ASAN.US

The ceiling is large in absolute terms but mostly a "grow an existing pie" story, not a brand-new market — and Asana captures only a fraction of it. Collaborative work management is a broad horizontal category tied to how all knowledge work gets coordinated, not a niche vertical. Asana's own S-1 cited IDC data putting the combined collaborative-applications and project-and-portfolio-management markets at about $23 billion in 2020, growing toward $32 billion by 2023, against roughly 1.25 billion global information workers. Those figures are dated and vendor-mediated, but the shape holds: the addressable pool is tens of billions of dollars and structurally tied to a secular shift toward digital, cross-functional coordination.

The honest read, though, is that this is a pie that already exists and is already contested, not a market Asana is creating. The need for work coordination predates Asana; the company's contribution was a cleaner, more structured way to do it, replacing the "work about work" that used to live in email, meetings, and spreadsheets. Even today the report notes Asana's filings warn that customers can keep relying on email, spreadsheets, messaging tools, and rival suites — i.e., the substitutes are free or already-owned, which is the opposite of a greenfield category with no incumbent alternative.

The one place Asana is arguably creating something new is the framing of an "operating system for human-agent teams" — the Work Graph as structured context for AI agents to act inside real workflows, monetized through consumption-based AI Studio rather than seats. That is a genuine attempt to expand the pie's definition. But it is aspirational: AI Studio exited FY2026 at just over $6 million ARR against a $790.8 million revenue base, so the "new market" portion is real in direction and tiny in size. From an LTGG lens, the TAM ceiling is high enough to matter but does not, by itself, support a 5x — because Asana is one of several players splitting an established budget line, and its share of that ceiling is what's in question, not the ceiling's height.

Doubling revenue in five years is a stretch, not a base case — it would require a clear AI-led reacceleration that has not yet shown up in the numbers. A double from the FY2026 base of $790.8 million means about $1.58 billion by roughly FY2031, which implies a sustained ~15% CAGR. Asana is currently growing at 9% (FY2026) and 9.5% in Q1 FY2027, and management's own FY2027 guidance of $855.5 million to $863.5 million implies only about 8%–9% growth. To get from high-single-digits to mid-teens for five straight years, growth would have to inflect upward rather than continue drifting down — and the report's whole thesis is that the seat-based engine is maturing, not accelerating.

On the volume-vs-price-vs-new-business question, the report is clear that the old volume lever is fading. Seat expansion is the historical driver, but dollar-based net retention has fallen to 96% overall (97% core), meaning the average existing cohort is barely growing net of churn; management explicitly flags PLG as a roughly two-point drag on ARR growth as AI changes how buyers discover software. Price/tier mix still helps at the high end — $100,000-plus customers grew 12% to 817 — but that cohort is small. The credible incremental engine is new business in the form of consumption-based AI: AI products were 17% of net-new ARR in Q1 FY2027, above the ~15% target pace, and AI Studio customers show better seat expansion and retention. RPO up 23% year over year (to $518.1 million), well ahead of reported revenue, is the most encouraging forward signal that bookings could feed faster revenue later.

Net: a double is possible only under the report's optimistic scenario, where AI and StackAI lift expansion enough to push NRR toward 98% and AI sustainably exceeds 15% of net-new ARR. The base case the report models is far more modest — roughly 8%–11% growth with margin gains doing more work than the top line. So under Baillie's "at least double in five years" test, Asana screens as a probable miss unless the AI monetization curve bends sharply upward; the more likely outcome is 1.4x–1.6x revenue over five years, not 2x.

The second curve is consumption-based AI monetization — the Work Graph as the runtime for "human-agent teams" — and it exists today only in embryonic form. The first curve is clear and now maturing: replace ad hoc coordination with structured, seat-priced work management, which built a $790.8 million FY2026 revenue base at 89% gross margin. The report frames the second curve as the harder, more ambitious claim: becoming the governed workflow layer on which humans and AI agents actually run the business, with AI Studio introducing usage-based pricing into what was an almost entirely per-seat model.

Crucially, the second curve does already exist as a shipping, revenue-generating product rather than a slide — which is more than many SaaS "AI stories" can say. AI Studio exited FY2026 above $6 million ARR and grew over 50% quarter-on-quarter in Q4 FY2026; AI products reached 17% of net-new ARR in Q1 FY2027; and AI Studio adopters show stronger seat expansion and retention. The May 2026 StackAI acquisition (about $75 million upfront cash plus equity earn-outs) is meant to extend agentic workflow orchestration across enterprise systems and, per management, accelerate the AI roadmap by more than a year. So the engine is built and idling, not absent.

