Description

We believe Atlassian (TEAM) is the most terminal value accretive SaaS transition we have ever observed. We believe now is a particularly opportune time to own the stock, as it enters the steepest portion of the alpha-curve. 

Why the opportunity exists 

• We estimate post-transition FCF of $20-25/share vs. consensus at ~$10/share: few investors have focused on the post-transition earnings power of the business because it will take ~4-5 years to be fully reflected in the financial model. Our post-transition estimates are built on a SKU by SKU and customer type by customer type analysis (see below).

• We believe Atlassian is entering the sweet-part of the alpha-curve: we expect subscription revenue growth rate to bottom in 2HCY23, followed by a steep multi-quarter acceleration in growth rate in CY24 and sustained best-in-class growth through CY25. The current window of opportunity reminds us of  the steepest portion of alpha capture in our experience with ADBE and ADSK subscription transitions (see below).

  • We believe Atlassian’s core end markets have been in recession for ~2-3 quarters, leading to repeated resets on near-term growth expectations. While a cyclical recovery will likely be a tailwind to growth in CY24/CY25, we do not believe we need a material cyclical recovery for subscription revenue to re-accelerate.

• Given the noise in the financials due to the SaaS transition, Atlassian screens as optically expensive on the “Rule of 40”. We believe this leads some investors to avoid the stock as it is perceived to be high-risk, or actively short the stock along a basket of “expensive software”.   

We leave valuation as an exercise to the reader. We believe the stock is currently trading at ~5-6x post-transition FCF per share, and one of the most compelling risk-adjusted returns in the market today. 

This is the most terminal value accretive SaaS transition we have ever observed 

We believe 80% of the legacy installed base will be repriced to ~8x pre-transition ARR or more, whereas the market expects only ~16% of the installed base to experience a material repricing of 3-4x. 

The Company announced a forced transition of its installed base from on-premise (and license-maintenance pricing) to Cloud (and ratable SaaS pricing) at its FY21 analyst day. Existing customers were granted a 3-year grace period to transition and multi-year discounts to incentivize the switch, and management explicitly guided to larger customers taking the full 3-year grace period before switching. 

We believe the multi-year duration and management downplaying the pricing uplift made the market complacent to the most accretive SaaS transition in our history of analyzing publicly traded software companies. 

The market perception: there is an accretive ~3-4x recurring revenue lift for legacy on-prem customers with >500 seats, but this represents only 16% of the installed base, and for the other 84% of the business the accretion is a much more pedestrian ~0.9x-1.3x. 

We believe there are three ways the market is underestimating the embedded FCF power of the installed base: (i) Large customers, where the accretion is concentrated, are only 16% of logos but we estimate ~80% of the installed base seats, (ii) management cherry-picked examples to demonstrate a 3.7x recurring revenue accretion for large customers; we believe the real blended number is ~8x assuming standard SKUs, and (iii) the Company only recently adopted best practices in pricing and packaging by introducing “Premium” and “Enterprise” SKUs, which we believe can lift that ~8x materially higher. We believe the stickiness of the platform and the low starting point of legacy pricing ($0.50-$3.00 per user/month for larger customers), compared to enterprise software peers commonly at $20-$200 per user/month, provides a pricing umbrella for the company to execute on this transition with only modest incremental churn. 

(i) The ARR lift applies to 80% of the installed base, not 16%. Market participants do not model the Company by customer segment. Only during the FY21 analyst day did Management reveal the infographic below, which shows 16% of customers (with >500 seats) would experience a material lift to recurring revenue. However, the number of customers (e.g., logos) is the wrong unit of analysis as pricing is calculated on a per-seat basis. We estimate these 16% of legacy customers represent ~80% of the installed seats.  

Source: Atlassian FY21 Analyst Day 

(ii) Blended recurring revenue uplift for large customers is likely 8x or more, not 3.7x, even if we assume the cheapest Cloud SKU. We believe Management cherry-picked seat count in their FY21 Analyst Day presentation. In the large customer example below, a customer with 2,200 seats is chosen to demonstrate a 3.7x lift to legacy maintenance revenue. This conveniently ignores that legacy pricing is flat within a user tier (e.g., the $25k legacy pricing in the example applies to all customers in the 2,001 – 10,000 user tier). For a customer with say, 4,000 or 8,000 users, the recurring revenue lift on Cloud is 9x-16x. 

