COGS Headcount

What it measures

Total count of full-time-equivalent (FTE) personnel whose primary responsibility is product delivery or direct customer service — the two delivery functions whose cost lands in cost of goods sold: Customer Support and Engineering & Technology. Aggregated across every product category (Software, Professional Services, Physical Products) and revenue model (Recurring, Non-Recurring), reported as a snapshot at month-end. Part-time and committed fractional roles are prorated to FTE. Excludes Sales, Marketing, General & Administrative, Product Management, and R&D — those sit above the gross-margin line and are measured as operating headcount, not COGS headcount.

Why it matters

COGS headcount is a leading indicator of whether your delivery engine scales. Unlike COGS expense — which swings with compensation, bonuses, and currency — headcount is stable and forecastable, so it isolates the structural question from the noise. Finance cares because delivery payroll is typically the largest component of cost of goods sold, so this count moves gross margin directly. The board cares because the gap between headcount growth and revenue growth is the clearest read on operating leverage: if COGS headcount outruns revenue, you are building inefficiency into your unit economics; if revenue outruns it, you are compounding leverage. It also forces the harder conversation — a business that needs bespoke support and engineering for each customer hits a scaling ceiling long before one that builds that scale into the product.

How to read it

Read COGS headcount as a ratio, never as a raw number. Pair it with revenue (COGS headcount per \$1M of revenue) and with total headcount (COGS HC as a share of the company). A declining ratio of COGS headcount to recurring revenue signals improving leverage — you are serving more revenue per delivery FTE. Track the trend against your hiring plan: are you staffing ahead of revenue or behind it, and is the gap closing? Always segment by revenue model and product line, because the blended number hides the story — Non-Recurring services often carry a much heavier headcount-to-revenue ratio and may not scale, so 20% headcount growth against flat Non-Recurring revenue is a problem even if the company total looks healthy. Use trailing-twelve-month averages when discrete hiring cohorts make any single month-end spike misleading.

What good looks like

Watch-outs

• Double-counting shared or split roles. If one engineer splits time between customer support and product development, count them once — as fractional FTE against each function — not as a full head in both. Double-counting inflates COGS headcount and masks true leverage.
• Lumping revenue models together. Blending Recurring and Non-Recurring COGS headcount into one number hides the trend that matters most. Non-Recurring services typically require a heavier headcount-to-revenue ratio and may not scale; always segment by revenue model and product line.
• Comparing raw headcount across companies. A self-serve SaaS firm and a services-heavy firm at the same revenue can carry very different COGS headcount. Always normalize by revenue, gross margin, or customer count before drawing conclusions.
• Treating budgeted positions as headcount. Open requisitions are not people on payroll. For forecasting and comparison, use the period-end actual count, not the hiring plan.
• Inconsistent FTE normalization. Counting a 30-hour-a-week contractor as 1.0 FTE one quarter and 0.75 the next creates phantom swings unrelated to real hiring. Fix the FTE basis and apply it every period.

Worked example