The honest caveat — and why this is "exists but unproven" rather than "exists and ready" — is scale. Over $6 million of AI Studio ARR against an $800 million-plus revenue base means the second curve is roughly 1% of the business; it is moving the direction of the story well before it moves the size. And there is a structural tension the report stresses: if AI lets customers do more with fewer seats, it can lift relevance while capping the seat-based first curve, so the second curve has to grow fast enough to outrun the cannibalization it may cause. For Baillie's purposes, the second curve clears the lowest bar — it is real and commercial today — but it is far too early to underwrite as the thing that "takes over" in five years; that remains a hope supported by promising early mix data, not a demonstrated handoff.

The moat is real but only moderate — soft switching costs from the Work Graph and cross-functional adoption — and over the next 3–5 years it is more likely to hold or narrow than to widen. Asana's core competitive advantage is architectural: the Work Graph is a relationship-preserving data model linking tasks, goals, people, dependencies, and now AI agents across departments, rather than storing work in siloed "containers." That creates two sources of stickiness. First, the product gets more useful the more teams inside one organization share it, so it embeds across approvals, planning, and cross-functional reporting in larger enterprises. Second, accumulated workflow history compounds into a soft switching cost, and it gives AI agents structured context that a blank chat box lacks.

But the report is candid that this moat is weaker than a true system-of-record. Unlike accounting ledgers, HR records, developer infrastructure (Jira/Confluence), or ITSM systems, a collaboration-and-execution layer has more substitutes and lower switching pain — Asana itself has warned for years that buyers can fall back on email, spreadsheets, messaging, legacy tools, or rival suites. The competitive set proves the squeeze: Atlassian has far more scale ($1.787 billion Q3 FY2026 revenue, ~24% growth) and deeper developer/ITSM gravity; monday.com is growing 24% with materially stronger retention (110% net dollar retention overall, 115%–116% in high-spend cohorts versus Asana's 96%); Microsoft can bundle Planner, Project, Teams, and Copilot at near-zero incremental price; and ServiceNow owns more mission-critical workflow at the high end.

On the 3–5-year trajectory, the forces cut both ways but lean negative for moat width. Widening it requires the AI bet to land — if customers prefer governed agents embedded in existing workflows over disconnected copilots, the Work Graph's context advantage could deepen lock-in and the moat genuinely broadens. Narrowing it is the base risk: if AI makes generic workflow creation cheap and suite vendors use distribution to compress standalone value, the differentiation erodes. The 96% DBNRR and the management-acknowledged two-point PLG drag are early evidence the moat is currently being tested, not widening. For Baillie, this is a "good enough to earn, not strong enough to assume" moat — the kind that supports a decent business but is too contestable to underwrite a decade of compounding pricing power.

Asana shows moderate-to-decent reinvention genes and a notably honest posture toward bad news — but it is being graded on this premise right now, because its core seat-based business is already being disrupted by AI. The implicit setup of the question is live, not hypothetical: AI is simultaneously (a) the technology threatening to compress seat-based collaboration software and (b) the very channel weakening Asana's PLG funnel, which management quantifies as a roughly two-point drag on ARR growth. So the test of whether Asana "has the gene to reinvent itself when its core is disrupted" is happening in real time, and the early evidence is mixed-positive.

On the reinvention gene, the constructive signs are concrete. Rather than defend the seat model passively, Asana is layering on consumption-based AI Studio (over $6 million ARR exiting FY2026, 17% of net-new ARR in Q1 FY2027), reframing the Work Graph as a runtime for "human-agent teams," and it acquired StackAI in May 2026 to pull its agentic-workflow roadmap forward by more than a year. It also executed a hard operating reset — R&D fell to 38% of revenue in FY2026 from 47%, sales and marketing to 51% from 58% — reaching its first non-GAAP operating profit and $56.7 million of FY2026 non-GAAP operating income. That is a company willing to remake both its product surface and its cost base under threat, which is the behavioral signature of reinvention capacity.