We believe management cherry-picked the 3.7x lift example to show a consulting-firm analysis that the move to Cloud is a win-win for customers on TCO due to FTE headcount and infrastructure savings offsetting higher Cloud pricing. When prices shifted by the FY22 Analyst Day the following year (due to additional price increases), the presentation quietly updated the 2,200 seat example to 1,750 to continue to continue to show the new pricing as win-win.

Source: Atlassian FY21 Investor Day. Red square emphasis ours. 

Source: Atlassian FY22 Investor Day. Red square emphasis ours. 

(iii) Atlassian has a significant opportunity to improve product packaging in its Cloud SKUs (bundling features, SLAs, enhanced support) to price-discriminate at the Enterprise level. Only in the last few years has the Company adopted industry best practices by launching and emphasizing “Premium” and “Enterprise” SKUs. If large customers adopt these more expensive SKUs, the pricing uplift more than doubles. Our conversations with Atlassian resellers suggest they are currently observing ~70-80% pro-mix in early conversions to Cloud. Our base case assumes ~40-50% of seats will eventually convert to premium SKUs. 

This misunderstanding of the installed base monetization opportunity is the largest driver of our variant perception that the Company can generate $20-25 of FCF/sh post-transition rather than ~$10. Our numbers initially seemed too good to be true compared to TEAM’s $765mm pre-transition FCF (~$3/sh). ADSK’s SaaS transition lasted ~7 years and FCF moved from ~$500mm to ~$2.4bn (incl. LT Deferred benefit). ADBE’s ~7-year transition moved ~$1bn of FCF to ~$5bn. However, both ADSK and ADBE had orders of magnitude lower embedded price increases in their transitions and lower underlying organic growth drivers; we believe these precedents suggest our math could ultimately prove conservative.

The strongest evidence we can find to refute our hypothesis: Customers may more strictly audit their licensing needs during the Cloud transition. Buyers will more critically examine if infrequent users need a license at all now prices, while still cheap on a relative basis, are no longer low on an absolute basis. 

We incorporate a 10% reduction in total seat count in our transition modeling. Our channel checks and survey work suggest seat reductions at transitioning customers are currently running <5%. Qualitatively, we have heard that Atlassian’s total bill is so far down the CFO’s list of costs, especially relative to Jira’s importance, that it generally escapes scrutiny. For instance, NOW has 1.2x the number of F500 customers as Atlassian but 5.5x the number of $1mm+ ARR customers. 

Source: Company Disclosures

Our transition modeling assumes no material discounting through ELAs, consistent with Atlassian’s sales practices pre-transition. Our channel checks suggest Atlassian’s increasingly traditional enterprise sales motion may selectively engage in ELAs with modest discounts off list at the largest customers, but only if the discussion entails an order of magnitude increase in seat count (e.g., going wall-to-wall) relative to pre-transition. 

Source: Illustrative PF for Price Change

Illustrative of Atlassian’s margin potential at scale, we believe that if the Cloud transition happened at the snap of a finger overnight, TEAM today would be generating 50% adjusted EBIT margins and 60%+ FCF margins. While this is a highly stylized pro forma (we don't underwrite to these types of margins due to ongoing investment in opex), we believe it highlights the exceptional unit economics of the embedded install base.

Organic growth (ex-pricing) has best-in-class unit economics, is more durable than the market anticipates, and benefits from looming vendor consolidation 

Unlike some subscription transition stories where structurally challenged businesses pull pricing as a final lever for value capture, we believe Atlassian has multiple vectors to drive 20% or better organic growth, at high incremental returns, over the next 5+ years. Near term, our work suggests Atlassian is one of the primary beneficiaries of a looming multi-year cycle in software vendor consolidation and cost rationalization.  

Atlassian’s second act: Expanding from the Developer to broad IT usage (~2012-present)

Atlassian initially launched the predecessor of Jira Service Management (JSM) to the broader IT department in 2013 after discovering that nearly 40% of customers adapted Jira to handle IT department service requests. By 2018, this SKU had become the fastest growing product in Atlassian’s history, prompting management to increase its GTM investment. To round out organic R&D, the company acquired OpsGenie (2018, $295mm) and tucked in Mindville (2020, undisclosed), the latter of which was a 3rd party application built on the Atlassian Marketplace, before launching the present form of JSM to directly address the ITSM market opportunity in late 2020. 

