Playbook › Playbook
Swapping equity comp for cash comp is neutral on owner earnings and looks like progress
Claim
SBC-adjusted free cash flow is defined FCF − SBC. Move $X of compensation from equity to
cash and both terms fall by $X:
(FCF − X) − (SBC − X) = FCF − SBC
The metric is constructed to be indifferent to compensation currency. That is the whole point of it, and it is why the swap cannot show up as an improvement.
What does move, loudly and in three different directions at once:
| Measure | Direction | Reads as |
|---|---|---|
| SBC / revenue | down | discipline ✅ |
| Non-GAAP operating margin | down | deterioration ❌ |
| GAAP operating income | up | inflection ✅ |
| Diluted share count | down/flat | genuine ✅ |
| SBC-adjusted FCF | unchanged | nothing |
Three of those four headline movements are artifacts of the same accounting choice, and they point in contradictory directions — which is precisely why the swap generates confused coverage and why it must be identified before any of the four numbers is interpreted.
Evidence — Atlassian FY27 guidance, disclosed 2026-08-06
The shareholder letter (p.20) states the mechanism outright, which is unusually honest and is what made it detectable:
"In FY27, as part of our ongoing effort to reduce stock-based compensation as a percentage of revenue, we will change the compensation mix to include a greater proportion of cash, and a lower proportion of equity, which we expect to reduce non-GAAP operating margin in FY27 by approximately three ppts."
| Line | FY26 actual | FY27 guided | Δ |
|---|---|---|---|
| SBC | $1,605.1M (24.42% of revenue) | $1,411.1M (19.0%) | −$194.0M |
| Incremental cash comp (3.0ppt × $7,426.7M) | — | +$223.0M | +$223.0M |
| Total compensation cost | $1,605.1M | $1,634.1M | +$29.0M (+1.8%) |
Zero of the $194M SBC decline is a reduction in total compensation cost. More than 100% of it is a transfer into cash opex. Absent the swap, SBC dollars would have risen $29M on 13% revenue growth.
The $29M excess is itself diagnostic. A clean same-year swap makes the two equal. The gap appears because cash is paid 100% in the year granted while the SBC saving is only the current year's slice of grants never made on a multi-year vest. The cash cost front-runs the expense saving, so year one of a swap makes owner earnings slightly worse, not neutral.
The other half: the ratio improves on the denominator
Atlassian's stated target is "reduce stock-based compensation as a percentage of revenue." Total comp went 24.42% → 22.0% of revenue — a real 2.4ppt improvement — bought entirely with 26% revenue growth, not with cost discipline. A percentage-of-revenue target is satisfiable by the denominator alone, and in a fast-growing company it usually is.
What the swap IS worth — do not dismiss it
This note is a caution against miscrediting, not against the policy. The swap is genuinely good for the holder, for a reason unrelated to earnings:
- It ends the dilution treadmill. Atlassian spent $1,800.5M on buybacks against $1,319.1M of FCF in FY26 — 136% of FCF, funded by draining the balance sheet, purely to offset SBC.
- Paying cash directly is cheaper and more honest than issuing stock and buying it back, especially when the buyback happens at a depressed price. It converts a hidden, share-count cost into a visible, income-statement one.
- It is the opposite of the behaviour that disqualified ZS, HUBS, MDB and PD in the March 2026 software screens ([[software-saas]]), where SBC was allowed to run and the share count with it.
So: a governance improvement, not an earnings one. Credit it in the moat/management column and score exactly zero of it in the cash flow column.
Rule
When a company guides SBC down, ask immediately: is total compensation cost falling, or is the currency changing? The tell is a simultaneous unexplained decline in non-GAAP operating margin. If non-GAAP margin falls by roughly the same percentage of revenue that SBC falls, it is a swap and SBC-adjusted FCF will not move.
Cheap detection, in order:
1. Δ SBC (as % of revenue) versus Δ non-GAAP operating margin. Roughly equal and opposite → swap.
2. Compute SBC + incremental cash comp and compare to prior-year SBC. Flat or up → swap.
3. Only then read the SBC ratio.
Why this matters more from 2026 onward
SBC discipline has become a stated priority across software after the 2026 de-rating. Expect more companies to run this play, and expect the sell-side to report the ratio improvement and the margin decline as two separate stories. The margin "miss" and the SBC "win" will frequently be the same journal entry.
What would falsify this
A company reducing SBC without a corresponding rise in cash compensation — i.e. total compensation cost genuinely falling, through headcount reduction or lower per-head comp. That is a real improvement and it does move SBC-adjusted FCF. The two cases look identical in the SBC line and are distinguished only by what happens to cash opex, which is why step 2 above is not optional.
Related
[[software-saas]] — records SBC-versus-FCF as the single most common disqualifier in the field, which is what makes correctly reading an SBC decline worth its own note. [[principle-primary-source-beats-vendor]] — the mechanism here was disclosed only in the shareholder letter, not in the press release, and appeared in no news summary.
History
- 2026-08-07 — established from Atlassian's FY27 guidance, found during that name's
/analyze. Caught only because the non-GAAP margin guide-down (30.4% → 25.0%) was large enough to force an explanation, and the explanation was in the shareholder letter rather than the earnings release. A smaller swap would not have prompted the question.