On how it treats mistakes and bad news, the report's evidence is genuinely favorable. Management's disclosures are unusually frank: it openly states PLG is a headwind and bakes the two-point drag into guidance rather than hiding it; it concedes AI monetization is "not yet material" against the revenue base; and it disclosed that a June 2025 flaw in its Model Context Protocol implementation potentially exposed certain customer-instance data — an admission of a self-inflicted security miss rather than a cover-up. The March 2025 combination of soft guidance plus the founder-CEO succession was delivered straight, even though the stock fell more than a quarter. The limiting factor is power to reinvent, not willingness: the moat is only moderate and the AI second curve is still ~1% of revenue, so the honesty and the pivots are real but not yet proven sufficient. For Baillie, this dimension is a relative bright spot — candid, self-aware management actively reinventing under disruption — but the jury is out on whether the reinvention is big enough to matter.

Management is strongly long-term-minded and deeply bound to the company — arguably the highest-conviction dimension here — though that same binding comes wrapped in a governance structure that ordinary shareholders should discount. On long-termism and interest alignment, the evidence is clear. Co-founder Dustin Moskovitz built Asana around product conviction, design discipline, and a long horizon, and put his own capital behind it repeatedly: he participated in a 2022 private placement and his founder buying cushioned the stock floor at several points. The report explicitly notes the company has demonstrably sacrificed current profit for the long run — it ran non-GAAP operating losses for years (negative $58.1 million in FY2024, negative $40.8 million in FY2025) to fund R&D and go-to-market before reaching its first non-GAAP profit, and it is still investing into the AI pivot (the StackAI deal, AI Studio) rather than maximizing near-term earnings. That is exactly the willingness to defer profit for years 3–10 that Baillie prizes.

The succession is a nuanced positive, not a red flag. Moskovitz announced in March 2025 he would retire as CEO but remain chair and focus on product vision and AI; Dan Rogers took over in July 2025, bringing enterprise go-to-market experience from LaunchDarkly, Rubrik, and ServiceNow — the right background for the "sell higher in the enterprise, turn AI into recurring dollars" phase. So the founder stays bound to the long-term vision while a fit-for-purpose operator runs execution. The early operating evidence under Rogers (four straight quarters of NRR improvement, 11.5% Q1 FY2027 non-GAAP margin, RPO up 23%) suggests the handoff is being managed competently.

The honest deduction is governance and dilution. The dual-class structure gives Class B shares ten votes each, and the 2026 proxy shows directors and executives as a group controlling 84.6% of total voting power — so outside shareholders have little formal influence even if management acts rationally. The 2022 irrevocable proxy neutralized the votes on the placement shares but did not dismantle the underlying control. And alignment is partly funded by dilution: stock-based compensation was $214.8 million in FY2026 (larger than the year's GAAP net-loss improvement) and management guides SBC to stay in the low-20s as a percentage of revenue in FY2027 — an improvement path, not a solved problem. For Baillie, this nets out as a standout on founder long-termism and skin in the game, tempered by a real governance discount: the people running Asana think in decades and are heavily invested, but a controlled board means you are trusting them to moderate dilution on shareholders' behalf.

On the first premise — indispensability — Asana would be genuinely missed but is closer to a high-value "nice-to-have" than a must-have system of record; on the second premise — social and regulatory sustainability — its growth is benign and easily sustainable. Net: moderate on indispensability, strong on sustainability. Take the disappearance test first. For organizations that have wired Asana into approvals, planning, and cross-functional reporting, losing it tomorrow would hurt — historical workflow context in the Work Graph compounds, and 96% overall / 97% core dollar-based net retention shows most customers keep paying and broadly expanding. That is real stickiness; this is not a trivial tool.

But the report repeatedly stresses the ceiling on indispensability. Unlike an accounting ledger, an HR system, developer infrastructure, or ITSM — where ripping the product out breaks the business — Asana is a coordination and execution layer with abundant substitutes. The company itself warns customers can fall back on email, spreadsheets, messaging, legacy tools, or rival suites; the 96% DBNRR (below monday.com's 110%) and the management-acknowledged two-point PLG drag both signal that, at the margin, customers can and sometimes do consolidate or downsize away. The report's own framing — the central risk is "relevance," not margin, and the fear is Asana becoming "a nice-to-have coordination layer rather than a must-have operating layer" — is essentially a direct verdict that indispensability is moderate. If Asana vanished, most customers would be inconvenienced and would migrate, not paralyzed.