Equity market participants are skeptical of Atlassian’s expansion to ITSM because ServiceNow (NOW) is perceived as the category-killer. Since its re-launch in 2020, JSM customer count has grown to 40,000+ from 25,000 in ~18 months, almost exclusively from mid-market wins replacing legacy vendors such as BMC Remedy and Ivanti.  Both NOW and TEAM management consistently say there is enough market for two players to bifurcate the enterprise and mid-market.

While competition between NOW and TEAM does not appear zero-sum today, we believe it is a question of when that will change rather than if. We hypothesize it is far more likely Atlassian succeeds in moving up-market than NOW moving down-market. JSM is a textbook example of a disruptive innovation that today offers 80% of the functionality at 20% of the price, with a clear roadmap to improve its value proposition upmarket over time (e.g., only recently achieved HIPAA compliance and is still working on FedRAMP, two critical enterprise level certifications). We also expect Atlassian to benefit from its fully integrated product suite for software developers and IT users. ITIL, the governing framework for ITSM, recently recommended as an industry best practice that ITSM software tightly integrate with developer toolchains. This reminds us of the change in guidance ITIL made for ITILv3 in 2007 that initially catalyzed an acceleration of growth for NOW, only this time making the move in favor of Atlassian. 

We estimate JSM today represents ~20-25% of revenue and is growing ~10% points faster than Company average. 

Atlassian’s third act: Expanding from IT to the Enterprise (~2020-Present)

We believe Atlassian is a few years into a high-priority investment cycle to expand to business users, reminiscent of management’s initial investments in expanding throughout IT nearly a decade ago. While still early, we believe this opportunity is significant, with the potential to become a large driver of organic growth in 3-5 years. 

Source: Company Disclosures

The market remains hyper-skeptical due to competition against cash-burning collaboration startups such as Asana, Monday.com, SmartSheet and Airtable. The market believes Atlassian, championed by the IT department, is poorly positioned due to the “consumerization” of enterprise IT. Analysts point to Slack winning the battle for chat or the iPhone winning enterprise adoption despite protests from IT. We believe this type of pattern recognition reasons by incorrect analogy. While IT had an opinion on the chat app and phones utilized by business users, it was primarily a political or philosophical battle that did not materially impact the IT department day-to-day. Conversely, Jira is the de facto system of work and how developers and IT departments collaborate daily to get their own work done. If organizations are standardizing on one enterprise-wide work platform, we believe it is unlikely developers and IT allow business users to dictate that Jira be ripped and replaced. 

While Atlassian is admittedly still refining its selling motion outside of IT, there are green shoots that a looming multi-year cycle of vendor consolidation can benefit Atlassian. We scraped thousands of collaboration customers to look at instances of mid-market and enterprise customers adopting new business user collaboration tools. Our analysis showed that 18% of the time point-product vendors (Asana et al.) were the incumbent when Jira was adopted by business users, the incumbent would be ripped out completely, vs. Jira being replaced <2% of the time if it was the incumbent when one of the point product vendors was adopted. We believe this is evidence Atlassian can benefit from an increase in vendor consolidation. 

Source: Our analysis/scraping.

A word on business quality and the durability of competitive moats…

Much ink has been spilled on the network effects of grassroots collaboration software leading to hypergrowth and user adoption flywheels. But viral adoption is a double-edged sword that can sow the seeds of eventual product commoditization. For instance, Zoom and Slack drove viral growth by removing almost all friction to adopt a new video or chat agent. But while network effects allowed them to quickly capture market leadership, they could not sustain a durable window of competitive advantage when Microsoft fast-followed with the same tactics.

Jira’s grass-roots adoption curve with software developers has been a slow-boil over two decades and can’t be as easily growth hacked. Nor are the “big bang” conditions of its creation easily replicated. Jira was born in 2002, a year after the publication of the Agile Manifesto, precisely when incumbent software vendors remained focused on the traditional waterfall software development that dominated for the 20 years prior to 2002. And Jira enjoyed a multi-year window where it faced little competition from new market entrants during the long winter following the collapse of the .com bubble.

