This is the second post in our series of insights from OPEXEngine’s 2026 benchmark release, drawing on our proprietary database combined with the BVP Nasdaq Emerging Cloud Index.

At a Glance

• Sales & Marketing (S&M) is the most consistent cost target across all cohorts; the $500M-$1B cohort cut S&M by 15 pp as a percentage of revenue in 2025.
• Larger companies cut across all three functions: S&M, R&D, and G&A. Smaller companies saw a more uneven pattern.
• R&D reductions vary widely by cohort, from -7 pp for the $10M-$50M cohort to flat for the $50M-$100M cohort, and do not yet suggest uniform, AI-driven engineering savings.
• Among larger companies, $1B-$5B, G&A was the biggest lever: down 19 pp, far outpacing reductions in S&M and R&D.
• Operating margins improved in parallel: the $500M-$1B cohort improved from 17% to 25%; the $100M-$500M cohort improved from 7% to 11%.

The first post in this series showed that larger SaaS companies are expanding margins while smaller companies show more growth stabilization. The natural follow-up question: where is that margin improvement coming from?

Our latest benchmark data answers this clearly, but with an important caveat: where companies are cutting differs significantly by size. There is no single cost-reduction playbook.

Operating Margin Broadly Follows the Profitability Story

Operating margin follows the same pattern as EBITDA margin.

Larger companies drove the clearest gains: the $100M-$500M cohort improved operating margin from 7% to 11%; the $500M-$1B cohort from 17% to 25%. Smaller companies saw little movement. The $10M-$50M cohort slipped from -5% to -6%, while the $50M-$100M cohort went from -1% to 0%.

Among larger companies in the BVP Nasdaq Emerging Cloud Index, the $1B-$5B cohort improved from -2% to 0%, and companies above $5B from 12% to 15%.

The pattern holds across both datasets: profitability gains are concentrated in larger cohorts. Revenue scale, combined with expense discipline, is translating into better margins at the top of the market.

Where the Cuts Are Coming From

_{SaaS companies are reducing expenses, especially in Sales & Marketing.}

S&M reductions were the most consistent across cohorts, with larger reductions as companies scale.

R&D was more mixed. The $10M-$50M cohort cut R&D by 7 points, the largest R&D reduction of any cohort, while the $50M-$100M cohort held flat. Larger companies cut R&D more moderately, with no consistent pattern across company sizes.

G&A varied in a different way. The $10M-$50M cohort increased G&A by 4 points, likely reflecting the growth stage and investment in operational infrastructure. The remaining cohorts all reduced G&A by 2-3 points.

Larger Software Companies Cut Differently

Among larger software companies in the BVP Nasdaq Emerging Cloud Index, G&A dominated, declining 19 percentage points as a percentage of revenue in the $1B-$5B cohort, while R&D fell 5 points and S&M just 1. For companies above $5B, cuts were more balanced across functions: S&M down 3 points, R&D down 4 points, and G&A down 6 points.

The contrast with OPEXEngine’s majority private-company database is notable. While companies up to $1B led with S&M cuts, the $1B-$5B cohort found its biggest efficiency gains in G&A. This likely reflects the maturation of finance, HR, legal, and systems infrastructure at that scale: gains that take time to build but can be significant once in place.

R&D Doesn’t Tell an AI Story Yet

The data shows no uniform, across-the-board reduction in R&D costs that would clearly indicate AI-driven productivity gains are translating into lower operating costs.

Some cohorts cut significantly: the $10M-$50M cohort cut R&D by 7 points, and the $100M-$500M cohort by 8. Others barely moved, like the $50M-$100M and $500M-$1B cohorts. R&D declined 4-5 points for $1B+ cohorts, which is meaningful but not dramatic.

That doesn’t mean AI isn’t impacting R&D, but it does mean that productivity gains aren’t translating into lower costs. Anecdotal evidence also points to AI efficiency gains being cannibalized by increased AI tooling and token costs.

We expect R&D as a percentage of revenue will likely keep shifting as AI tools mature.

What This Means for SaaS Operators

The expense data reinforces a few things that aren’t always obvious from the headline profitability numbers.

S&M is the primary efficiency lever, especially at scale. The $500M-$1B cohort reduced S&M by 15 points in a single year. That points to a structural reassessment of go-to-market spend. For companies at smaller scales, the question is whether current S&M investment is generating proportionate returns.

G&A efficiency tends to arrive later. The $1B-$5B cohort’s 19-point G&A reduction reflects years of systems and process investment paying off. Smaller companies won’t see that same opportunity yet, but building the operational foundation now — in finance, HR, and legal — sets up future efficiency gains.

R&D investment is a strategic choice, not just a cost. The variation across cohorts reflects real differences in strategy: some are reducing cost, others are protecting or redirecting investment. Cutting R&D to hit a margin target without understanding what it supports is a risk that the expense data alone can’t surface.

The margin improvements visible in 2025 are real, but they’re built on specific expense decisions that vary meaningfully by company size. The strongest performers found efficiency in the right functions for their stage, not uniform cuts across the board.

OPEXEngine benchmarks are based on annual data from 250+ SaaS operational and financial metrics, validated directly with participating companies. All data is confidential and anonymized. The BVP Nasdaq Emerging Cloud Index provides a public-market comparison based on publicly available financials.

‍