On the second premise, sustainability is a clear strength and arguably a point in Asana's favor versus flashier growth names. Helping teams coordinate knowledge work carries no meaningful societal harm, no addictive-engagement or attention-extraction dynamic, and no obvious regulatory target on the business model itself — growth here does not depend on externalizing costs onto society. The only sustainability caveat the report surfaces is operational, not ethical: the June 2025 flaw in Asana's Model Context Protocol implementation potentially exposed certain customer-instance data, a reminder that pushing AI and cross-system integrations adds data-security and trust obligations. That is a manageable execution risk, not a structural regulatory threat. For Baillie, this dimension splits: the social/regulatory license is firmly intact and durable, but the "how badly would customers miss it" half lands as moderate — valued, sticky, replaceable — which is exactly the soft spot the whole Watch rating circles.

The unit economics are attractive and structurally SaaS-like — high, stable gross margins with real operating leverage now emerging — but the cash is still substantially recycled into stock-based compensation, so per-share economics improve more slowly than the operating model suggests. Start with gross margin: it sat at 89% in FY2026 (90% FY2024, 89% FY2025, 88% in Q1 FY2027), essentially flat and firmly in best-in-class software territory. Incremental software delivery is cheap relative to subscription revenue, so the incremental gross economics are excellent and do not deteriorate at scale.

The more important recent change is below gross profit, where Asana is finally showing the operating leverage the model always promised. R&D fell to 38% of revenue in FY2026 from 47% in FY2025, and sales and marketing to 51% from 58%, while total operating expenses actually declined slightly in dollars despite higher revenue — the report calls this a genuine inflection, not cosmetic cost-cutting. The result: non-GAAP operating income swung to positive $56.7 million in FY2026 (from negative $40.8 million in FY2025), Q1 FY2027 non-GAAP operating margin reached 11.5% (a 720 bp year-over-year gain), and operating cash flow climbed from negative $17.9 million (FY2024) to $14.9 million (FY2025) to $90.4 million (FY2026). Capital intensity is trivial — FY2026 property and equipment was $3.8 million and capitalized internal-use software $9.6 million — so on an owner-earnings basis the business converts close to its adjusted free cash flow ($34.4 million in Q1 FY2027), not its still-negative GAAP earnings.

The honest deduction is where the earned cash goes, and the answer is partly "back to employees." Stock-based compensation was $214.8 million in FY2026 — larger than the year's GAAP net-loss improvement and dwarfing the $56.7 million of non-GAAP operating income — and management guides SBC to stay in the low-20s as a percentage of revenue in FY2027. So the operating turnaround is real, but a meaningful slice of the "profit" is funded by dilution rather than dropping to per-share value; GAAP net loss was still $189.0 million in FY2026. There are no buybacks or dividends; the residual cash mostly sits on a strong balance sheet (over $350 million after StackAI) and funds the AI pivot. For Baillie, the unit economics screen as a genuine strength on the cost-to-serve and incremental-margin axes — they clearly improve at scale — but the cash-return quality is only moderate until SBC trends down, which is why the report flags dilution as a central, unresolved valuation issue rather than a footnote.

A 10-year 5x from $6.92 is possible but not probable — it needs a chain of conditions to ALL hold, and several of them run against the report's base case, which is precisely why the rating is Watch rather than Buy. A 5x takes the share price to roughly $35 and, on a stable-ish share count, the market cap from about $1.65 billion toward $8 billion-plus. For that to happen, essentially every link in the following chain has to break favorably:

• AI monetization scales from token to engine. AI Studio (over $6 million ARR exiting FY2026, ~1% of revenue) and AI products (17% of net-new ARR in Q1 FY2027) must grow into a material, durable revenue stream — not just lift demos and internal efficiency. This is the linchpin and is currently unproven at scale.
• Growth reaccelerates and at least doubles revenue. From the $790.8 million FY2026 base, revenue needs to compound well above today's ~9% — realistically into the mid-teens or better — so the AI second curve must outrun seat maturity rather than merely offset it.
• Retention inflects up, not sideways. Dollar-based net retention (96% overall, 97% core) has to climb toward and through 100%, reversing the maturity that management flags via the roughly two-point PLG drag.
• Margins expand without starving the product. Non-GAAP operating margin (11.5% in Q1 FY2027) must move durably higher and SBC must fall from the low-20s as a percentage of revenue so the gains reach per-share value rather than diluting it.
• The multiple re-rates sharply. Forward EV/revenue must expand from today's ~1.5x–1.6x to a growth-software level (the report's optimistic case assumes 3.0x), which only happens if the market re-underwrites Asana as a winning AI-workflow platform.
• Competition and consolidation don't compress it first. Atlassian's scale, monday.com's faster growth and stronger retention, and Microsoft's bundle all have to fail to cap Asana's economics over a decade — a tall order in a fragmented, low-barrier category.