Jira is highly customizable, leading some users to complain that the software is clunky. This slowed adoption but is in part what makes it sticky. Plug-and-play software spreads easily but is also easy to rip-and-replace. At its best, Jira becomes a customized database and a system of record of all the work and major projects within an enterprise, broken down to atomic tasks. This is where Jira’s customizability shines. And while network effects drove grassroots adoption, it’s the customizations that make Jira sticky. We are careful not to push this logic too far. While tempting to compare this to canonical systems of record, replacing Jira is painful but not the “open-heart surgery” of replacing an ERP system that houses live production data.  We think the more apt analogy is Salesforce. CRM’s own “big bang” moment came in the mid-2000s. It launched Force.com to open its platform to 3rd party developers, just as its core product gained traction as a system of record.  This defied convention from dominant software franchises at the time such as Oracle and Siebel, who actively sought control of their application stack, to the point of litigating against 3rd party developers. CRM’s timing, by luck or by genius, coincided with the launch of the iPhone and increasing awareness of “app stores” and tapped into a wave of discontent from a cottage industry of 3rd party enterprise software developers.  

Almost every software company claims to investors to have crossed the chasm from point-product to platform, but we believe the list of true enterprise software platforms is highly exclusive. Vibrant 3rd party developer platforms are rare because they create a durable flywheel: end-customers receive a rich option of niche, customized software add-ons that may not be profitable for the platform owner to focus on, all built on the same application stack, data objects and UI; 3rd party developers building on the platform have faster time to market and lower CAC selling to a captive and growing customer base; the platform owner receives a 10-25% royalty but more importantly creates durable customer lock-in that is difficult to disrupt.  

Source: Our estimates/scraping.

While many have tried, we believe Atlassian is one of less than a handful of companies that has achieved platform status in B2B software.  We don’t expect this progress to be linear. Atlassian’s current transition to Cloud is reminiscent of Salesforce’s own pivot to mobile a decade ago, a period that saw a temporary slow-down in developer growth as Salesforce re-platformed. Atlassian’s Marketplace counts 4,000+ apps that have been built by 25,000+ developers over the last decade. We estimate 3rd party developers generate ~$700mm in annual revenue on Atlassian’s marketplace. 70%+ of Atlassian’s customers have paid for at least one app on the marketplace and when a customer buys a marketplace app, their propensity to churn is cut in half. 

While investors focus on Atlassian’s unique go-to-market strategy, we think it can be replicated. Instead, we believe Jira’s high level of customization, enhanced by its platform ecosystem, is what creates material switching costs and a durable competitive advantage window. 

The three most likely ways we’re wrong

(i) Microsoft Competition

Microsoft is often cited as a competition bogeyman and perceived to be especially formidable given its dominant position in the developer toolchain (Visual Studio, Github, Azure, etc.) and the business user (Office365, Sharepoint). While Microsoft hasn’t pursued scorched earth competition to date, our work suggests they consider Atlassian’s end markets an important strategic investment vector.  Our analysis suggests the fear of MSFT is currently unfounded in the business user segment. Our customer scraping analysis separates out customers investing heavily in Microsoft Teams (as a proxy for customers going “all-in” on Microsoft) and their spend intent relative to control group. Companies investing heavily in Microsoft are far more likely to increase their spending on Atlassian. In contrast, they are far less likely to spend on other productivity software (Zoom, Slack). 

Source: Our analysis/scraping.

While still positively correlated to Microsoft growth, we do find some evidence of Microsoft rate-limiting Atlassian’s opportunity to consolidate the developer toolchain. For instance, customer intent to invest aggressively in Atlassian is materially higher (1.5-1.8x ratio) for companies increasing investment in dev tools. However, if customers adopt MSFT’s Github, they are slightly less likely to expand Atlassian usage than if they adopted Gitlab (GTLB). Similarly, customers in our scrape investing heavily in public cloud were positively correlated to more aggressive investment in Atlassian. But adopters of AWS or GCP were more likely to be investing heavily in Atlassian than those investing aggressively in Azure. 

Source: Our analysis/scraping.

