5 Software Valuation Metrics: How AI Has Raised the Bar
- How AI has changed the way buyers and investors weigh the five core metrics for valuation
- What a strong metrics story looks like for founders preparing to sell or raise capital
AI has changed the way buyers and investors evaluate the five core metrics that drive software valuations: recurring revenue growth, retention, gross margin, market size, and sales efficiency. Its widespread adoption has reshuffled which metrics matter most and redrawn the line between a strong performance and a weak one.
Founders working from an older valuation playbook may be measuring themselves against a bar that has already moved.
Below is a summary of what buyers have historically looked for and how we are seeing that change in current processes.
Recurring Revenue Growth
Growth Leads, and the Definition of Recurring Is Expanding
Historically
Because a valuation reflects a company’s ability to generate future revenue, growth has always played a central role. Buyers preferred contractual revenue that is easy to predict over transactional revenue, and revenue quality was largely judged by one question: what percentage of revenue is SaaS? Professional services and implementation revenue was viewed as one-time.
Growth rate also mapped fairly cleanly to valuation, with companies growing 0 to 20% a year trading at roughly 1x to 5x trailing revenue, 20 to 55% at 3x to 10x, and 55% or more at 5x to 15x or higher.
Now
There is a lot of growth in the market right now, and companies are reaching scale faster than they used to. A business that is not growing invites the question of why not.
In our observation, here is how growth maps to multiples today: companies growing 0 to 20% a year trade at roughly 1.5x to 3x trailing revenue, 20 to 55% at 3x to 6x, and 55% or more at 5x to 10x or higher.
What has changed is the second half of that equation: what counts as recurring. AI has brought a wave of usage-based pricing, from cost-plus token models that add a fixed margin to the vendor’s token cost, to packages of credits, API calls, or tokens that bill more once the package runs out. Usage typically spikes at adoption, moderates, then grows steadily as the product reaches critical mass.
Buyers are more comfortable with this revenue than they were with implementation and services revenue because the margin is better. But they are asking a harder question about it: how sticky is it? When a customer gets nervous, token spend is among the first things to be throttled back. A fixed monthly contract simply gets paid. In our view, highly recurring contractual revenue is still worth prioritizing as a path to a strong business.
Two signals buyers are watching closely:
- Seat-based pricing is not a good place to be. The premise of AI is fewer people doing more, and everybody knows it.
- New bookings matter more than total growth. If all of your growth comes from upselling and cross-selling existing customers, with no net new logos, buyers read that as a sign that new customers no longer think the product matters.
Neither signal is difficult to get ahead of. Founders who can clearly show where new revenue is coming from, and how durable it is, should be able to answer the question before it gets asked.
Retention
Gross Retention Is Now a Proxy for AI Risk
Historically
Retention has long measured revenue predictability and product-market fit, and it influences growth, gross margin, and sales efficiency. Good retention varies by industry and customer size, but buyers and investors have consistently rewarded above-average rates.
Now
Retention remains critical, but now buyers are focusing more closely and consistently on gross retention, which counts only the revenue lost to churn and downgrades and ignores any expansion. With AI tools widely available, every renewal gives a customer the option to buy new AI products, re-up with the current vendor, or adopt that vendor’s own AI capabilities.
Strong gross retention shows the software is still valuable in that environment and that frontier AI tools have not replaced it to the point that customers are leaving. As a result, gross retention has become even more important in calculating valuation.
In our observation, here is where the gross retention bar sits today:
- 95% has always been strong, and at least 90% is becoming the minimum.
- 80-89% might have been enough to complete a transaction 18 to 24 months ago. Today, that is unlikely.
A number that cleared a process two years ago may not clear one today, which makes gross retention worth tracking by cohort well before a conversation with a buyer starts.
Gross Margin
From Benchmark to "It Depends"
Historically
Gross margin in software can often exceed 80%, which puts the business model in a class of its own. A high margin signals an opportunity for profitability at scale and the ability to reinvest earnings in growth, and it also served as a heuristic for whether a company’s value came from software. The lower the margin, the more likely a company is to be valued on EBITDA rather than revenue, which can materially impact valuation.
Now
Gross margin has become an "it depends" metric, to the point that it may no longer belong among the headline five. Two forces are pulling in opposite directions:
For traditional software businesses, gross margins are rising because of AI and automation. A customer support team, historically a large part of cost of goods sold, can now be replaced by automation. A strong margin is now table stakes, and we are seeing buyers expect gross margin above 80% as a minimum for these companies. Gross margin used to be a quick way to tell whether a company’s value came from software rather than services. That shortcut no longer tells buyers much.
For businesses whose products consume significant volumes of frontier model tokens, those tokens are a new line in cost of goods sold that did not exist two years ago, when COGS was mainly hosting and customer service. These companies may report gross margins nowhere near the traditional benchmark and still be strong businesses. For them, the evaluation depends on factors other than the margin profile.
It is too early to say whether AI is pushing more businesses toward EBITDA-based valuation. A revenue multiple almost always works mathematically, because nearly every business has revenue. An EBITDA multiple does not, because a company may have little or no EBITDA to multiply. Founders would be well served to be aware of their EBITDA multiple and factor it in, unless the business is growing 80% or more, in which case a revenue basis is more likely.
The useful exercise is knowing which of these two pictures your business presents and being ready to explain it. Buyers are not applying a single standard here, which could leave more room than usual to make the case.
Market Size
TAM Now Comes With Market Risk
Historically
Buyers have always placed higher valuations on companies with a larger total addressable market (TAM), because the market needs to be large enough to grow the business further. Market size has been especially important to investors who need to generate a return within a three-to-five-year hold.