Are these realistic? Individually each is plausible; jointly, over ten years, they are demanding, and the report's modeled outcomes underline the gap. Its optimistic 12-month scenario implies only about 65%–80% upside (multiple to 3.0x, margins toward 12%+), and its stated expected annualized returns are roughly 1%–2% conservative, 10%–12% base, and 19%–21% optimistic over a three-year path. A ~19%–21% annualized return sustained for a decade is what a 5x requires — i.e., the optimistic case would have to become the realized case and persist. That is the tail, not the center.

What does today's price imply? The opposite of a priced-in moonshot. At ~$6.92 and ~1.5x–1.6x forward EV/revenue — a steep discount to faster-growing peers — the market is pricing credible stabilization with skepticism about the AI bet, not disaster and not triumph. The report's fair-value work centers the conservative case at roughly $6.9–$7.1 (essentially today's price), the base case at $9.0–$9.5, and the optimistic case at $12.0–$12.5; the pre-mortem sees a downside toward $3–$4 if AI fails to monetize and the multiple compresses to ~1.0x. So the current price embeds modest expectations and a wide outcome cone. For Baillie's 10-year-5x test, Asana is a low-probability candidate: the asymmetry is interesting because the entry multiple is already compressed, but the blue-sky path requires the whole favorable chain to hold, and the report deliberately declines to bet that it will.

The market mostly "can't see far enough" rather than failing to understand or looking down on Asana — it is pricing the maturing old business too literally and the AI second curve too skeptically, and it is right to demand proof. This is the rare case where the discount may already be roughly fair, so the "mispricing" is modest, not glaring. Asana is not misunderstood: it is a well-covered, transparently-reporting mid-cap with a recognizable product. Nor is the market merely snobbish about a small-cap — the skepticism is fundamentally grounded. What the market is doing is extrapolating the visible trend (growth decelerated to 9%, DBNRR at 96%, PLG a roughly two-point drag) and discounting the less-visible optionality (AI Studio over $6 million ARR, AI at 17% of net-new ARR, RPO up 23% versus revenue up 9.5%, four straight quarters of NRR improvement) because that optionality is early and unproven against an $800 million-plus base.

The "can't see far enough" framing is the report's own: it argues the stock "may be pricing the old business model too literally and the new one too skeptically." Three forces reinforce the short-sightedness. First, an anchoring scar — the March 2025 selloff (down more than a quarter after combining soft FY2026 guidance with the founder-CEO succession) reset sentiment from "founder-led growth story" to "transition asset," and the market still treats every operating gain as early evidence, not proof. Second, a genuine scale mismatch — AI is moving the direction of the story well before it moves the size, so screens and quant signals see ~9% growth, not a 1%-of-revenue product compounding 50% quarter-on-quarter. Third, a structural fear the market keeps front-of-mind: that AI helps customers do more with fewer seats, so even healthy AI engagement could cap the seat-based engine — making investors reluctant to pay up for AI claims that might cannibalize the core. Layered on top are a governance discount (84.6% insider voting control) and high SBC (low-20s of revenue), both real reasons to demand a lower multiple.

The narrative inflection point — what would force a re-rating — is specific, not thematic. The report points to a combination rather than a single print: two-plus more quarters of NRR improvement with large-customer growth holding above 10% year over year; AI products sustaining above 15% of net-new ARR while management begins to quantify the revenue impact clearly (turning AI from "mix" into measurable dollars); evidence that the 23% RPO growth converts into durable revenue rather than noise; and visibly lower SBC as a share of revenue in FY2028 planning so the profitability story becomes credible per share. Crucially, the report notes Asana "does not need to become expensive to work — it only needs to become less cheap": because the entry multiple is so compressed (~1.5x–1.6x forward EV/revenue), even a modest re-underwriting of the AI-and-retention story could move the stock without heroic growth. For Baillie, the answer to "why hasn't the market realized this" is therefore honest and humbling: largely because there isn't yet enough to realize — the market is sensibly cautious on downside structure and probably a touch too cautious on upside optionality, which is a watch-list setup, not a screaming inefficiency.