Pre-Revenue Valuation Ceiling: Why Comparable Round Data Beats Berkus and Scorecard Methods Every Time

Pre-revenue valuation is the process of setting a price on a startup that has no revenue, no customers, or too little of either to run a discounted cash flow. The Berkus Method and Scorecard Method were built to solve that problem with structured judgment, but comparable round data — pulled from Carta, PitchBook-NVCA, and AngelList — now sets a market ceiling that no rubric-based valuation can override. If your Berkus number lands above the local seed median, no investor will honor it, no matter how many boxes you check.

This guide walks operators, founders, and angel investors through why comparable round data beats Berkus and Scorecard methods every time in a 2025 fundraising environment, with specific pre-revenue valuation numbers, a step-by-step benchmarking framework, and an example that shows exactly how the ceiling gets enforced in real deals.

The Berkus Method: A Structured Guess Frozen in the 1990s

Dave Berkus, the American angel investor who chairs the Tech Coast Angels' family of funds, published the Berkus Method in the mid-1990s as a reaction to the fact that fewer than one in a thousand startups meet their financial projections in the periods planned. He wanted a starting valuation that did not depend on a founder's spreadsheet. The Angel Capital Association's own update, "After 20 Years: Updating the Berkus Method of Valuation," restates the current form: five elements of risk, each worth up to $500,000, capped at a $2.5 million pre-money valuation.

The five Berkus buckets are:

Sound idea — up to $500,000 for basic value of the concept
Prototype — up to $500,000 for reducing technology risk
Quality management team — up to $500,000 for reducing execution risk
Strategic relationships — up to $500,000 for reducing market risk
Product rollout or sales — up to $500,000 for reducing production risk

The method has three structural problems for a 2025 raise. First, it caps the pre-money valuation at $2.5 million, which is well below where actual seed rounds price today. Second, it treats every industry the same, so a pre-revenue biotech and a pre-revenue AI infrastructure company get the same ceiling despite radically different capital requirements. Third, it does not adjust for geography or vintage — a Berkus number computed in San Francisco in July 2026 is identical to one computed in Tulsa in January 2019.

Practical takeaway: Use the Berkus Method as a sanity floor, not as a market price. If your Berkus number exceeds the local seed median, the market — not the rubric — decides.

The Scorecard Method: Better Structure, Still Anchored to a Comp

Bill Payne, a US angel investor and Kauffman Foundation instructor, published the Scorecard Valuation Methodology in 2001 and revised it most recently in 2019 through the Angel Capital Association. Unlike Berkus, the Scorecard Method is explicitly a top-down comparison. You start with the median pre-money valuation of pre-revenue companies in the target company's region and sector, then multiply that median by a weighted factor score.

The Payne 2019 factor weights are:

Strength of the management team — 0 to 30%
Size of the opportunity — 0 to 25%
Product or technology — 0 to 15%
Competitive environment — 0 to 10%
Marketing, sales channels, partnerships — 0 to 10%
Need for additional investment — 0 to 5%
Other factors — 0 to 5%

Payne is explicit that the strength of the management team is and always will be the most important factor in valuing a pre-revenue startup — the largest single weight. This is a real improvement over Berkus, because it forces the investor to pick a comp and then argue for a premium or discount against it. But the entire method still collapses to one number: the median pre-money valuation you plug in at the start. If that median is stale, borrowed from the wrong sector, or borrowed from the wrong geography, every downstream weighting is arithmetic on a bad anchor.

Practical takeaway: The Scorecard Method is only as good as the comparable it uses. If you cannot cite the specific dataset, quarter, and sector behind your median, you are not doing Payne's method — you are doing Berkus with more decimals.

What the 2025 Data Actually Says About Pre-Revenue Ceilings

Three datasets set the actual pre-revenue valuation ceiling in 2025 — Carta's cap table dataset, the PitchBook-NVCA Venture Monitor, and AngelList's H1 2025 State of Venture report. Together they cover tens of thousands of live rounds and are the same data institutional seed funds use to defend their offers.

Carta's State of Pre-Seed Q3 2025 report shows that median SAFE valuation caps in 2025 sit around $10 million for pre-seed rounds between $250,000 and $1 million, and around $15 million for rounds between $1 million and $2.5 million. Carta's State of Private Markets Q3 2025 puts the median pre-money valuation on new primary seed rounds at $16 million, up 14% year over year, on a median cash raise of about $4 million.

The Q3 2025 PitchBook-NVCA Venture Monitor, published in October 2025, reports a median VC pre-money seed valuation of $13.6 million with an average of $15.8 million as of September 30, 2025. It also flags a dispersion signal every founder should note: the median pre-seed AI and ML pre-money valuation reached $28 million with an average of $45 million — roughly double the non-AI comp for the same stage.

AngelList's H1 2025 State of Venture report puts median pre-seed pre-money at $10 million and median seed pre-money at $20 million, held flat from 2024. Y Combinator's Standard Deal page currently walks through a conversion example at a $15 million post-money cap, which is the anchor most YC batches negotiate against.

Practical takeaway: Before you compute any Berkus or Scorecard number, write down four data points: the Carta median for your stage this quarter, the PitchBook-NVCA median for your sector, the AngelList median for your geography, and the YC post-money cap. That is your ceiling grid.