Now
The importance of market size has gone up significantly, and it is best evaluated alongside market risk. Concern over TAM was at the heart of the SaaSpocalypse: the fear that frontier models such as Claude or GPT could do everything software companies were selling, leaving them no market at all. Since then, AI’s effect on TAM has played out in three ways.
Some markets have contracted. Content is a clear example. Software that produces content served a large market for a long time. Demand for content is still there, but what customers will pay for that software is largely gone.
Some markets face pricing pressure. This is the middle ground. It is not the end of the world, but buyers are focused on it. Consider a business whose customers pay $30,000 a year, renew consistently, and keep signing good deals. That remains a good business. But a plan to grow those same customers to $80,000 a year is now viewed more skeptically. At some price point, a customer may build the capability itself, use AI for a cheaper alternative that is good enough, or, facing a seven-figure annual cost, hire a small team to do the work.
The result is concern about price caps for application and vertical software that did not exist when building an alternative was impractical. The customers still exist, but what they are willing to pay has come down. Some industries are not susceptible to this, particularly where pricing is modest and customers consistently get the value.
Some markets have expanded. Software TAM has always been best measured by what potential customers would actually spend, not by inflated headline figures, and that spend was confined to the software budget. Businesses offering AI solutions that deliver outcomes can now reach the services budget as well, where some companies pay consultants tens of millions of dollars. Markets that were software TAM can now include professional services TAM. These tend to be more legacy, less efficient industries where a new competitor can take meaningful share.
Buyers are crediting that expansion when it shows up in the numbers, such as software that lets customers stop paying a consultant or accountant. At the same time, TAM has become a bigger risk point in processes and is less well-defined than it has been for years. In our experience, market size and market risk have decided the outcome of a significant number of recent deals, in both directions.
Market risk cuts both ways, and founders in categories where AI is expanding the opportunity have a stronger story to tell than they did a year ago.
Sales Efficiency
More With Less, and a New Focus on Top of Funnel
Historically
Buyers look at sales efficiency, measured through LTV:CAC and similar metrics, to see how good a company’s sales and marketing engine is at creating value. A track record of acquiring customers at a much lower cost than their lifetime value supports a higher valuation, while costs closer to the value generated signal that returns will diminish sooner.
Now
Sales efficiency should be improving across the board. AI-generated funnel building has made customer acquisition costs much lower, and doing more with less is the central promise of AI.
That is especially meaningful for bootstrapped founders, whose constraint has always been the raw dollars they can deploy; greater efficiency with those dollars can translate into faster growth. Because the improvement is broad, it is expected everywhere.
Buyers are also paying more attention to top-of-funnel marketing metrics as part of their AI risk evaluation, much as they do with gross retention:
- Are people still searching for this category?
- Does the business come up when they do?
- Are prospects still visiting the website and filling out lead forms?
This matters because retention can mask a problem. A business with three- to five-year contracts may show strong retention simply because customers have not yet come up for renewal, even as prospects who no longer plan to buy that kind of solution stop filling out lead forms. Buyers want to know whether leads, MQLs, and SQLs are holding up.
These are metrics most founders already have at the ready, but looking at them the way a buyer would, well before a process, puts them in a stronger position when the time comes.
A Note About Scale and Maturity
Revenue scale shapes how buyers interpret each metric. Growth of 60% means something different at $2 million in ARR than at $10 million. Longevity works alongside it. A track record of more than three years lends legitimacy that buyers and investors like.
With so many young companies growing quickly, buyers are asking how sustainable and how real that growth is. That is somewhat fair, though there are many excellent three-year-old businesses.
A related dynamic is brand. Customers are seeking safe havens with trusted brands and bigger companies, because everyone is promising everything with AI and skepticism is high. A vendor that has been in business for 20 years and has an established customer base can look like the safer choice.
Another notable shift is speed. We are seeing many companies reach $7 million to $10 million in ARR in two to three years, compared with four to five years previously, driven by AI.
The Subjectivity of Valuation Has Increased
Buyers and investors do not all work from the same valuation model. Even when every buyer reviews the same data, the range of valuations can be broad, and intangible factors have a big influence on the outcome.
What has changed is that the range has widened at both ends. The top has moved higher, while businesses that would have been much more attractive a year ago are now noticeably less so. Disagreements among buyers tend to concentrate on binary market risk questions that did not exist before, such as whether accountants will exist in five years. A strategic buyer might say yes while a private equity firm says no, or the reverse.
Proprietary data, system-of-record status, and payments still carry significant weight as intangibles. The more a business can be a point of integration for other systems, the better, which is why system-of-record status is so important. A killer feature set for a vertical market, or a product that is simply better, is no longer a moat. A few functionality advantages are not enough; the story has to go further.
What a Strong Metrics Story Looks Like Now
Out of context, metrics can send the wrong message, and that context now runs through AI. In our view, a strong narrative today looks like this:
- Growth: Strong contracted ARR growth.
- Gross margin: Expanding across the board because functions such as customer support have been streamlined.
- Gross retention: Holding steady in the 90s. It does not necessarily need to improve.
- Net retention: Rising modestly, ideally because a new AI feature is where new revenue is coming from.
- Market size: A market where people will always be needed, such as compliance, where getting it wrong can mean legal consequences and customers are less willing to trust automation, or a market with momentum, such as software that helps build data centers.
- Sales efficiency: Efficient automated content producing good MQLs, automated qualification producing SQLs, and a small team of strong closers. A business operating that way is likely dropping a meaningful amount to the bottom line.
For founders unsure where their business lands on these metrics, that assessment is worth having early, well before a process begins. Vista Point Advisors is available to help founders assess where their business sits and understand their options.
This material and the opinions contained herein are for general information only and are not intended to provide specific advice or recommendations for any individual or entity.