How we value technology companies

Why the P/E ratio misleads

A growing software company spends money today to have revenue in three years: sales, engineering, customer acquisition. Those costs hit the income statement immediately; the benefit arrives later.

The result: the faster a healthy company grows, the worse its earnings look. A P/E ratio then penalises precisely what creates value. Conversely, a company can flatter its earnings in the short term by stopping investment — and sell its future in the process.

So we value on revenue, growth and capital efficiency.

The starting point: EV/Revenue, normalised for growth

Enterprise value to revenue only means something when set against the growth rate. 6× revenue at 45% growth is an entirely different proposition from 6× revenue at 8%.

We therefore always compare within a peer group of similar growth profile — not against a sector average.

Rule of 40

Growth (%) + margin (%) ≥ 40

The metric captures the fact that growth and profitability are interchangeable: 50% growth at −10% margin is as healthy as 15% growth at +25% margin.

Which margin is meant is usually left unsaid — and that is not a detail. The rule is commonly attributed to Brad Feld. Strictly speaking, the venture capitalist made it public on 3 February 2015; he did not devise it. In his post he explicitly credits an unnamed late-stage investor. And he wrote "profit" — not free cash flow margin. The free cash flow version in circulation today is a later mutation of market practice, not part of the original formulation. Analyst reports now put EBITDA margin, adjusted EBITDA margin or operating margin in the same slot.

For a growing software company the gap between those measures is easily ten points. Two houses can report 33 and 45 for the same company, and both are calculating correctly.

Where it misleads:

• The second term is not defined anywhere — without knowing which margin was used, the number cannot be read
• One-off effects in the cash flow margin (prepayments, tax refunds)
• Stock-based compensation, which separates accounting expense from cash outflow
• Very small companies where a single large customer dominates the numbers

So we show how the 40 is reached and what it was calculated with, not just whether. In detail: Rule of 40.

What actually signals quality

None of these thresholds is prescribed anywhere. They are conventions of market practice, formed around large American cloud vendors — no standard mandates them.

Net revenue retention, and why a blanket threshold is worthless

Net revenue retention is the most informative of the four. It measures whether revenue from the existing customer base grows without new customers. Above 100%, the company grows even if it wins no new business at all.

The thresholds in circulation, however, come mostly from listed cloud vendors selling to large enterprises. Applied to a small software company, they measure the wrong business. The defensible benchmarks come from the annual surveys of private SaaS companies; SaaS Capital, KeyBanc/Sapphire and Benchmarkit independently arrive at the same order of magnitude.

So the median sits just above 100%, not at 115%. And it splits further once you separate by customer size:

Hence the rule we read the number by: a threshold only means something inside its segment. A vendor selling to small customers at 97% is sitting exactly on its own benchmark. An enterprise vendor at 97% has a problem. Same number, opposite conclusions.

Alongside this sits the Bessemer Venture Partners framework — 100% good, 110% better, 120%+ best. It is citable and widely used, but it is a target drawn from venture capital practice, not a market average. Useful as a bar to clear; not a benchmark.

One further point tends to get lost: NRR is not defined in any accounting standard — neither IFRS nor US GAAP. It is not even a "non-GAAP financial measure" under SEC rules, for which a reconciliation to a reported figure would be required. Every issuer defines it itself, every issuer may change that definition, and nothing is audited. Many companies disclose exactly that in the risk section of their annual report. We therefore do not adopt an NRR without the definition footnote that goes with it.

Gross margin is a fork in the road: below 70% you are usually looking at a services business with a software veneer — and that belongs on services multiples, not software multiples. The difference is a factor of three or more.

The small-cap exception

The framework above comes from the SaaS world. A large share of technology small caps is something else:

• Hardware and semiconductors → cyclical; value on EV/EBITDA across the cycle, not on revenue multiples at the peak
• Project businesses → order book and margin per project, not recurring revenue
• Pre-revenue (AI, deeptech) → there is nothing to multiply. We treat these like an explorer: technology, team, cash runway and dilution — and we state explicitly that no defensible valuation is possible

In all three cases the Rule of 40 simply does not apply. Calculate it anyway and you get a number without meaning.

What this method cannot do

• It does not detect technological disruption. A model showing 30% growth still shows 30% growth while a competitor is making the product obsolete.
• Multiples are sentiment. They move with rates and risk appetite. The same company traded at 20× revenue in 2021 and 4× in 2023 with no operational change.
• Accounting latitude in capitalising development costs distorts comparability.
• The benchmarks are foreign. The median, quartiles and segment figures above come from US surveys of private software companies. For a German small cap they are a reference point, not a yardstick.

What we do not do

We do not compare revenue multiples across companies with different growth rates. We do not strip out stock-based compensation. We do not quote an NRR threshold without the customer segment it applies to. And we do not value a services business on software multiples just because the company calls itself a platform.

Sources

• Brad Feld: The Rule of 40% For a Healthy SaaS Company, 3 February 2015. Feld writes "profit" and credits the rule to an unnamed late-stage investor. feld.com
• SaaS Capital: What Is a Good Retention Rate for a Private SaaS Company? — median, quartiles and the breakdown by customer segment for private software companies. saas-capital.com
• Bessemer Venture Partners: State of the Cloud — origin of the "good 100%, better 110%, best 120%+" framework. A target, not a market average. bvp.com
• The order of magnitude of the median is independently confirmed by the annual surveys from KeyBanc/Sapphire and Benchmarkit.

What changed in version 2

11 August 2026. We checked the figures on this page against the primary sources. Two of them were wrong.

• The NRR thresholds were not calibrated. Version 1 said "above 110% good, above 120% strong". Those are the numbers of large listed cloud vendors selling to enterprises, applied to small companies. The median in private software is 101 to 102%, and the lower quartile is 97%. Under the old table a vendor selling to small customers would routinely have looked weak while sitting exactly on its own benchmark. Replaced by a reading per customer segment.
• The Rule of 40 was misattributed and misdefined. Brad Feld made it public on 3 February 2015; he did not invent it, and credits an unnamed late-stage investor. He wrote "profit", not free cash flow margin. Version 1 presented the free cash flow version as the definition. It is a later mutation of practice.
• New: the clarification that NRR is defined in no accounting standard, and a list of sources.

No published valuation is affected. aktienanalyse.online operates in "start without an analyst" mode: we publish neither a fair value nor price targets. Not a single valuation was issued under version 1 of this page that relied on the old figures.

Version 2 · 11 August 2026