SaaS Startup Plan Parameters 2026: The Benchmarks to Put in Your Financial Model

Every early SaaS financial model is built from the same dozen assumptions: how many visitors become customers, how many customers stay, what it costs to win one, what margin is left, and how much money arrives at seed. Founders usually fill those cells with numbers from a blog post or a gut feeling. Investors then check them against benchmarks that founders have never seen.

This guide gives you those benchmarks, each with the sample it comes from. Use the median as your base case, the quartiles as your bear and bull cases, and say in the model which benchmark you used. "We assume 16-month CAC payback, the 2025 median for B2B SaaS" ends the argument faster than any spreadsheet formatting.

1. Growth: what "normal" looks like now

The median B2B SaaS company grew revenue 20% in 2025, and that median has fallen for four straight years (Aleph x Benchmarkit, 2026 panel of 342 companies). A separate survey of private SaaS companies with a median ARR of $23.3M put median ARR growth at 15% in 2024 (KeyBanc Capital Markets & Sapphire Ventures, n=59 reporting KPIs).

At the very start, the numbers are harsher. Only 3.3% of software startups reach $1M ARR within one year of their first paying customer (ChartMogul, billing data from more than ten years of startups).

What this means for your model: if your plan shows 3x year-over-year growth, you are modelling a top-decile outcome. That can be right, but say why: a specific channel, a specific contract, a specific buyer. A base case built on the median is not pessimism. It is what most companies actually do.

2. The funnel: visitors to paying customers

ChartMogul surveyed 200 B2B software products in January 2026, most of them between $1M and $10M ARR. Results per 1,000 website visitors:

• Free trial, no card: 3.6 paying customers.
• Free trial, credit card required: 10.5 paying customers. Fewer people sign up, but many more of them pay.
• Freemium: 5.5% of free users convert to paid. The median free-to-paid rate across all products is 8%.

A quick check for your model (our calculation from the ChartMogul figure): 100 new customers from a no-card free trial means roughly 28,000 website visitors. If your traffic plan does not show where those visitors come from, the funnel is the weakest part of your model.

3. Retention: the number that decides whether growth compounds

Net revenue retention (NRR) is what a cohort pays you a year later, after upgrades minus churn and downgrades.

• Median NRR for B2B SaaS companies with at least $250K ARR: 82% (ChartMogul, about 2,700 companies, 2025).
• Median NRR for B2C SaaS in the same dataset: 49%. A consumer subscription loses half its cohort revenue in a year.
• Median NRR for B2B SaaS below $5M ARR in the Aleph x Benchmarkit panel: 94% (n=230, 2025).

The two B2B medians disagree by 12 points (our calculation), and that is the honest picture: ChartMogul draws on billing data from a broad base of small companies, while the Aleph panel is made up of companies that chose to report to a benchmark. For a pre-seed model, take the lower figure as your base case. If you are B2C, model churn as the main cost line, not a footnote.

4. CAC payback: how long each customer is a debt

CAC payback is the number of months of gross-margin revenue needed to recover the cost of winning a customer. From the 2026 Aleph x Benchmarkit panel (n=198 reporting, 2025 actuals):

Segment	CAC payback
Median, B2B SaaS	16 months
Top quartile	6 months or fewer
Bottom quartile	24 months or more
ACV below $5K	11 months
ACV $50K–$100K (enterprise)	22 months

Public companies are slower. Across US-listed software companies with revenue above $50M, a payback measure based on reported filings came to a median of 34.7 months in 2025 (SEC EDGAR XBRL, n=90). That number is computed differently (from total sales and marketing spend and new gross profit), so do not compare it one-to-one. It shows the direction: payback gets longer as companies move upmarket and pay for full sales teams.

For the model: an enterprise motion with a $50K+ contract and an 11-month payback is an assumption you will have to defend. A self-serve product under $5K ACV with 22-month payback is a warning sign.

5. Gross margin: what is left after serving the customer

• Software gross margin in B2B SaaS below $5M ARR: 72%, rising to 86% at $50M–$100M ARR (Aleph x Benchmarkit, 2025).
• Usage-only pricing models have the lowest median gross margin: 62%. When you resell compute, including AI inference, the cost grows with every customer.
• US-listed software companies (revenue above $50M): median 74.1% in 2025, with the middle half between 64.4% and 79.8% (SEC EDGAR XBRL, n=110).

The familiar rule of thumb about very high software margins describes mature companies. An early-stage plan near the 72% median for companies below $5M ARR is realistic. If your product calls a paid model API on every request, start lower and show the path up.

6. Team efficiency

The median B2B SaaS company generated $193K of ARR per employee in 2025, up from $150K a year earlier (Aleph x Benchmarkit, n=96). Use this to sanity-check the hiring plan: divide planned ARR by planned headcount for each year, and if the result sits far below the median, investors will notice.

On the combined growth-and-profit test, the median Rule of 40 score for companies at $1M–$5M ARR was 33% (High Alpha, 800+ respondents, 2025).

7. Fundraising: what seed looks like in 2026

• Median US seed round: $3.0M in the first half of 2026 (PitchBook-NVCA Venture Monitor). Startups on Carta reported a higher median of $4.5M at a $23.9M post-money valuation in Q2 2026, because Carta counts only priced rounds.
• Median trailing revenue at the time of a US seed round: $363K in 2025, up from $156K in 2021 (SVB, with PitchBook data). At Series A the median was $3.3M.
• Time from seed to Series A: median 1.9 years for US startups on Carta in Q4 2025.
• Only 3% of seed companies raised a Series A within 12 months (SVB, 2025 graduation cohort). Within two years, 19% of non-AI startups that raised seed in Q1 2023 had reached Series A (J.P. Morgan, PitchBook data).

What this means for your plan: a seed round has to last about two years, because that is the median gap between seed and Series A. Plan the runway so that the company reaches Series A-level revenue on seed money. Seed investors now expect revenue already in hand.

A base-case parameter sheet

Model cell	Base case (median)	Source and sample
Annual revenue growth	20%	Aleph x Benchmarkit, n=342, 2025
Paying customers per 1,000 visitors (free trial)	3.6	ChartMogul, n=200, Jan 2026
Net revenue retention, B2B	82%	ChartMogul, ~2,700 companies, 2025
CAC payback	16 months	Aleph x Benchmarkit, n=198, 2025
Gross margin, below $5M ARR	72%	Aleph x Benchmarkit, 2025
ARR per employee	$193K	Aleph x Benchmarkit, n=96, 2025
Seed round, US	$3.0M	PitchBook-NVCA, H1 2026
Revenue at seed, US	$363K	SVB, 2025
Seed to Series A	1.9 years	Carta, Q4 2025

What these numbers do not tell you

• Survivorship. Benchmark panels contain companies that are alive and willing to report. The startups that ran out of money do not appear in any median.
• Geography. Most panels are US-heavy (High Alpha: 69% US respondents). Seed sizes and revenue at seed are US numbers. European and emerging-market rounds are usually smaller.
• Segment. A median across all B2B SaaS mixes low-priced self-serve tools with large enterprise contracts. Where a benchmark gives a cut by ACV or ARR band, use the cut, not the overall median.
• Definitions. "Churn," "CAC" and "gross margin" are calculated differently across sources. When you cite a benchmark, use that source's definition in your own model.

Build the model on medians, show the quartiles as scenarios, and name the source in each cell. Then the investor conversation can be about your company instead of your assumptions.