Adam Schwab: Why Atlassian is one of the world’s most overvalued businesses
Atlassian now has a market value of $99 billion. Incredibly, Australia’s fourth most valuable business has become a meme stock.
A month ago, Atlassian was one of the most overvalued shares in the world. Its market capitalisation was $73 billion. It now has a market value of $99 billion. Incredibly, Australia’s fourth most valuable business has become a meme stock.
Atlassian’s share price remarkably rose after it released its June earnings results which to a traditional investor, appeared to be roundly terrible.
The business lost money during 2022, lots of money: US$614.1 million ($862 million) during the year to be precise (this was a slight improvement on the prior year’s loss). Revenue grew by 34% with cloud revenue growing and server revenues falling. Atlassian’s growth is decent but not amazing (a big chunk of the revenue growth came from expansion revenue rather than new customers).
Bear in mind it’s not overly difficult to grow sales if you’re not bothering about generating a profit. Great businesses are able to consistently increase their return on equity — that is, the more money you invest in them, the more money they make (very few businesses have this delightful characteristic).
Atlassian focus its investor presentations on what they call non-IFRS metrics, essentially ignoring classic accounting metrics in favour of its own Dr Frankenstein version of profit, specifically, free cashflow. A free cashflow metric once was a great way to keep companies honest — businesses that report really high profits but terrible operating cash flows were often misreporting their real profitability. For example, businesses might try to claim all the contracted payments under a 20-year contract as revenue (like Worldcom did) or use dodgy acquisition accounting.
Atlassian ironically does the reverse thing, essentially using a dubious measure of cashflow as a proxy for EBITDA. The problem with this is that Atlassian’s cashflow doesn’t include equity payments to employees and Atlassian pays a huge amount of equity-based compensation to employees — around $1 billion a year.
Paying staff with equity is a fantastic way to align and reward team members, but it has its shortcomings (especially if a business’ share price is dropping).
But regardless of the merits of equity-based compensation, to ignore wages in any sort of performance metric it completely misleading.
Investors have long held similar views on excluding depreciation from financial results (depreciation is an annual charge which is created when a business buys an asset which lasts a number of years).
Deprecation itself is a non-cash expense, that is, a business pays upfront for an asset, say a motor vehicle in year one, the depreciation charge is spread over the lifetime of that vehicle, say five years.
Warren Buffett once noted, “not thinking of depreciation as an expense is crazy. I can think of a few businesses where one could ignore depreciation charges, but not many. Even with our gas pipelines, depreciation is real — you have to maintain them and eventually they become worthless (though this may be 100 years).”
If a business is capital intensive (that is, they need to buy lots of things like factories buildings and manufacturing equipment to make money), ignoring the impact of depreciation in the business’ profit is the financial reporting equivalent of stealing someone’s wristwatch. Nevertheless, EBITDA remains a preferred metric for many investors. Ignoring a huge chunk of wage costs in reporting a profit-like measure as Atlassian does is more like running off with your best mate’s wife and moving into their house. To make matters worse for investors, paying wages in the form of equity has a double cost if a business’ share price is increasing rapidly as the cost of the dilution to existing shareholders is multiplied.
If you look at Atlassian’s Statement of Cash Flows, it looks kind of OK — operating cash flows increased in 2022 from $1.185 billion to $1.24 billion (a rise of around 5%). The big problem for Atlassian is that it increased ‘non-cash’ payments to staff from $544 million last year to $996 million. The business also had a weird non-cash impairment of derivatives of $563 million.
As this columnist noted recently, despite being around for two decades, Australia’s fourth most valuable home-grown business not only appears to be incapable of making real money it even gloats about its financial chicanery. In their letter to shareholders, Atlassian founders, Mike Cannon-Brookes and Scott Farquhar boasted:
The inept investors who value Atlassian at $100 billion lap up this financial reporting equivalent of prison gruel. Speak to any technologist and they roundly hate using Atlassian’s old and clunky flagship Jira product (although Jira is very well-priced and relatively sticky). And despite Atlassian’s comical boasts that they “have a line of sight to US$10 billion in annual revenue based on our current products and core markets”, the customer acquisition data doesn’t look great.
This is from Atlassian’s shareholder letter:
In the last half of FY21, the business added 30,000 new customers (from a lower base). In the last six months of FY22, Atlassian only added 16,000 new customers. So customer growth dropped from 17% to only 7% (this was in-line with operating cashflow growth). Not only is Atlassian rapidly expanding its costs base, it is doing so while the number of new customers it acquires drops.
In short, Atlassian looks like a business that is quickly approaching its growth peak with a 20-year old flagship product. Despite several attempts, Atlassian hasn’t been able to come up with another ‘Jira’ (its flagship workflow management product). Steve Jobs went from Apple 2, to Macintosh, to iMac, to iPhone to iPad to iWatch — poor old Atlassian never managed to get past the Apple 2. It’s had more duds than Renny Harlin.
Despite these problems, Atlassian is one of the world’s most overvalued businesses by almost any metric. Even though it loses money, Atlassian trades on a multiple of price to sales of a comical 25 times. Stern did a comparison of price-sales multiples in January 2022, noting that the multiple for the entire market was 2.88 and for software (this was before the bubble popped) was 16 times.
Microsoft, a business that actually makes money and continues to grow is trading on a multiple of around 10 times sales, Salesforce is on 6.6 times, Netflix trades on a multiple of 3.5 times, Dropbox which is a similar enterprise software business is on only 4.2 times. Even Tesla is only on a price-sales multiple of 14 times. All those businesses have far better products than poor old, loss-making Atlassian. If Atlassian was valued at three times sales (a reasonably fair multiple given its aging product, tepid growth and inability to generate real profits) that would value the business at around $10 billion — a 90% decrease on the current price.
My prediction: Atlassian will start to dramatically reign in costs in the next six months, hiring will slow, and then the business will make a large redundancy round (essentially removing less talented engineers and product managers which it hired in the last two years before their equity vests). This will allow the business to start moving towards profitability. It will also slow growth, which will lead to a re-rating of the business.
Benjamin Graham famously observed that in the short term, the market is a voting machine, but in the long term, it’s a weighing machine. For now at least, Atlassian remains the most popular kid in class, just don’t pull out those scales.
