Buying a SaaS Business: What to Pay, Based on What Buyers Actually Paid

The median US software, SaaS and app business sold for $625,000 over the five years to 2025, on median annual revenue of $483,261 and median owner earnings of $208,866. That is 2.99 times owner earnings and 1.29 times revenue. If you are buying one, that is the number to anchor on, and it is a long way below the ARR multiples every valuation article quotes.

Almost everything written about SaaS valuation describes a market you are probably not buying in. The 5x to 12x ARR figures are real, and they come from venture rounds, growth equity deals and public comparables, where the company has a management team, an engineering organization and investors who will not sell below a mark. The owner-operated product doing $400,000 a year, built and still maintained by one founder, trades somewhere else entirely. This is what that market actually paid.

How much should I pay for a SaaS business?

Price it on seller discretionary earnings, not ARR. Software businesses that sold went for 2.38x SDE at the lower quartile, 3.12x at the median and 4.00x at the upper quartile. Divide the observed median sale price by the observed median owner earnings and the multiple that actually cleared the market is 2.99x. On $200,000 of verified owner earnings that is a fair opening range of roughly $476,000 to $800,000, with $598,000 as the middle.

Owner earnings here means net profit plus the founder's own compensation, benefits, interest, depreciation and any one-time costs. For a business at this size that is the right base, because what you are buying is an income stream that currently pays one person. Revenue multiples are a cross-check, not the method. The full quartile spread, the source data and the arithmetic behind it sit on our SaaS business valuation page.

You pay the upper quartile for a specific set of things: churn low enough that revenue holds without constant selling, customers spread widely enough that losing the largest one is survivable, and a codebase somebody other than the founder can maintain. You pay the lower quartile when any one of those fails.

Why are SaaS asking prices so much higher than what businesses sell for?

Because most of the listings never sell. Software sellers ask a median of 4.35x owner earnings. The market pays 3.12x. Divide the sold multiple by the reported 0.89 average sale to ask ratio and you recover what the sellers who actually closed were asking: 3.51x. So of the 28.3 point gap between asking and sold, only 11.0 points is negotiation. The other 19.4 points is selection. Sellers asking well above 3.51x are not negotiated down, they simply sit.

That distinction matters when you are deciding which listings to spend time on. A software business listed at 4.35x is not a business you will talk down to 3.12x. It is a business that will still be listed next year. The ones that trade were priced near 3.5x from the start.

Software carries the largest selection penalty of the four sectors we have run this arithmetic on. Restaurants come out at 17.8 points of selection and 10.0 of negotiation, accounting practices at 9.3 and 3.0, and gas stations invert it entirely at plus 3.8 and zero. Software, at 19.4 and 11.0, has both the widest wishful-thinking gap and the most room at the table.

One more thing the record shows, and it cuts against instinct: the software businesses that sold were smaller than the ones still on the market. Sold businesses carried median revenue of $483,261 against $602,500 for current listings, and median owner earnings of $208,866 against $275,010. Buyers picked smaller and simpler over larger and more entangled, by roughly 20 to 24 percent on both measures.

Can you buy a SaaS business with an SBA loan?

Yes, and for this size of deal the SBA 7(a) program is the usual route. At a ten year term, 10.5 percent, a 1.25x debt service coverage requirement and a 10 percent equity injection, the median $208,866 of owner earnings supports a purchase price of about $1,146,000, or 5.49x. The market pays 2.99x. Financing is not the binding constraint on software prices the way it is in fuel retail, so you have real headroom, roughly 1.8 times the clearing price.

The catch is the replacement hire, and this is where most software deals get decided. If the founder writes the code and you do not, the lender underwrites the earnings after the salary of whoever replaces them. Every dollar of that salary comes straight out of your borrowing capacity.

Run that to its crossover and the number is about $95,100. If the business needs a replacement costing roughly that much, the SBA ceiling lands exactly on the 2.99x the market pays, and above it debt service stops covering the market price. So the practical question for a non-technical buyer is not whether you can get a loan. It is whether the product needs a $60,000 contractor or a $120,000 engineer, because that single fact moves your ceiling by more than two turns of earnings.

