Three methods, one answer
Nobody values a technology asset with a single method. Each of the three standard approaches answers a different question, and a defensible range comes from seeing where they agree and understanding why they disagree.
Discounted cash flow asks what the business is worth on its own economics. Comparable companies asks what the public market pays for businesses like it. Precedent transactions asks what acquirers have actually paid. When all three cluster, the answer is probably right. When one is an outlier, the reason is usually informative.
Discounted cash flow
Project free cash flow and discount it back. Straightforward in principle and highly sensitive in practice.
Use five to seven years rather than the three to five typical for a mature business, because a growing software business has not reached a steady state inside three. Build the growth from the bottom up: addressable customers, win rates, pricing and expansion behaviour. A growth rate asserted at the top level is an assumption dressed as an analysis.
Discount rates for businesses in this range commonly fall between twelve and twenty percent, and the choice within that band matters more than most of the operating assumptions.
The number to watch is terminal value, which frequently accounts for sixty to eighty percent of the total. When that much of the answer depends on a perpetual growth rate chosen by judgement, the method is telling you less than its precision implies. Run it, but treat the output as a range rather than a figure.
Comparable companies
Public trading multiples give a market reference, adjusted for the differences between a listed business and a private one.
Select peers on business model, growth rate, margin profile and end market rather than on sector label. Two vertical software businesses with different retention profiles are not comparable, whatever industry classification they share.
Then apply the discounts that actually exist. Private assets trade below public ones for size and liquidity, and a business at thirty million of revenue does not command the multiple of one at three hundred million with the same growth rate.
For subscription businesses the Rule of 40, being growth rate plus margin, has become the shorthand the market uses to sort quality. It is crude and it is what buyers reference.
Precedent transactions
The most directly relevant of the three, and the hardest to source well.
Look at transactions in the same sub-sector within the last two to three years. Older comparables reflect a different cost of capital and are misleading in either direction.
Adjust for what the headline number conceals. A multiple that includes an earnout is not the same as one paid in cash at completion, and a transaction where the seller rolled over equity is a different economic event from a clean exit. Where the terms are not disclosed, the multiple is an estimate rather than a data point.
Where a technology asset actually lands
Revenue multiples for middle-market technology assets vary widely by vertical. EdTech and other businesses with seasonal or consumer revenue sit toward the bottom, vertical B2B SaaS sits in the middle, and AI platforms with defensible data sit at the top.
The spread inside each vertical is wide, and the width is the interesting part. Sector explains where the band sits. What decides the position within it is retention, customer concentration, the credibility of the standalone cost base, and how cleanly the asset separates.
The metrics that move the position
Net revenue retention above 110% moves an asset up its band, and below 90% moves it down or out. It is the single strongest predictor in this size range.
Gross margin above seventy five percent signals a business that scales without proportional cost.
Customer acquisition payback under twelve months is strong for a middle-market business, and under eighteen is acceptable. Beyond that, growth is consuming capital faster than it returns it, which matters more when capital is expensive.
Ratio of lifetime value to acquisition cost above three suggests the growth is worth funding.
What the exercise is for
Not to produce a number. To produce a range you can defend, and a clear account of which assumptions the range depends on.
That matters because the number an asset actually achieves is set by competition among buyers, not by a model. Valuation work tells you whether an offer is reasonable and what to argue about. It does not tell you what the asset is worth, because that is decided in a room with several bidders in it.
Editorial Team · Published for orientation, not as advice on a specific transaction. Any figure cited is orientation, not a valuation. See market notes.