Why Comparable Round Data Beats Berkus and Scorecard Methods Every Time

The comparable rounds ceiling wins for four specific reasons that Berkus and Scorecard cannot overcome:

Rubric methods are not price-discovered. A Berkus checklist can produce any number the checker wants. A PitchBook median is the actual clearing price where thousands of investors and founders agreed to trade equity. When you show a lead investor a Berkus number of $6 million and the Q3 2025 PitchBook median for your sector is $13.6 million, the number that survives the term sheet is the market comp, not your rubric.
Rubric methods have no vintage or sector adjustment. The Q3 2025 PitchBook-NVCA data shows the pre-seed AI premium is roughly 2x the non-AI median. Berkus and Scorecard both blind you to that dispersion because their inputs are the same regardless of what your company does. Comparable data forces you to price the sector premium honestly.
Rubric methods ignore round structure. Carta's data shows that in 2025 the SAFE cap for a $250,000 raise is materially different from the priced-round pre-money for a $4 million raise. Berkus and Scorecard produce one number and say nothing about instrument, size, or dilution — the three variables that actually govern founder outcomes.
Rubric methods anchor high and lose the round. The Ewing Marion Kauffman Foundation's Angel Investor Performance Project, which tracked 3,097 angel investments, reported a 2.6x cash return and roughly 27% gross IRR — but 52% of individual exits returned less than the invested capital. Angels who overpay at entry cannot make up the loss on the winners. Rational investors walk from a pre-money that violates the comp median, and that is why the comp always wins.

Practical takeaway: Rubrics are for internal reasoning. Comparable round data is for setting the number you write on the term sheet. Use the rubric to argue why you deserve the top quartile of your comp band — never to argue past it.

Step by Step: Build the Comparable Round Ceiling in an Excel Template

Here is the exact process a founder or angel should run in a spreadsheet model before opening a raise. This is the same workflow institutional investors use, and it takes about 30 minutes if you already have the source PDFs open.

Pull the four source medians. Open the current Carta State of Pre-Seed and State of Private Markets pages, the current PitchBook-NVCA Venture Monitor PDF, and the current AngelList State of Venture report. Record the median pre-money for your stage, the average, the 25th percentile, and the 75th percentile in a single Excel tab called Comps.
Apply sector adjustments. If you are in AI or ML, add the sector premium PitchBook-NVCA publishes for that sector (roughly 2x on Q3 2025 pre-seed data). If you are in life sciences or hardware, apply a discount using the sector table in the same report.
Apply geography adjustments. Use the AngelList geography breakdown to discount out of San Francisco, New York, and Boston medians. Rest-of-US pre-seed medians typically sit 20-30% below Bay Area medians for the same stage.
Compute the Scorecard number using the adjusted median. Apply Bill Payne's weights (management 30%, opportunity 25%, product 15%, competitive 10%, sales 10%, capital need 5%, other 5%) against the adjusted median from step 3, not against a national average.
Compute the Berkus number as a floor check. Run the five Berkus buckets. If the Berkus number is above your Scorecard number, your rubric is too generous. If it is below, use it as a walk-away floor.
Set your ask in the top quartile of the adjusted comp band, not above it. If your adjusted Carta median is $16 million and the 75th percentile is $22 million, ask for $22 million with a walk-away at $16 million. Never open above the 75th percentile of comparable rounds — that is the ceiling investors will not honor.

Practical takeaway: A pre-revenue valuation model is a two-tab Excel spreadsheet: comparables on tab one, rubrics on tab two. The number you send to investors comes from tab one. Tab two is your defense of the premium.

Worked Example: A Pre-Revenue B2B SaaS Startup in July 2026

Imagine an unrevenued vertical AI SaaS startup based in Austin with a working prototype, two experienced founders, one signed design-partner LOI, and no revenue. Here is how the three methods produce three very different numbers.

Berkus: Sound idea $400K + prototype $400K + team $500K + strategic relationships $300K + rollout $0 = $1.6 million pre-money. This is the number you get from the rubric alone. It is well below any current seed comp and would leave meaningful money on the table.
Scorecard against a national median: Adjusted comp median of $16 million (Carta Q3 2025) × 1.05 weighted factor score = $16.8 million pre-money. Better, but the comp median is national and untuned for sector.
Comparable-round ceiling with AI premium: Q3 2025 PitchBook-NVCA pre-seed AI median of $28 million, discounted 20% for Austin vs. Bay Area, gives an adjusted comp of $22.4 million. The 75th percentile at $32 million becomes the ceiling. Ask $22 million pre-money with a walk-away at $16 million.

The Berkus number and the market ceiling differ by more than an order of magnitude on the same company. That is the size of the mistake founders make when they price a raise off a rubric instead of comparable round data.

Conclusion: Templates Turn a Two-Day Process Into a 30-Minute Model

Pre-revenue valuation is not a philosophy problem; it is a data problem. The Berkus Method and the Scorecard Method are useful thinking tools, but the number that survives negotiation is the number the market has already cleared for companies at your stage, in your sector, in your geography. Every quarter Carta, PitchBook-NVCA, and AngelList publish that number for free.

The pain point is that stitching those four data sources into a defensible valuation model — with sector adjustments, geography adjustments, Scorecard weightings, Berkus floor checks, and a live comp table — takes hours the first time and reliably breaks the next time you have to update it. This is exactly the workflow a purpose-built pre-revenue valuation Excel template solves. Load the current Carta and PitchBook-NVCA medians, pick your sector and geography from a dropdown, and the spreadsheet produces your walk-away floor, your target ask, and your ceiling on one tab. That is the difference between opening a raise with a rubric and opening a raise with a market comp — and it is the difference between a term sheet and a rejection.