Day91 Brief — Issue 005

A · read — Release and call read. The billings figure is third-party and marked inline; everything else is the company's own.

FACT. Sprinklr's Q2 FY27, quarter ended 31 July 2026: revenue of $213.7 million, up 1%. Total remaining performance obligations of $1.03 billion, up 11%. Current RPO of $614 million, up 3% — leaving roughly $416 million beyond twelve months. Subscription net dollar expansion was 102%. CFO Manish Sarin stated that total RPO outgrew current RPO because several large renewals and expansions carried terms extending up to five years, and that average contract duration rose by more than two months for the second consecutive quarter. Billings were reported at $183.5 million, down 8.5% year over year, per third-party coverage of the quarter; that figure is not in the release language located and should be confirmed against the 10-Q. The company completed a $125 million accelerated share repurchase and ended with $453 million of cash and no debt.

READ. This is the exact inverse of the Workday shape in Issue 004, and it arrives eight days later. There, duration was compressing at a company growing 12.8%. Here, duration is extending at a company growing 1%. Longer terms push dollars past the twelve-month line, which is why total RPO can grow four times faster than current RPO and billings can fall in the same quarter — the invoices are spread out even as the commitments get bigger. Duration is not a health signal in either direction. It is a contract-structure signal, and it moves the three headline metrics in three different directions.

SO WHAT. Never read RPO, cRPO and billings as three independent reads on demand. Compute the spread between total and current RPO growth first; if it is wide, duration moved, and every other metric in the release needs re-reading with that in mind. Then ask whether your own longer terms were bought with discount — a five-year renewal at a concession is a duration extension and a price cut at the same time, and only one of those shows up in the backlog.

→ Sprinklr Q2 FY27 earnings call coverage — RPO, duration, buyback

A · read — 8-K exhibit opened. The guidance and the officer change are in the same filing as the beat.

FACT. GitLab's Q2 FY27, quarter ended 31 July 2026: revenue of $286.3 million, up 21%. The release leads with record gross bookings and net ARR growth exceeding 40%. Dollar-based net retention was 117%. Total RPO was $1.2 billion, up 16%; current RPO $744.7 million, up 20%; calculated billings up 24%. GAAP operating margin was negative 20%, against negative 8% a year earlier. Non-GAAP operating margin was 15%, against 17%. Operating cash flow was negative $3.1 million, against positive $49.4 million. Q3 revenue guidance is $281–$283 million — 1.2% to 1.9% below the $286.3 million just reported. The same 8-K discloses that Chief Accounting Officer Simon Mundy will resign effective 16 September 2026, with the CFO assuming the principal accounting officer role.

READ. Four growth rates in one document: net ARR above 40%, billings 24%, revenue 21%, total RPO 16%. They descend in order of how far the metric sits from audited revenue. The furthest one is the headline. Sequentially declining revenue guidance after a quarter described as exceptional is the company's own answer to which number to trust. Operating cash flow swinging $52 million the wrong way on 21% revenue growth is the collections view of the same thing. And an accounting officer departing in the same filing is not evidence of anything on its own, but it is a fact that belongs in your file rather than in a footnote nobody opened.

SO WHAT. Rank every growth rate a vendor publishes by distance from recognised revenue, and quote the closest one. When a competitor's deck leads with a bookings or net-ARR figure, ask for the revenue guide for the next quarter — if it is below the quarter just reported, the bookings number is describing timing, not trajectory.

→ GitLab 8-K Ex. 99.1, Q2 FY27 — SEC → GitLab Q2 FY27 results — Investor Relations

A · read — Release opened. Zscaler publishes both the reported and the ex-acquisition figures itself.

FACT. Zscaler's Q4 FY26, year ended 31 July 2026: revenue of $898.2 million, up 25%. ARR of $3,771 million, up 25%, of which $246 million was net new ARR in the quarter. The release then states that excluding Red Canary, which contributed $141 million of ARR, ARR grew 20% to $3,630 million and net new ARR grew 17% rather than 24%. Non-GAAP operating margin was a record 24.3%. FY27 guidance is ARR of $4.396–$4.426 billion, which is 16.6% to 17.4% growth on the reported $3,771 million base. The company also announced a restructuring affecting approximately 3% of employees with charges of $30–33 million.

READ. Two of the four headline growth figures move by five and seven points once the acquisition comes out, and the company published both versions in the same paragraph. Note which base the FY27 guidance is set against: growth of about 17% on the reported figure, against organic growth of 20% just delivered. That is a deceleration guide, issued alongside a record margin and a 3% headcount reduction. The CFO's framing — growth broadening beyond users into non-seat-based solutions — is the substitution story stated as strategy rather than as risk.