(ii) The organic growth opportunity is large but may come at poor incremental returns 

There are multiple SaaS companies that relied on self-service sales and failed to move up-market, in part due to insufficient investment in a traditional enterprise salesforce. If Atlassian does make the necessary investments, it will compete with cash-incinerating collaboration competitors such as ASAN (-82% GAAP EBIT margins), so unit economics may be inferior to TEAM’s historic 35%-40% FCF margin, when almost all its sales were low-cost self-serve or channel led. 

Our base case financial forecast assumes TEAM invests significantly to build out a traditional field sales force during our forecast period. There were three primary ways we triangulated the ramp and cost of this investment: 

1. Our bottom-up field sales build: we assume TEAM adds ~$1bn of enterprise sales costs over our forecast period. Key assumptions within the analysis include $1mm of ARR per quota-bearing field salesperson, a span of control of two FTE resources (pre-sales, sales engineers) per QBS at an all-in average compensation of ~$750-800k per QBS cluster. We believe this rolls up to a direct annual sales-quota of ~$1.3bn of incremental ARR without considering existing field sales capacity and the continued contribution from self-service and channel partners. 

2. Adobe as a historical analog: Following its acquisition of Omniture in 2009, Adobe embarked on a significant investment in a field-sales force to move up-market to a C-Suite level sale for its Digital Marketing business unit.  Over this investment period ADBE added ~$1bn of S&M expense, of which we estimate ~$500mm+ was directed to building out the enterprise salesforce.

3. SaaS benchmarking: LinkedIn data scraping suggests 11% of Atlassian’s FTE base is in a sales role, compared to ~22% for its comp group.  “Right-sizing” Atlassian’s salesperson headcount to match the mix of its Enterprise software peers suggests an additional ~$300mm of pro forma investment required today, or ~$500-600mm over our forecast period adjusting for continued overall headcount growth.  

Source: LinkedIn data scrape

(iii) Management may never give us the margins 

While this is possible, we believe the risk is amongst the lowest of founder-led tech companies we cover. The co-CEOs/founders describe themselves as “ruthless capital allocators.” They are unique in our software coverage in self-identifying in that manner and in their execution of that identity.  

Co-CEOs Scott Farquhar and Mike Cannon-Brookes co-founded Atlassian in their early twenties in 2002. They jointly own ~43% of the equity (88% voting power). Culturally, Atlassian was built different from the recent cohort of software companies. Atlassian’s frugal nature was borne out of necessity rather than deliberate strategy. The Company was bootstrapped on $10,000 of credit card debt in 2002, profitable by year 3, and didn’t raise external capital until 2010. The Company generated 30%+ FCF margins pre-Covid and prior to the forced subscription transition. 

Internal R&D has shown tangible evidence of high returns. Jira Service Management, the fastest growing product in the company’s history, was the result of an internal hackathon and has increased from 0% of revenue to 20%+ in the last decade. 

Management’s track record of M&A is mixed, in part because the Company has shown a willingness to rapidly curtail investment and admit defeat if a growth vector is not demonstrating attractive unit economics or return on investment. 

In 2012, the Company acquired HipChat, an enterprise chat tool. Management spent considerable internal resources and external goodwill with investors in positioning the product as a key growth vector.  Most tech executives would likely have succumbed to sunk cost fallacy and escalation of commitment, even as competitor Slack started to run away with the market. Instead, the founders shut down HipChat in 2018 by selling the division to Slack for a nominal sum. Despite HipChat’s failure to win the market, we estimate the Company realized an attractive return on the HipChat acquisition. More importantly, we believe the management team has consistently demonstrated a healthy respect for capital scarcity and a focus on high incremental ROIC growth rather than growth at any cost (see also:  the “ruthless” exit of Jitsi, an open-source competitor to Zoom in 2018, and the more recent de-emphasis of dev-chain tools such as BitBucket). 

“We concluded that, although this market remains large, the additional investment required to compete effectively [against Slack] is unlikely to generate returns that are comparable to those of our other products and the opportunities around us.” – co-CEO, July 2018 earnings call, commenting on the decision to sell HipChat to Slack.

Catalyst

we expect subscription revenue growth rate to bottom in 2HCY23, followed by a steep multi-quarter acceleration in growth rate in CY24 and sustained best-in-class growth through CY25.