Built by Claude FableSalesforce -- How the Business Works
Salesforce is best understood not as clouds but as three stacked layers, each monetised differently. The investment question is whether value is migrating down the stack.
| Revenue line | FY27Q1 | % of total | Y/Y | Gross margin |
|---|---|---|---|---|
| Subscription & support | $10,593M | 95% | +14% (+12% cc) | 82% |
| Professional services & other | $540M | 5% | +1.5% | -14% |
| Total revenue | $11,133M | 100% | +13% (+12% cc) | 77% |
Professional services is deliberately loss-making -- $617M of cost against $540M of revenue, a negative 14% gross margin. It is a land-and-expand and implementation-assurance vehicle, not a profit centre. All the economics live in subscription, where gross margin is roughly 82%.
Salesforce revised its disaggregation this quarter into two buckets. Understanding what sits inside each is essential, because composition -- not underlying demand -- drives the headline growth rates that the whole bull/bear debate turns on.
Composition matters: Sales and Service are the two ~$10B-scale franchises and grow faster than the bucket average. Marketing (ExactTarget) and Commerce (Demandware) are legacy assets management names as a drag on every call. Slack grows double digits. The +7% is therefore a blend, not a clean read on AI seat erosion.
Composition matters here too: this is where acquired Informatica revenue ($444M in FY27Q1) lands, flattering the growth rate. But Tableau also sits here and was flagged for "increased softness" -- so the bucket prints +23% despite a declining asset inside it.
| Region | FY27Q1 | % of total | Y/Y cc | Prior-qtr cc |
|---|---|---|---|---|
| Americas | $7,233M | 65% | +11% | +9% |
| Europe | $2,754M | 25% | +12% | +13% |
| Asia Pacific | $1,146M | 10% | +12% | +13% |
Growth is unusually uniform at +11–12% cc across all three regions, and the Americas actually accelerated from +9% to +11%. That argues the deceleration is product-mix driven, not a regional demand problem -- a meaningful point in management's favour.
The strategic logic of every acquisition and product launch is a single loop. Understanding it explains why Informatica was bought and why Slack is not a write-off.
across the enterprise
+136% Y/Y
proprietary context
+111% Q/Q
on top of seats
Where the flywheel can stall. The loop assumes step one keeps turning. If agents displace the seats that generate the data, the input to the whole cycle shrinks. This is precisely why "Sales and service saw year-over-year seat growth with humans and agents both expanding" is the most important sentence management said this quarter -- it is the claim that step one is still intact. See thematics.
| Product | What it does | Pricing unit | AI exposure |
|---|---|---|---|
| Agentforce Sales | Pipeline, opportunity and forecast system of record | Per seat + premium SKU | High both ways -- seats at risk, but agents need the pipeline data |
| Agentforce Service | Case management, routing, knowledge | Per seat + per resolution | Best-monetised -- case deflection is directly billable as AWUs |
| Marketing (ExactTarget) | Campaign orchestration, journeys | Per contact / send volume | Named drag; competitive loss to Adobe/Braze predates AI |
| Commerce (Demandware) | B2C/B2B storefronts | % of GMV | Named drag; Shopify pressure at the low end |
| Slack | Messaging; increasingly the agent supervision surface | Per seat | Re-rated upward -- MCP hit 1M users in six weeks |
| Data 360 | Customer data platform; Zero Copy federation | Consumption (credits) | Pure tailwind -- metered on data, not people |
| Informatica | Data integration, governance, lineage, MDM | Consumption + subscription | Pure tailwind -- makes agentic actions auditable |
| MuleSoft | API integration layer | Per API call / core | Neutral-positive; agents need tool access |
| Tableau | BI and visualisation | Per seat | Negative -- flagged for softening bookings; natural-language query disintermediates dashboards |
The pattern is clear and is the crux of the whole thesis: every product priced per seat is contested, and every product priced on consumption is a tailwind. The mix shift from the former to the latter is the single variable that decides whether this business compounds.
| Source | Strength | Evidence |
|---|---|---|
| System of record | Very strong | Customer data, pipeline and case history live here; replacing it means re-plumbing the revenue organisation |
| Configuration switching cost | Very strong | Years of bespoke objects, workflows and integrations per org; ~8% attrition despite sustained price increases |
| Contracted backlog | Strong | $33.6B cRPO covers 73% of the FY27 revenue guide before the year starts |
| Ecosystem | Strong | AppExchange, systems-integrator channel, large certified-admin labour pool with career lock-in |
| Distribution into the base | Strong | Over 50% of Agentforce and Data 360 bookings come from existing customers; 98 deals above $1M net new AOV |
| Governance / auditability | Emerging | Informatica lineage answers "which agent changed this record, on what authority" -- the gate for regulated deployment |
| Pricing unit | Vulnerability | Revenue scales with human headcount -- precisely the variable agentic AI is meant to reduce |
Salesforce holds roughly 20–21% of the global CRM applications market, larger than the next several vendors combined. The front office is a genuine oligopoly: Salesforce, Microsoft Dynamics, Oracle and SAP, with ServiceNow, HubSpot and Adobe adjacent in specific lanes.
The threat that matters is not feature parity, it is bundling. Microsoft can attach Dynamics and Copilot to an existing E5 enterprise agreement at near-zero incremental price. That is a pricing attack, not a product attack, and Salesforce has no equivalent bundle to defend with. Against it, the switching cost on a mature Salesforce org is measured in years -- which is why the attack has compressed Salesforce's growth rate rather than caused outright share loss. The observable evidence for this reading: attrition has stayed near 8% throughout.
| Gross margin | 77% blended; ~82% on subscription |
| Sales & marketing | 34% of revenue -- down from 35%, the main margin lever and the remaining runway |
| R&D | 14% of revenue -- engineering headcount roughly flat near 15,000 for two years |
| G&A | 7% of revenue |
| Capital intensity | ~1.5% of revenue -- capex was just $145M in FY27Q1 |
| Contract acquisition costs capitalised | $2,065M current + $2,920M noncurrent; $584M amortised in the quarter |
| Billing seasonality | Heavily Q4-weighted; Q1 collects it (receivables fell $14,339M to $5,080M) |
| Unearned revenue / Q1 billings | $20,363M / $7,179M (+4.3% Y/Y) |
The model is asset-light in the extreme -- capex of 1.5% of revenue against a 34.7% FCF margin -- which is what allowed free cash flow to more than triple while revenue merely doubled. It also means the AI transition is funded from the income statement rather than the balance sheet, unlike the hyperscalers, so there is no capex supercycle to swamp FCF.
Billings grew just +4.3% Y/Y against revenue of +13%. Billings is noisier than revenue and distorted by contract duration and the Informatica addition, so it is not conclusive on its own -- but it is running well below revenue growth, and the cRPO figure partially masks it. If the 2H reacceleration is real, billings should inflect first. This is the cleanest forward indicator on the page and it is not yet confirming.