SO WHAT. For any vendor that acquired in the trailing twelve months, build the organic figure before you use their growth rate as a benchmark. If they publish it themselves, use theirs. If they do not, that absence is itself the finding, and it is worth putting in the deal file.

→ Zscaler Q4 FY26 results — Investor Relations

A · read — Both releases opened. The methodology sentence is Asana's own published definition.

FACT. Asana's Q2 FY27, quarter ended 31 July 2026: revenue of $216.4 million, up 10%, above the high end of guidance. Overall dollar-based net retention of 97%, against 96%; Core cohort 98%, against 96%; the $100,000-plus cohort 98%, against 95%. Customers spending $5,000 or more grew 7% to 26,778. Asana's published methodology states that its reported dollar-based net retention rate equals the simple arithmetic average of its quarterly rate for the four quarters ending with the most recent fiscal quarter. FY27 revenue guidance is $858.5–$863.5 million, approximately 9%. The CFO stated the company sees a meaningful opportunity to build consumption- and outcome-based revenue streams alongside seats, and Agentic Work Management brings AI Teammates, AI Studio and Asana Dash to every paid tier in Q3.

READ. A trailing four-quarter average moves slowly by construction, so a one-point improvement in the reported figure implies a larger move in the underlying quarter — in either direction. It also means the number cannot tell you what happened in July. Separately, every cohort remains under 100%: the base is still contracting, just less. And the two product announcements point opposite ways. AI moving into every paid tier is bundling, which lifts perceived value and lowers separately attributable AI revenue. Consumption and outcome pricing alongside seats is unbundling. Both were announced in the same release.

SO WHAT. Check whether your own net retention is a point-in-time figure or a trailing average before you compare it to anyone's. If it is an average, compute the single most recent quarter separately and look at that. A smoothed metric is the right thing to publish and the wrong thing to steer on.

→ Asana Q2 FY27 results — Investor Relations → Asana Q2 FY26 release — published NRR methodology and prior-year cohorts

B · unpulled — 8-K dated 2 September referenced in coverage; the filing itself not opened. Two sources give different Azure figures — resolved below by arithmetic, not by picking.

FACT. Microsoft announced on 2 September 2026, via an 8-K and investor presentation, that it will report Azure revenue as a separate product-and-service revenue line from fiscal 2027, and will consolidate three operating segments into two: Agents and Infra, and Devices and Consumer. Restated figures put Azure at $101.9 billion for fiscal 2026. Coverage of the June quarter conflicts: one figure is $29.42 billion at roughly one-third of total revenue, another $20.71 billion at 27%, both citing 42% growth. Derived: at $29.42 billion the remaining three quarters sum to $72.5 billion, an average of $24.2 billion, which is consistent with a business growing 42%; at $20.71 billion the other three quarters would average $27.1 billion, higher than the fourth, which is not. The $29.42 billion figure is the one that reconciles. The company also stated it will no longer separately report costs and operating margins for the three former segments, and Azure does not become a reportable segment.

READ. A disclosure added and a disclosure removed in the same filing. Investors get the Azure revenue line they have asked for since 2015 and lose segment-level cost and margin detail that let them infer where profit actually sits. Issue 004 logged Autodesk retiring its Design and Make revenue line in the same quarter it closed a $3.575 billion acquisition. This is the larger version of the same move, and it is being reported as transparency.

SO WHAT. When a vendor restructures reporting, write down both columns — what appears and what disappears — before the first restated quarter prints. After one cycle the old basis is gone and nobody reconstructs it. This one is worth doing now because two years of recast history were published alongside it.

→ Microsoft segment change and Azure disclosure, 2 September 2026 — CNBC

A · read — Release read. The share-count effect is derived from disclosed EPS and the disclosed 8% reduction.

FACT. Docusign's Q2 FY27, quarter ended 31 July 2026: revenue of $875.7 million, up 9%, including approximately 1.3 percentage points of foreign exchange benefit — roughly 7.7% without it. Non-GAAP diluted EPS was $1.16 against $0.92. The company repurchased $306.5 million of stock, reducing diluted shares 8% year over year to 193 million. Derived: holding the disclosed EPS and share figures, non-GAAP net income rose about 16%, so roughly ten of the twenty-six points of EPS growth came from the smaller share count. Current contract liabilities fell to $1.58 billion from $1.63 billion at 31 January 2026. IAM was 15.1% of total ARR, up from 12.6% in Q1. Management guided digital add-on revenue to be a roughly 1.0% full-year revenue headwind as customers move onto subscription plans.

