heavily on customer acquisition. It did all that before confirming it could fulfill orders at a profit. “We didn’t know if we could ever get our cost of goods low enough,” its CEO said. It ended in a discounted acquisition.

One venture capitalist interviewed for the study put it bluntly: “Growth solves many problems, but unit economics is not one of them.”

“The companies that ultimately scaled did not chase growth at the outset,” Hallen says.

The Pattern Is Older Than the Study

The research sharpens a warning that has been around for years. The Startup Genome Report studied more than 3,200 high-growth tech startups and found that 70% had scaled prematurely in at least one area: team, spending, marketing or product. Startups that scaled properly grew about 20 times faster than those that scaled too soon. That data dates from 2011, but the mechanism hasn’t changed.

Founders describe the same pattern when they explain their own failures. In CB Insights’ analysis of 111 startup post-mortems, 38% cited running out of cash or failing to raise more, 19% a flawed business model and 15% pricing or cost problems. Running out of cash is usually the last symptom. Spending ahead of proven economics is often the cause.

A fair caveat: six companies in one industry is a small sample, and the full HBR article is behind a paywall. The findings come from deep case research rather than a large statistical survey. Still, they match what bigger datasets and many founders’ own accounts describe.

What This Means

Hallen and his colleagues suggest a new benchmark to sit alongside product-market fit: “product–market–profit fit.”Customers want what you sell, and you make money every time you sell it. Here’s how to apply it, whether you run a venture-backed startup or a 10-person firm adding a new service line.

Do the math on one unit. Choose your unit: an order, a client, a subscription, a project. Add up everything it costs, including acquisition, delivery, support, and refunds. If one unit loses money, ten thousand will lose ten thousand times as much.
Cap growth on purpose while you learn. A waitlist, a limited launch area, or a pilot with a handful of clients can look like timidity. In practice, it’s a cheap experiment. StyleStar’s waitlist bought it time to fix its model before scale made fixing expensive.
Simplify before you multiply. Each extra product, tier, or channel adds costs that are hard to see at small volume. The firms in the study that succeeded cut their offering down before they expanded it.
Build capability only on a proven base. Systems, hiring, and automation should copy something that already works. One of the struggling firms added services before fixing weak logistics, and the new layer made the old problem worse.
Treat “we’ll fix margins at scale” as a warning sign. Sometimes scale does lower costs. Often it just makes the losses bigger. If the plan depends on economics you haven’t seen yet, test them before you hire for them.