What should I check before buying a SaaS business?

Verify that the recurring revenue is recurring, that the earnings are real, and that the product survives the founder leaving. In that order. Ask for Stripe or payment processor exports covering at least 24 months, monthly cohort retention rather than an annual churn average, revenue concentration by customer, and bank statements you can reconcile against the claimed deposits.

The reconciliation step is the one people skip and the one that changes prices. Claimed MRR and money that landed in the account are different quantities, and the gap is usually annual prepayments, failed rebills and refunds. Work through the statements line by line for the two most recent years, and if you are taking the business over you will want that history in your own books anyway, which is a good reason to move the prior year of banking into QuickBooks before you close rather than after.

On the product side, ask who has deployed to production in the last six months, whether there is a staging environment, and what happens to the service if nobody touches it for ninety days. A single founder who has been the only committer for four years is a real risk even when the revenue is clean, and it is the difference between the lower and the upper quartile of the price range above.

Concentration deserves its own look. A business at $400,000 of revenue with one customer at 30 percent of it is not a $400,000 business, it is a $280,000 business with a call option. Price it that way, or structure the difference into an earnout.

What multiple do SaaS businesses actually sell for?

On earnings: 2.38x at the lower quartile, 3.12x at the median, 3.31x on average and 4.00x at the upper quartile. On revenue: 0.97x, 1.51x, 1.70x and 2.43x. Those are closed sale prices across 2021 to 2025, not asking prices, and they describe owner-operated software rather than venture-backed companies.

It is worth knowing how far that sits from the numbers in circulation. Apply the commonly quoted 5x ARR to the $483,261 of revenue the median sold business carried and you get $2,416,305, against the $625,000 that changed hands. At 8x it is $3,866,088, at 12x it is $5,799,132. The overstatement runs from 3.9 to 9.3 times. The ARR multiple is not wrong, it is just measured on a different population, and a seller quoting it at you is quoting a market neither of you is in.

Check the base of any figure you are handed, too. Quoted multiples for this category are often drawn from the asking distribution rather than the sold one, and on software those are 4.35x versus 3.12x on earnings and 2.59x versus 1.51x on revenue. That single confusion is worth more than a turn of earnings on a typical deal. The same pattern across other sectors is in our SDE multiples by industry table, and what multiple a business sells for covers how to place a specific business inside a published range instead of defaulting to its midpoint.

Is buying a SaaS business a good investment at these prices?

At 2.99x owner earnings, a software business pays back the purchase price in about three years if earnings hold flat, which is a better yield than most main street categories. The risk is not the price, it is whether the earnings hold at all. Software has the highest owner margin of any online and technology category at 43.2 percent, and the worst average sale to ask ratio at 0.89, and both facts come from the same source: high margins with fragile, easily replicated products.

The five year record inside this dataset makes an unusual point about what buyers reward. Between 2021 and 2025 the median software business that sold went from $651,989 of revenue at a 24.6 percent owner margin to $290,238 at 83.2 percent, and the median sale price still rose 45.8 percent, from $562,500 to $820,000. Buyers paid substantially more for less than half the revenue, because the earnings were higher. If you are evaluating a target that cut a costly acquisition channel and shrank, that is not automatically a decline.

Compare across the category set before you conclude software is cheap or expensive. On the effective multiple, meaning median sale price divided by median owner earnings, websites and ecommerce sit at 3.06x, graphic and web design at 3.06x, all online and technology businesses at 3.04x, software at 2.99x, IT and software services at 2.83x and cell phone and computer repair at 1.93x. Software is mid-pack on price and top of the table on margin, which is the trade you are actually making.

Median time on market was 169 days, so there is no rush in this category. The listings clear slowly, the ones priced above about 3.5x mostly do not clear at all, and the businesses that trade are the smaller, cleaner ones. Run the target's numbers through a business valuation calculator before you make an offer, so the range you are negotiating against comes from closed transactions rather than from the seller's spreadsheet.

Transaction figures throughout are from BizBuySell software, SaaS and app business valuation benchmarks covering businesses sold between 2021 and 2025. Multiples described as effective or market clearing are computed here from the published medians. Last updated September 2026.