READ. Three separate things are lifting the headline: currency, buybacks, and a product mix shift that management itself flags as a drag. The underlying constant-currency growth is under 8% and the ARR guide is 8.5–9.0%. Deferred revenue falling while revenue grows is the same signal Autodesk showed in Issue 004 from the billings side.

SO WHAT. When benchmarking EPS growth against a public comparable, strip currency and share count first. Both are disclosed, both are large here, and neither has anything to do with whether the product is winning.

→ Docusign Q2 FY27 results — press release

C · no document — Price is stated by Nvidia to the dollar. No revenue figure for Hugging Face exists in any document located, so no multiple is computable — by us or anyone.

FACT. Nvidia announced on 3 September 2026 an agreement to acquire Hugging Face for $12,930,300,000. Reporting of the statement puts $11.9 billion to shareholders and up to $1 billion as an equity-based retention program for employees joining Nvidia, with closing expected in the first half of next year. Hugging Face was last valued at approximately $4.5 billion in 2023. No annual recurring revenue, revenue, or bookings figure for Hugging Face appears in the announcement or in any filing located.

READ. An exact price to the dollar and no denominator. Every multiple published this week is someone's estimate wearing a decimal point. The disclosed structure is the more useful fact: the retention pool is about 7.7% of consideration, carved out and named, which is the same shape as Klaviyo's up-to-$17 million team purchase in Issue 004 at a different order of magnitude.

SO WHAT. Do not carry an ARR multiple for this deal into any board material. If you need a comparable for an AI platform acquisition, use the disclosed retention share of consideration instead — it is a real number and it tells you what the buyer thinks it is actually buying.

→ Nvidia announcement — stated consideration → Deal structure and prior valuation — CNN

A · read — Stated in Zscaler's own non-GAAP explanation.

FACT. Zscaler's Q4 FY26 release states a non-GAAP tax rate of 23% applied to all prior periods, aligning with the enactment of the One Big Beautiful Bill Act, with the revised rate applying prospectively. FY27 guidance assumes a 21% non-GAAP tax rate. Issue 004 recorded Okta moving its fixed long-term projected non-GAAP tax rate from 26% to 21% effective 1 February 2026.

READ. Two identity and security vendors, two weeks, two changes to the assumption sitting underneath every non-GAAP EPS figure they publish. Zscaler restated prior periods, which preserves comparability; Okta applied its change prospectively, which does not.

SO WHAT. Before comparing any two vendors on non-GAAP EPS this cycle, check both tax assumptions and whether prior periods were restated. Add it as a standing line in your competitive tracker — it will change again.

→ Zscaler Q4 FY26 release — non-GAAP tax rate

A · read — 8-K exhibit opened.

FACT. MongoDB's Q2 FY27, quarter ended 31 July 2026: total revenue of $771.8 million, up 30%, described by the company as the highest growth in several years. Atlas revenue grew approximately 29%. Enterprise Advanced and other revenue grew approximately 36%. Full-year guidance was raised, with the second-half raise attributed mainly to Atlas.

READ. For a company whose entire narrative is migration to the managed cloud product, the self-managed line outgrowing the cloud line by seven points is the more interesting sentence in the release. It suggests workloads are being placed where the customer controls the infrastructure — which is what you would expect if AI workloads are driving data-locality and cost requirements.

SO WHAT. If you sell a cloud product against an on-premises or self-managed alternative, check whether your own self-managed line is reaccelerating before you retire the sales motion for it.

→ MongoDB 8-K Ex. 99.1, Q2 FY27 — SEC

A · read — 8-K exhibit opened.

FACT. Samsara's Q2 FY27, quarter ended 1 August 2026: revenue of $508.4 million, up 30% and 29% in constant currency. Net new ARR of $134.1 million, up 28%. Ending ARR of $2.125 billion, up 30%. Customers with ARR over $1,000,000 generated over $500 million of ARR, growing over 50% year over year for the third consecutive quarter. GAAP EPS of $0.03, the fourth consecutive quarter of GAAP profitability.

READ. The counter-example the rest of this issue needs. Thirty percent growth at $2.1 billion of ARR, with the largest cohort compounding at over 50%, and no acquisition or currency adjustment required to get there. Whatever is compressing duration and flattening expansion elsewhere is not universal.

SO WHAT. When a sector narrative forms, find the company it does not describe and work out what is different about the buyer, not the product. Here the difference is that the spend is tied to physical assets and operations rather than to headcount.

→ Samsara 8-K Ex. 99.1, Q2 FY27 — SEC

Every claim above carries the state of its own sourcing. Derived figures state the calculation and the inputs. If a link is wrong or a number is misread, reply and it is corrected in the next issue with the correction stated plainly, and logged publicly.

Day91 · Document-sourced revenue signals. AI-drafted, human-approved.