WeWork: The Metric That Rose While the Cash Fell

What happened

WeWork signed long leases on office floors, fitted them out and resold the space by the desk on short terms, growing revenue from $436.1 million in 2016. Its August 2019 public filing showed 527,000 memberships across 528 locations and a 2018 net loss of $1.93 billion. For the quarter to June 2019 it reported a positive contribution margin of $169.8 million, in the same three months the net loss was $638.1 million. Six weeks later it withdrew the registration, and in November 2023 the company filed for Chapter 11.

The case, start to finish

Nothing was concealed. The metric simply pointed at the one floor of the building that was working.

Lease long, sublet short

Take a fifteen-year lease on an office floor. Spend heavily to fit it out. Sell the desks by the month. The gap between a long fixed obligation and a short cancellable one is the entire business, and it is also the entire risk, and the two halves do not appear on the same page of any financial statement.

For most of a decade the model looked like it was working. Revenue grew from $436.1 million in 2016 to $1.82 billion in 2018, and the S-1 filed in August 2019 showed 527,000 memberships across 528 locations in 111 cities. The demand was real, and it is worth being clear about that, because the collapse invites a lazier reading. Average commitment length more than doubled, from roughly eight months at the end of 2017 to more than fifteen months by mid-2019, and committed revenue backlog rose from $0.5 billion to $4.0 billion over the same stretch. This was never a business without customers.

The number that was invented

What WeWork needed was a measure that made the model look like it worked, and it chose one called contribution margin: what a location earns after the costs of running that location, leaving out head office, marketing, the cost of opening new locations, and depreciation. It then published a second version with a non-cash accounting adjustment removed as well.

For the three months to June 30, 2019 that measure was positive $169.8 million. In the same three months the net loss was $638.1 million. The filing also reported adjusted EBITDA, a wider view but still not the full bill, at negative $297.2 million on the same basis, or negative $524.1 million with the adjustment left in. Four figures, one quarter, one company, and every one of them correctly calculated.

Nothing was hidden. The S-1 disclosed the metric, the reconciliation, the losses and the obligations. That is what makes this case useful rather than merely a scandal: the decision that mattered was made by readers, not by the company. What the metric did was direct attention to the one layer of the business that looked healthy, and away from the layer where the money was going. A margin defined by adding back the cost of growing cannot tell you whether growing is affordable.

The two sides of the lease

Set the backlog against the obligation it exists to service. On one side, $4.0 billion of committed revenue from members. On the other, $47.2 billion of committed lease payments, nearly twelve times as much, on US contracts averaging about fifteen years against member commitments averaging about fifteen months. The backlog had grown eight-fold in eighteen months and was still under a tenth of what it had to cover.

Then go to the cash flow statement, which is the one part of the document that cannot be redefined. Operations consumed $198.7 million in the first half of 2019. Investing consumed $2.36 billion more, because fitting out new space ran through investing cash flow rather than the income statement. The gap was filled by raising $3.43 billion from financing. Stripped of names, the business ran on the capital markets staying open.

In 2018, location operating expenses alone were $1.52 billion against $1.82 billion of revenue, before a dollar of head office, marketing, growth spending or depreciation. And only 30% of open locations were mature as of June 1, 2019, so occupancy had to keep rising while the fleet kept growing.

What to read instead

Six weeks after the S-1 appeared the company asked the SEC to withdraw the registration statement, saying it no longer wished to go public at this time. Chapter 11 petitions followed in November 2023.

The transferable habit is small and works on documents far less dramatic than this one. When a company invents a measure, read what the measure leaves out, because that is the direction it is pointing away from. Then go to two places that resist renaming: the cash flow statement, and the obligations. A number can be honestly calculated, fully reconciled, plainly disclosed, and still be the wrong thing to look at. If you run something smaller and find yourself quoting a figure to your own board that excludes a cost you are certain to keep paying, you have built the same instrument, just at a scale where nobody will write about it.

Timeline

• 2010–2018 WeWork signs long leases on office floors, fits them out, and resells the space by the desk on short terms. Revenue grows from $436.1 million in 2016 to $1.82 billion in 2018.
• August 14, 2019 The S-1 is published. It shows 527,000 memberships and 528 locations in 111 cities, and a 2018 net loss of $1.93 billion.
• Quarter ended June 30, 2019 The filing reports a positive contribution margin of $169.8 million for the same three months in which the net loss was $638.1 million.
• September 30, 2019 The company asks the SEC to withdraw the registration statement, saying it no longer wishes to go public at this time. The IPO is dead six weeks after the S-1 appeared.
• November 6, 2023 WeWork Inc. and subsidiaries file Chapter 11 petitions in the District of New Jersey.

You're in the owner's chair

August 2019. You are deciding whether to buy the IPO. The S-1 shows a positive contribution margin, a committed revenue backlog that grew eight-fold in eighteen months, and 70% of locations not yet mature. It also shows a $1.93 billion net loss for 2018. Which of these do you build your answer on?

