Operating Expenses

What it measures

The period sum of every cost the product and engineering organization consumes to design, build, and maintain the product: R&D, product, and engineering payroll with benefits and equity loaded in, the contractors and freelancers who augment that team, and the software, dev infrastructure, and facilities they use. It excludes cost of goods sold (the cost of delivering the product to customers), sales and marketing (the cost of acquiring them), and general and administrative overhead (finance, HR, and legal) — those sit in separate lines on the Income Statement.

Why it matters

Operating Expenses is the investment line of a software business — the dollars you spend building a product that does not yet exist in order to earn revenue that has not yet arrived. Leadership and the board read it as a bet on the future: too little, and the roadmap stalls while competitors ship past you; too much, and you burn cash faster than the product can pay it back. As a share of revenue it is the cleanest read on operating leverage — whether each new dollar of revenue costs you less to produce than the last. It feeds EBITDA, burn, and runway directly, so every headcount plan and fundraising forecast runs through it.

How to read it

Read OpEx two ways at once: in absolute dollars and as a percentage of revenue. Rising absolute OpEx is healthy when revenue rises faster — that is operating leverage, and the ratio falls. The warning sign is OpEx growing as fast as or faster than revenue, which means the product org is getting more expensive per dollar earned. A scale-up growing revenue 50% year over year should be able to hold OpEx growth well below that and watch the ratio compress. Always decompose any move into payroll versus contractors versus tools: one strategic engineering hire is investment, but five overlapping SaaS subscriptions and a creeping contractor bill are usually leakage. Strip one-time spikes — a reorg, a severance, a short consulting engagement — before you trend the line, or you will mistake noise for a structural shift.

What good looks like

Watch-outs

• Conflating OpEx with COGS. Roll cloud hosting or customer support into OpEx and you overstate gross margin and misread your unit economics — COGS scales with customer volume and belongs above the margin line, while OpEx is mostly fixed headcount in the short term.
• Booking cash spend instead of accrual. Recognize each cost in the period it is incurred, not the period it is paid. Bonus timing, a deferred-comp payout, or a large settlement can swing cash-based OpEx and break every period-over-period comparison.
• Treating contractors as someone else's line. Outsourced or freelance engineering belongs in OpEx at its fully loaded cost. Hiding it in COGS or CAC flatters apparent productivity and conceals the real cost of building the product.
• Ignoring loaded labor cost. Booking base salaries without benefits, payroll taxes, and stock-based compensation understates true OpEx by 25–50% and inflates margins that do not exist — load every head before you trend the line.

Worked example