Last year, I called the co-working space provider, WeWork: The Uber of Dumb Investments. US$10bn struck me as a completely unsustainable value for the unlisted company. I stand by my claim, but the company recently raised US$430m from Chinese investors at a price that implied the company’s valuation is now more like US$15bn!
The Financial Times, while maintaining a veil of impartiality, highlighted the absurdity of that valuation by comparing it to one of Manhattan’s landmark buildings. The GM Building’s current valuation is about US$2,400 per square foot, for outright ownership. WeWork, which owns no property, is valued at US$3,000 for every square foot that it leases.
But I think there’s an even starker comparison available. UK competitor Regus plc has an enterprise value of a little under £3.0bn, or US$4.2bn. It has 2.3m members worldwide. That’s an enterprise value per customer of US$1,825. That strikes me as about right given typical customer rents and retention.
It’s far from a perfect comparison, WeWork’s management would probably be insulted that we called Regus a competitor. But WeWork’s US$15bn valuation and 50,000 customers give it an enterprise value of US$300,000 per customer. The comparison isn’t perfect, but is relevant.
There is a slim chance that WeWork might grow (and grow and grow and grow) into its valuation. But I’ll point out a few important factors:
- Regus is also growing pretty damn fast at the moment
- New competition could and should emerge (you could replicate WeWork’s existing position today for but a small fraction of US$15bn)
- Network effects are nowhere near as strong for WeWork as for, say, Uber
- Growth ain’t free.
The smartest thing WeWork’s management could do today is raise as much money as it can while the going is good. US$430m is a start, but if I were running the show I’d be getting much greedier. Perhaps it could buy Regus. It would also be wise to stay unlisted while it tries to raise capital, away from the reach of short sellers, non-believers and other rationalists. Raising big licks of capital won’t justify the current valuation, but it will add real intrinsic value for existing shareholders (it won’t stop recent investors from losing money, unless they can repeat the trick at higher and higher prices).
The probability that WeWork’s new Chinese investors won’t eventually regret their recent commitment? I’d put that at significantly less than 5%. This is one of the craziest valuations I’ve ever come across, and that’s saying something.
Want to know more about the style of investing we practice at Forager? Check out our page what is value investing
I work in a co-working space, so I feel I have something to offer here… USD3000 per square foot – WHAT? Co-working spaces are a great idea, and I hope they catch on (I love mine!), but USD3000 per square foot is patently ridiculous. If I had to guess, I would say the average person in my office uses 16 to 25 square feet each, let’s run with 20 just for argument’s sake. At $660 per month, this works out to $33 of revenue per square foot per month, or $396 per year per square foot. Those are AUD too… Now, no doubt prices in Silicon Valley are higher than prices in Melbourne – but not every city can be like Silicon Valley, and even if it were, I still don’t see how you could justify such an outrageous price…
All that being said, I still feel Atlassian is probably even more overvalued. It has traded as high at 700 times earnings and is currently at 573 times. PE ratios are no doubt poor valuation tools (I view them are pricing tools not valuation tools), but I don’t think I know of a single example of a 500+ PE which has ever been justified.
Great article and I agree that an enterprise value of $300k per customer seems very, very expensive.
Patrck, with regards to paying high P/E multiples, you could have purchased the Altria Group for more than 600x P/E on January 1, 1973 and still earned a 10% CAGR (incl. div reinvested) through December 31, 2015. The MSCI World (also div reinvested) generated a 9.2% CAGR over the same period. Of course, this is the exception rather than the rule, but there are phenomenal franchises out there with such strong business characteristics (moats, ROIC, growth runway) that you can pay extremely high multiples today and still generate great compounding returns over time. However, it’s hard to argue WeWork’s has developed such a moat around its business already, so I’d agree with the conclusion that it seems extremely expensive today.
Also be aware of the potential for large distortion from expensed research and development from youngish tech companies. Certainly not to suggest Atlassian is cheap (I’ve not looked). But calendar 2015 EBIT was ~$4m, after something like $175m of R&D was expensed. R&D is almost 50% of revenue. Is that sort of R&D intensity going to be present in 5-10 years? I have no idea about the answer, but it is an important question for shareholders.
Atlassian at 11x historic revenues, they’re not getting my money but it’s not impossible to imagine them growing into that. Plenty of tech companies have. WeWork at $3,000/customer looks like a more certain stretch to me.
Gareth, I most certainly feel Atlasssian’s current R&D will certainly remain, if not accelerate. It’s a necessary expense (or capital expense) for these types of companies. That is partly the problem. Just like a mining company, the have to continually spend to find the next new deposit or in this case improve their current product with ‘new’ bells and whistles. The cheque is never banked. There is always something new to spend money on.
Food for thought.
Oh no doubt Mike. And the absolute number for R&D spending will almost certainly continue growing. But 50% of sales is a very high figure for R&D, even for a tech company. Amazon’s is ~10-12%, Google’s 14-16%, Apple’s less than 5%, Facebook around 25%, Netflix around 10%. These are not perfect apples-with-apples comparisons by any stretch, and those numbers come from a data provider and may have errors. But my guess is that if Atlassian grows the way the founders want it to, it will be spending less than 50% of revenue on R&D a decade from now. There is at least the potential for revenue to grow fast and earnings to grow a good clip faster. Again, I’m not an investor though.