• The contribution margin — mature locations earn money, and most of the fleet has not matured yet
• The cash flow statement and the lease obligations
• The backlog — $4.0 billion of committed revenue, up from $0.5 billion, is real contracted demand

What WeWork owed, what it was owed, and what it had — June 30, 2019

• Future minimum lease payments owed to landlords: 47.2 $ billions
• Committed revenue backlog from members: 4 $ billions
• Cash and cash equivalents: 2.47 $ billions

US leases ran about fifteen years; member commitments averaged about fifteen months. The backlog grew eight-fold in eighteen months and was still under a tenth of the obligation it had to cover.

Business model

Take a fifteen-year lease on a floor. Spend heavily to fit it out. Sell the desks by the month. The gap between a long fixed obligation and a short cancellable one is the whole business. It is also the whole risk, and the two do not appear on the same page of any financial statement.

Revenue model

Monthly membership fees, about 83% of revenue in the first half of 2019, plus conference rooms, printing and other services. The company reported $1.54 billion of revenue in those six months, up from $763.8 million a year earlier.

Cost structure

Rent, fit-out and the cost of filling empty desks. In 2018, location operating expenses alone were $1.52 billion against $1.82 billion of revenue, before a dollar of head office, marketing, growth spending or depreciation. Fitting out new space ran through investing cash flow, not the income statement: $2.48 billion out the door in 2018 and $2.36 billion in the first half of 2019 alone.

Strategic challenge

WeWork needed a number that made the business look like it worked. It chose one called contribution margin, which counts what a location earns after the costs of running that location, and leaves out head office, marketing, the cost of opening new locations, and depreciation. It then published a second version of that number with a non-cash accounting adjustment removed as well. In the three months to June 30, 2019 that measure was positive $169.8 million. In the same three months the net loss was $638.1 million. The filing also reports adjusted EBITDA, meaning earnings before interest, tax, depreciation and a named list of other items. That is a wider view than contribution margin, but it is still not the full bill. That was negative $297.2 million on the same three months, or negative $524.1 million with the non-cash adjustment left in.

Key decision

The decision that mattered was made by readers, not by the company: which number to believe. The S-1 disclosed all of it: the metric, the reconciliation, the losses and the obligations. Nothing was hidden. What the metric did was direct attention to the one layer of the business that looked healthy, and away from the layer where the money was actually going.

What worked

The demand was real. Memberships reached 527,000 and average commitment length more than doubled, from roughly eight months at the end of 2017 to more than fifteen months by mid-2019. Committed revenue backlog rose from $0.5 billion to $4.0 billion over the same period. This was never a business without customers.

What failed

The arithmetic between the two sides of the lease. Against that $4.0 billion of committed revenue sat $47.2 billion of committed lease payments, nearly twelve times as much, on contracts averaging about fifteen years in the US against member commitments averaging about fifteen months. Operations consumed $198.7 million of cash in the first half of 2019 while investing consumed $2.36 billion more, and the gap was filled by raising $3.43 billion from financing. The IPO was withdrawn six weeks after the S-1; Chapter 11 followed in 2023.

Risk factors

A fixed obligation many times longer than the revenue that services it. Occupancy that has to keep rising while the fleet keeps growing: only 30% of open locations were mature as of June 1, 2019. Total dependence on outside capital to fund fit-out. And a downturn hitting members and landlords in opposite directions at once: the desks empty, the rent does not.

Lesson summary

When a company invents a measure, read what the measure leaves out: that is where it is pointing away from. Then go to the cash flow statement, which cannot be adjusted, and to the obligations, which cannot be renamed. A number can be honestly calculated, fully reconciled, plainly disclosed, and still be the wrong thing to look at.

Key data

• +$169.8M Contribution margin, quarter to 30 Jun 2019 (excl. non-cash lease adjustment)
• −$297.2M Adjusted EBITDA, same quarter, same exclusion
• −$524.1M Adjusted EBITDA, same quarter, adjustment included
• −$638.1M Net loss, same quarter
• $198.7M Cash used by operations, six months to 30 Jun 2019
• $2.36B Cash used by investing, same six months
• $1.82B Revenue, 2018
• $1.93B Net loss, 2018

Sources & basis

The company here is real and named, and nothing about it was invented to make the story land. The list below is where each fact came from — public filings, court records, published reporting — so you can open a source and check it against the sentence that used it.

1. The We Company, Form S-1 filed August 14, 2019 — consolidated statements of operations and cash flows, contribution margin and adjusted EBITDA reconciliations, $47.2 billion of future minimum lease payments, $4.0 billion committed revenue backlog, membership and location counts View source ↗
2. The We Company, Rule 477 withdrawal request for Registration Statement No. 333-233259, September 30, 2019 View source ↗
3. WeWork Inc., Form 8-K filed November 7, 2023 — Chapter 11 petitions filed November 6, 2023 in the US Bankruptcy Court for the District of New Jersey View source ↗